Hand Sanitizer, Startups and Still Waiting for $120,000 from Amazon

In March 2020 when COVID-19 struck, I was sitting at home in Cambridge thinking through how I would continue to make payroll and keep my teams busy at Sandymount Technologies and Point 5 Brewing. Inspired by a photograph in the Financial Times of a distillery in Europe that had converted to producing hand sanitizer, I thought about how I might quickly build a supply chain to sell hand sanitizer.

I had been shopping at Trader Joe’s and they had no hand sanitizer on the shelves. Any shops I called – Wholefoods, Target, CVS – were out of stock. If I could make it – I thought – I could sell it. It didn’t turn out to be that easy! I got onto social media and drew up a post on LinkedIn and Facebook asking friends if they would be interested in setting up a supply chain. Friends from MIT, Anton Hunt found a supply for alcohol, and Gina Brooks-Zak found a manufacturer that could convert to filling bottles. Together, our initial team drafted fliers touting the product and started to circulate them online and through connections. It was a chicken and egg problem. We would have the bottles and labels lined up, and then our alcohol supplier would suddenly sell the alcohol off to someone who could pay in cash. Conversely, at times we would have a large order lined up, but we would be scrambling to secure the alcohol.

Revenue started to come in with orders from homeless shelters, car washes, golf clubs and fire stations. At one point we cleared $100k in purchase orders and we started to take payment. The problem was, once we started to get in that money, we were on a clock to hit our promised delivery date, but we couldn’t start producing until we had $200k to buy all of the materials. Every day we couldn’t get to $200k in sales was another day of delay, another day where our supply chain of bottles and alcohol might fall apart because we didn’t have the cash to secure it in place. I didn’t want to spend any of the money we had received until we had the full $200k – it wouldn’t have been responsible to buy everything and then not be able to deliver.

It was a Friday towards the end of March. Anton, Gina and now Adam Weiner and Dana Hemmert from Sandymount, as well as my cousins John and Barry McGovern were doing everything to try and find a large buyer that would make the cashflow work. Chris Lazarte, who works with me at Sandymount helped me to set up an LLC for the operation because we didn’t have the time (never mind the expense) to set up a non-profit. Everything was ready to go, but large buyers kept on dropping out. Anton was offering all kinds of discounts to get a big buyer to jump but we couldn’t do it. I had to pull the plug. We refunded the money and e-mailed clients and suppliers apologizing. It was pretty bad – I had gained the trust of these buyers only to disappoint them. I didn’t want a situation where we were holding people’s money and not delivering the product.

The next day, Saturday morning, Anton Hunt gave me a call. One of his partners in a separate business had a huge demand for sanitizer but didn’t have a supply chain in place. Importantly, they had the funds to get the project rolling. A hastily pulled together deal over the weekend reignited the whole project. Anton and I were joking how this is the project that just wouldn’t die. For the next two months, our team held an online meeting every morning at 7.45 am. We would go over the logistics for the day, the orders of ethanol, tracking down bottles and bottle caps that were delayed, getting boxes, labels, hydrogen peroxide and glycerin ordered. It took us three weeks to get the production lines up to full speed – making over 40,000 bottles per day when we had two sites on-boarded. It was never plain sailing, there would always be some shortage of ethanol, some bottle caps that had been delivered without seals, or some federal alcohol permit that we had to start calling senators about to help us get it through – which they did!

Over the three-month period, the business was involved in producing over 1.2M bottles of hand sanitizer, selling primarily through Aero Healthcare in New York, partnering with SPEC Engineering in Chicago, Califormulations in Georgia, and then – towards the end – selling to Amazon for them to resell online. In June, it became pretty clear that the market was getting back to normal. We found it increasingly hard to find buyers of large quantities and it was time to wind the business down and direct efforts towards donations. The sign to move on was when Amazon decided not to place a further order beyond the 200,000 bottles they had purchased. We barely got in that order with Amazon before there was a surge of supply. I think we were the last supplier to be approved for 7-day payment terms, compared with the 90-day payment terms Amazon normally offers suppliers.

Over the next months, our team wound down operations, sold off any remaining bottles and alcohol, and started to direct donations to charities, with $500,000 donated as of the end of August, split between five charities: Fathers Uplift Inc.; Y2Y Network; Fresno Barrios Unidos; My Stuff Bags Foundation; and Urban Revival Inc., as well as 6,500 boxes donated to the Food Bank of Northern Indiana. The only thing we are waiting for now is for Amazon to pay out the remainder of their purchase order, which is for a bit over $120,000. It has been a real challenge working with the supplier portal on Amazon. Our last payment is now overdue by over two months and it isn’t clear how you resolve this kind of issue with a company like Amazon. We tried their dispute procedure, but it hasn’t worked to date. The company is very large, and it is hard to connect with a person who has the authority to fix the problem. Absent any other clear way to communicate with Amazon, I’ve pulled together a letter and sent it to Jeff@Amazon.com , sending hard copies to both him and also Jeffrey Wilke – who runs Global Consumer. $120,000 seems like a lot of money for us, but at least our team isn’t reliant on that money for income – it is destined for charity. I can imagine there are a lot of suppliers in a lot worse of a position than us.

I’m hopeful that the issue will be resolved because it’s a good outcome to have been able to provide hand sanitizer when it was needed, it’s a good outcome to have helped the charities, and I think it can be a good story for Amazon too. For the team involved, I think we’re lucky to have had this opportunity and that our skills and connections came together in a way that made the project work. It has been a huge highlight being able to work with Anton, Gina, Adam, Dana, Barry, Chris and John this year – and I hope I get to do many more projects with this level of excitement during the rest of my life.

Subscribe here to get my updates mailed to you once per month:

Signing you up…
Kindly check your e-mail to confirm your subscription.

Warren Buffett’s Well Worked Will


  • A flood of BRK shares to the market would put downward pressure on price.
  • Buffett’s donation program is spread out over time ensuring liquidity to the market is smoothed over time.
  • Buffett has been donating shares since 2007, so there should be no abrupt change in shares entering the market over time due to donations.

As a Berkshire Hathaway shareholder, my top concern is for the future of the company after Warren Buffett’s time. While the greatest loss to Berkshire will undoubtedly be Mr. Buffett’s business acumen, there is also the question of how the distribution of Berkshire shares in his will would affect the supply and demand – and hence the price – of Berkshire shares in the market. Fortunately, Mr. Buffett has thought through this issue carefully and, I believe, largely mitigated associated risks by i) spreading out the donation of shares over time and ii) gradually reducing the amount of shares donated each year so that the dollar value of shares entering the market via share sales by foundations is somewhat stabilised.

Background on Mr. Buffett’s donation structure

In 2006, Mr. Buffett announced that the vast majority of his shares in Berkshire would be donated to five foundations. The announced donation amounts, expressed as BRK.A shares, were as follows:

Recipient | Announcement Date >20062012Total
Bill and Melinda Gates Foundation333,333333,333
Susan Thompson Buffett Foundation33,33333,333
Howard G. Buffett Foundation11,6668,14719,813
The Sherwood Foundation (formerly Susan A. Buffett Foundation)11,6668,14719,813
NoVo Foundation (Peter A. Buffett)11,6668,14719,813
Total BRK.A Shares401,66424,441426,105

As per the above table, Mr. Buffett committed to donating the equivalent of approximately 426,105 of his A shares to these charities. In total, Mr. Buffett originally owned 474,998 BRK.A shares, meaning that about 90% of his Berkshire shares will be donated to these organisations.

Warren Buffett’s donation strategy

US charities by their nature are required (subject to certain allowances) to spend funds they receive within a relatively short space of time after receiving donations. Had Mr. Buffett decided to make the above donations in one fell swoop, that would have resulted in a huge surge of shares becoming available on the market. As an illustrative example: Considering the July 31st 2020 closing BRK.A price of about $300k per share, there would be about $128B of liquidity entering the market at once.

To avoid this issue, Mr. Buffett has spread out the donation of shares as follows:

  1. For many decades, Mr. Buffett pledge not to make any donations until the settlement of his estate. In 2006, he announced a change of course and started donating shares in 2007. By steadily donating shares through the rest of his life and then beyond, the potential for a sudden inflection point upon his passing was removed.
  2. The shares listed in the above table have been donated at a rate of 5% per year since the year following each announcement. Importantly, donations are made at the rate of 5% of the remaining balance after the previous year’s donation – not 5% of the originally pledged amount. This ensures that – while the stock price grows on average – the number of stocks donated per year falls. The result is that the dollar value of donations per year is somewhat stabilised – smoothening liquidity  which means that liquidity to the market is tempered, and, donations received per year by the charities too.

As of July 2020, Mr. Buffett has already donated 248,734 BRK.A shares – constituting just over half of his original stake. The donation of the rest of his shares will occur smoothly over time as per the plan described above. For further context, the amounts of shares being donated per year (about $2B per year in market value as of July 2020) are reasonable relative to the overall market capitalisation of the business (~$475B as of July 31st 2020) and relative to the total amount of shares being bought back by Berkshire Hathaway itself (on the order of single billions of dollars per year in recent years), which serves as one source of demand to mop up the supply of foundations selling.

Rest assured, Mr. Buffett has well planned the distribution of his shares to charities before and after his time.

Notes & Materials:

Buffett originally made a pledge measured in 2006 B shares, when 30 B shares were equal to 1 A share. Subsequently, B shares were split 50:1, resulting in a B to A relationship of 1,500:1 .

Spreadsheet formulated using Berkshire Hathaway information: BRK_donation_schedule.xlsx

Subscribe to my Monthly Newsletter:

Signing you up…
Kindly check your e-mail to confirm your subscription.


How hard is it to create an eCommerce store?

Over the weekend, I set up a new website for Point 5 non-alcoholic beer using Shopify. Previously, I had been using wordpress/woocommerce in a way that was taking a lot of time and money. Using Shopify made me realise and appreciate how web tools are now at a level where someone like me – without web coding experience – can get a fully functional online store up and running. The fact that Shopify offers a full service platform makes things very easy (e.g. hosting your website, providing a simple interface to design the website, providing payment solutions, providing integration with Google, Facebook, Amazon and Walmart shops).

Being able to easily build an eCommerce store like this is a very beautiful thing because:
1. Instead of spending time or money on coding a website, I can spend that time and money on improving the product and marketing the product.

2. Since I built the website myself, if there are things that need to be fixed, I can easily do them myself. Making quick changes like this is very valuable in the early stage of a business.

If any of you are thinking of doing an online store – even just for fun – I highly recommend it. The technology that is available nowadays is amazing and allows anyone to go from zero to store in less than a day.

Now on to the harder challenge of marketing and driving sales 🙂


Things I’m thinking about in July 2020

1. Sandymount Draft

You might remember seeing a video last year of Sandymount’s draft dispense technology, allowing beer or non-alcoholic beer to be dispensed at a bar from a tiny keg of concentrate, by adding CO2 and filtered water. Over the last months the Sandymount team has been working hard to develop a user interface that allows operators (and soon also consumers) to see the temperature, alcohol % and carbonation of the beer (or non-alcoholic beer) being poured in real time. Stay tuned for a video update.

2. Point5Brewing.com – Non-alcoholic Beer
Point 5 continues to sell online for home delivery – and now at a significantly reduced price point. Point 5 non-alcoholic beer is designed for an easy-drinking real beer taste and, if it is to make a dent in people’s lives as regular refreshment, it became clear that a lower price point was necessary. Much of what will allow this price improvement is a move to a canned format in the autumn, which will significantly reduce the weight and size of packages to be delivered. Until then, we’ve dropped the prices on bottles anyway, so pricing should stay consistent through the transition. www.Point5Brewing.com/shop is where you can find it.

3. ShopRespond.com – Hand Sanitizer Business
From the end of March until the end of June, the ShopRespond team managed to pull together a supply chain from scratch and produce over 1.5 million bottles of hand sanitizer. Over the last weeks, we have been winding down operations and are almost done with donating the proceeds from this initiative. Well done to Anton Hunt, Adam Weiner, Dana Hemmert, Gina Brooks-Zak, Chris Lazarte, John McGovern & Barry McGovern. Press release to follow shortly.

4. Risk Savvy – A book by Gerd Gigerenzer

A practical book on the benefits of keeping things simple when making complicated decisions.

5. You can now use a Quantum Computer

Why would you want to do this? Quantum computing is apparently still in its early days in terms of being useful for anything, so the best reason to do it is probably for the craic. IBM is making available a few quantum computers to mess around on (remotely). I’ll let ye know how trying it out goes. You can too here: https://quantum-computing.ibm.com/docs/guide/iqx/the-quantum-world

6. Rockets look like fun

Lovely technical video here on rocket design from Tim the everyday astronaut.

7. Internet may come from the sky instead of cables?

SpaceX and OneWeb are lobbing satellites into low orbit – low meaning close enough that they could send internet down to earth without it being too slow. One of the problems right now is how to cost effectively build receivers. There’s a decent article from CNBC.

What does this mean for Ireland’s broadband plan? Could we end up laying a ton of fibre only to find internet will come to us from the sky?

As per usual, I’m keen to chat or learn more on all of the above, reach out. Cheers.

Footnote: What happened to the Desalination Barge Concept? It’s still there lurking in the background for right opportunity/timing – and may continue to do so for years 😉


Making it easier for you to try non-alcoholic beer

Over the last few months, the Point 5 team has been selling on a third party online store. As of today, Point 5 is now also selling directly on our website at http://www.PointFiveBrewing.com . By selling on our own website this gives flexibility in a few ways:

We felt it was too expensive for new fans to taste Point 5

Many of you are tasting a non-alcoholic beer for the first time. Until now, it was $36 (when you include shipping) to buy a 12-pack. This is a steep price to try for the first time. To tackle that, we’ve done two things:

1. We’re making available a Starter Pack (3 bottles) for $9.99 – including shipping.

2. We’re making it a flat price of $29.99 for our 12-pack – including shipping.

Basically, we found it’s annoying to see the price of a product and then have to pay for shipping on top, so we’re simplifying to all-in pricing, with just state sales tax that would be added on top.

For those of you who do order the Starter Pack, you’ll also be getting a welcome note with a discount on your first 12-pack order. Keep an eye out for that.

Let me know your suggestions for how we can keep improving the website, and make the Point 5 experience more fun and enjoyable. Cheers, Ronan (+Dana + Ad!)

P.S. You can reach our “Make it a Point 5” campaign launch blog here.


A brief story about selling hand sanitizer online.

When the COVID-19 Response team (Anton, Ad, Dana, Gina, John, Barry) got together, one of the issues in getting hand sanitizer production started was finding the money to pay for all of the ingredients and bottles upfront. Thanks to a distribution partner we were able to get funds to kick production off. Selling online was hard at that point though because – for online sales – you more or less have to have the product ready (or ready soon) to be able to sell.

Now that we’ve been up and running a while, we’re in a position to start selling online and are doing that in 32 packs of 8 oz bottles on our new website http://www.ShopRespond.com . The thinking is that businesses will need these quantities as things start to open up again.

Right now we’re just selling to 48 states in the US and have 8 oz bottles. The hope is to expand out to doing other supplies soon – likely starting with 64 oz sanitizer bottles. Here is a quick video on the product.


Your ideas for making a lot of simple fabric masks

By now, many of us are wearing masks when outdoors. I’ve found it tricky to get a comfortable fabric mask, and I’ve been lucky that my housemate’s girlfriend made me one. It was much appreciated.

Some thoughts:

Not everyone’s head/face is the same size.

I imagine it’s tricky finding smaller masks to fit kids.

Wearing a mask is tricky because I wear glasses. They make the glasses wobbly and can lead them to fogging up.

It’s nice to have a mask that is a decent fit but also is loose enough to be able to talk – say, on a phone call.

Many masks squash my nose.

Some questions:

Have you any ideas on companies (maybe clothing companies?) that could make a large amount of simple fabric masks of different sizes? Say between 10,000 and 100,000 per week?

Are there simple fabric mask designs I should look at that address some of the issues above?

Let me know if you’d like to help finding out how to get a lot of masks made. If so, I should be able to get them on an online store that should open soon with the COVID-19 initiative I’ve been working on.


Ronan’s Live Pre-Stream to the Berkshire Hathaway Annual Meeting (Warren Buffett’s Company)

Warren Buffett is likely the world’s most famous living investor. Each year, his company (Berkshire) holds a meeting with over 40,000 people where he takes questions for about six hours. This year, that meeting takes place at 4 pm Boston time (9 pm Irish time) on Saturday May 2nd and will all be compressed into a one hour meeting by video.

Right before the meeting, I’ll be doing a 30 min pre-stream – starting at 3.30 pm Boston time and 8.30 pm Irish time – to cover:

A little background on how Berkshire/Buffett makes money, including a very rough tour of the company’s cash flow and balance sheet.

Some comments/reaction to Berkshire’s results from the first quarter of 2020 (which likely reveal what Buffett did (or didn’t do) with his money during the market crash due to COVID). Note that these will only be realised early morning on May 2nd.

A preview of a few questions that Buffett is likely to get asked during the 45 min Q&A session.

Some links that may be of interest:

  1. A warm-up video if you’d like to hear the latest on Berkshire and from Buffett.
  2. A the link to the actual live stream of the meeting – starting at 4 pm Boston time (9 pm Irish).
  3. A link to my live pre-stream commentary – starting at 3.30 pm Boston time (8.30 pm Irish).

Has your business got PPE for when the economy reopens?

It’s a thought that dawned on me when chatting with my mother today. I feel that people are so focused on figuring out when the economy will reopen that we haven’t started to plan yet for what will happen when things do reopen. Are we planning for what will be needed to go about daily life and work?

I’m not saying we should reopen soon. I’m just thinking that once some dates are announced around reopening, there might be another big rush to get masks, sanitizer etc.

I’m not sure how this will play out, but it’s something I’m thinking about as I plan for continued production of hand sanitizer. Will all businesses initially recommend wearing masks? Will there be guidance around hand washing or sanitizing? It seems like there is a good chance the answer is yes.


Taking large orders for 8oz hand sanitizer

Hi folks, Covid-19 Response LLC is taking orders for order sizes of 200,000 to 1,200,000 bottles of 8oz hand sanitizer.

The next available batch will ship on Friday May 8th.

Please e-mail StopCorona@Sandymount.com for inquiries.

Respond Hand Sanitizer ready for packing – 8 oz fliptop bottles


Respond Hand Sanitizer – 8 oz fliptop bottles


Some ideas I use for investing. Part 3.


Price matters to me when I am investing. The lower the price the more I invest. To get the best return, I don’t want to be too diversified. I want to have my eggs in a small number of baskets. Lastly, I buy and hold because this reduces the tax penalty of selling/buying stocks.

The Kelly Criterion.

The more I diversify across a wide range of stocks, the more I get the return of the market. To beat the stock market, I need to have an opinion on the price of specific stocks. To maximise my returns, I need to bet more when I am confident in my opinion, and bet less (or not at all) when I am not confident. Mathematically, this betting pattern is described by the Kelly criterion. This is what card counters use when playing blackjack. This is is how Warren Buffett invests. It is true that Berkshire is more diversified now than it was in the past, but Berkshire has a very large percentage of it’s assets in a small number of companies (like Geico, BNSF and, more recently, Apple). Berkshire has outperformed the market in the past – not by copying the market – but by making big bets on companies that massively outperformed the market.

I am still working on better ways to implement this kind of behaviour in buying stocks and I am not yet fully systematic. Take Facebook for example, which I consider good value at $178 per share. When the stock falls below this value (it has) I put a certain percent (say 3%) of my liquid assets in this stock. If the stock falls below 75% of this value – $134 per share, I would put in another 3% – and so on. It is important for me to be disciplined about setting rules and sticking to them. Stock prices jump around and I’m not able to predict whether stocks will go up or down at any given time. This approach is very conservative, so I’m not buying stocks very often.

Note: I also set limits and won’t put more than 10% of liquid assets into one stock (although I won’t sell if a stock appreciates by itself and goes above that threshold). I am considering loosening that.

Buy and Hold.

In the US, I pay capital gains tax (~15% + state tax) on profits when I sell a stock that I have held for more than one year. If I sell in less than one year, I pay ordinary income tax (more than double). Clearly this is a reason to hold for at least one year, although I plan to hold for much longer.

Capital gains tax is paid when I sell a stock after holding for one year. There is a critical mathematical point (well described in Charlie Munger’s Almanac) that as I delay selling, the impact of the tax penalty on my annual return becomes smaller and smaller. Take the example of a stock that grows in value by 10% each year, and pays no dividend. Consider a capital gains tax rate of 15%. If I sell the stock after one year, the capital gains tax will reduce my after tax annual return from 10% down to 8.5%. However, if I hold on to the stock for 30 years, the effect of tax is to take my annual return down from 10% to 9.44% instead. If I held the stock for ever, the effect of the tax would tend towards zero – this is effectively what Berkshire does with most of the major stocks it owns. That is Warren Buffett’s claimed holding period – forever. One last point from Buffett – I think about what stocks I would buy if I could only buy ten over my whole life – that forces me to be selective!

After tax annualised returns assuming 10% annual stock growth and 15% capital gains tax. Spreadsheet here.

Technical note on Berkshire: On the surface, the Berkshire structure looks very inefficient tax-wise. For example, Berkshire owns a large stake in Coca-Cola. Coca-Cola pays corporate tax (22% headline rate) on any profits it makes. Of those profits, it distributes dividends to its shareholders (Berkshire). Berkshire then pays tax on those dividends – so there is a double tax. However, the reality is not as bad than this. Yes, Coca-Cola pays tax on its profits, however, it only redistributes a portion of those profits as dividends – with much of the rest being reinvested in ways that grow the value of the stock. Berkshire may never sell its Coca-Cola stock, so the impact of the tax of selling that stock is largely mitigated. Other Berkshire stock investments (like Apple) spend most of their profits after tax on reinvestment or on stock buybacks – both of which increase the Apple stock price. Again, Berkshire doesn’t get impacted much by a double tax here because they are long term holders of stock. In short, for Berkshire, holding stocks reduces the disadvantage of being double taxed. For me, holding stocks provides an advantage that reduces the impact of capital gains tax.

Side note on tax policy: I think it is good that capital gains tax is set up in a way that discourages short term holding of shares because I think long-term thinking is healthy for the market.



Some ideas that I use for investing. Part 2.


I pick stocks using rules that select for companies that have a low levels of debt, that have a trusted CEO and are priced at a level that is cheap relative to how much cash they generate. The lower I feel the price is, the more money I invest. Lastly, I stick with my bets by buying and holding, which reduces the tax penalty of selling/buying. I’ll cover these last two points in Part 3. My strategy is more an exploration and learning process than a refined process and I expect will continue to evolve.

Low debt:

I don’t want to lose money. I especially don’t want to lose money when most people and businesses are losing money. To the extent I can lose less money than the average person during a downturn, I can take advantage of lower prices when others are selling. Mark Spitznagel provides mathematical reasons why losing money is bad, not just in itself, but also for average returns. I think about protecting the downside in buying stocks as well as finding good upside for the long run.

Companies with high debt tend to have stock prices that fall drastically if the economy takes a downturn. Debt seems sustainable when the market is doing well – in such times the market often values companies based on profits or even just revenue. In bad times, companies with high debt struggle. Compare Exxon Mobil and Occidental Petroleum – both hit by the recent market and oil price falls. The two companies have different market exposures, but a big cause of Occidental’s stock price is – in my opinion – the high level of debt that Occidental has (as a result of its recent acquisition of Anadarko). Occidental’s level of long term debt was about $39B at the end of 2019. The combined revenue of Occidental and Anadarko for 2019 was about $29B. For comparison, Exxon had about $26B in debt (less than Occidental) at the end of 2019, but is a company with revenue of $256B (almost ten times the size of Occidental).

Chart of Exxon Mobil and Occidental Stock Prices
Chart of Exxon Mobil and Occidental Stock Stock Prices during Coronavirus and the Oil slump in early 2020

Picking an “acceptable” level of debt is subjective. I screen out any companies that have a ratio of long term debt to adjusted cash-flow that is below five. In simple terms, if a company doesn’t generate enough cash annually to pay off it’s debt in five years (without compromising on necessary capital investment), I don’t want it. I find the long term debt on the company’s balance sheet. I calculate adjusted cash-flow by looking at cash flow from operations (on the statement of cash flows) and subtracting off depreciation (on the P&L or statement of cash flows). For software companies like Google, there isn’t much depreciation. For industrial companies, depreciation is a real cost because companies need to invest capital into equipment/land/assets in order to generate that operating cash flow. In other words, a portion of the operating cash flow is needed to pay for new equipment just to keep the business going as is, so I have to make an estimate of how to subtract that out. Note: I could subtract out capital spending, but that tends to fluctuate more year on year. Also, if the company is growing, capital spending will outstrip depreciation – so I would be subtracting out too much. Mature companies that are growing more slowly will have annual capital spending and depreciation that are closer in value – compare Southwest airlines with comparable capital spending and depreciation to Ryanair (more growth and more capital spending in 2019).

Of course, there is subjectivity in how you pick the adjusted cash flow and how many years you want to average over. Some companies that I considered passing this screen at the end of 2019 were: Ryanair, Southwest, Berkshire Hathaway, Amazon, Apple, Facebook, Google, Spotify, JetBlue.

Trusted CEO

My heuristic for a trusted CEO is simple. Either the CEO has to have a majority of their net worth in stock of the company (not options) OR Warren Buffet has to have invested in the company. This is a strict rule, but I think a good one to ensure alignment with stockholders. Of the list above this would leave the following as qualifying: Ryanair, Southwest (via Berkshire), Berkshire Hathaway, Amazon, Apple (via Berkshire), Facebook Spotify, (I include JetBlue as well because I think Buffett would own it but the market cap is too small for him to be able to invest).

There are other companies that could be in there, but I haven’t had enough time to read the annual reports of a lot more qualifying companies. I prefer to learn about a few companies at a time even if it limits me on the pool I’m picking from.

Low price relative to cash generated

Over the long run, I believe that cash generation matters. Quite simply, if a company is generating cashflow from operations, they can fund growth without having to issues shares or debt. If they are generating a lot of cash, they can either support fast growth, or they can distribute that cash (via dividends or stock buybacks). The decision of what is done with the cash is important, which is why a trusted CEO is important. The level of cash generated is also important.

For mature companies (with less than 20% year on year revenue growth), I screen for companies that have adjusted cash flow (see above) equaling 10% or more of the stock price. For example, if a stock is $50 per share, I want to see $5 per share of adjusted free cash flow. For growth companies (ones growing revenue by 20%+ per year), I screen for companies with adjusted free cash flow equaling about 7% or more of the stock price.

Note: This cash flow based screening is just one approach and it misses a lot of companies that could be good. For example – Amazon is at a high price relative to its cash flow, but could still be a good investment. I haven’t figured out a strategy for low cashflow companies that I’m comfortable with.

Summary and stock holdings

As of now (April 2020) the main stock holdings I have are Berkshire (10% of total assets), Southwest (5%), Ryanair (5%), Facebook (3.5%), JetBlue (2%). I have been holding Berkshire, Southwest and Ryanair for quite some time, and picked up Facebook and JetBlue recently as their prices fell. I’m not confident that aviation will come out well out of this (although low fuel prices and reduced competition due to bankruptcies may help), but I am staying the course. For clarity, this is about 25% in stocks at present. 6% gold index. 6% Bitcoin. The rest (63%) I hold in cash (since TBills went to zero).

Note on Berkshire Hathaway. It is a bit of work to figure out cash flow for Berkshire because it owns significant stakes in other companies whose cashflows are not reported in Berkshire’s consolidated financials. I have to go in to those companies (Apple, American Express, etc.) and figure out what share of cashflow can be attributed to Berkshire.


Some ideas that I use for investing. Part 1.


A simple way to invest is to split my money 50-50 between stocks and bonds – that is what I do with my 401k (retirement plan). For my other savings, I loosely stick to this 50-50 split, but I do pick specific stocks and I do that following certain rules.

Overall, there are two goals to how I am investing. My main goal is to avoid losing money. This is the majority of my strategy. My secondary goal is to invest in things that, in the long term, I think will do well. I use certain rules to identify what I think will be safe and will do well.

In short here is why I think it is very important not to lose money:

When things are going well with the economy, it is nice if my investments are doing well. However, I am more likely to have a job and so are my family. So getting a good return on my investments is nice but not essential.

When things are going badly with the economy, it is pretty annoying to lose money. On top of losing money, I have a bigger chance of losing my job and I need the money more than when the economy is going well.

In short, it is often better for me to avoid losing money when things are going badly then to make a lot of money when things are going well. Furthermore, if the economy is doing badly then many things gets cheaper to buy (e.g. houses, stocks), so if I haven’t lost too much money, I am in a good position to buy things.

The Ben Graham 50-50.

Ben Graham wrote a book called The Intelligent Investor. One of the investing strategies discussed is to invest 50% of money in stocks and 50% in bonds. Here is why I think this is a reasonable thing for me to do:

  1. Why not put more in stocks? Many people recommend putting much more in stocks. For example, Warren Buffett has suggested putting 90% in stocks (the S&P500 and 10% in bonds (although he doesn’t do that with his company’s money). I think that is a bad idea. With 90% of my money in stocks, the stock market could very well fall right before I need the money (say, for a mortgage down-payment or for a wedding or for retirement). Is it really worth rolling that dice to get some upside?
  2. Why not put more in bonds – wouldn’t that be safer? Keeping money all in bonds (or cash) also has risks, such as inflation or an increase in interest rates. If I had all of my money in bonds (especially of mixed or long durations) right now when the interest rates are near zero, the value of those bonds would plummet if interest rates rise to even 3 or 4%, never mind 10 or 15%. I’m not saying that will happen, it’s just a risk I don’t want to take.

When I look through the last 150 years, pure stock and pure bond portfolios have each had large falls. With a 50-50 portfolio, there are significant dips, but not nearly as bad. The other nice thing about holding a constant split of 50-50 is that when stocks down up relative to bonds, I’m buying more stocks. If bonds go down relative to stocks, I’m buying more bonds. By definition, I’m buying low and selling high.

Technical note: For stocks, a simple approach I use is to buy an index fund like SPY. For bonds, a simple approach I use is to buy a broad based bond index fund (that has bonds of different durations/lifetimes). This means I don’t have to think about what specific stock or bond to buy (and risk being wrong).

Avoiding risks I may not be able to avoid with just stocks and bonds.

There are things that can happen that are bad for both stocks and bonds. For example, very strong inflation of the US dollar might be bad for US stocks and US bonds. These are the kinds of things that are quite unlikely to happen, but if they do happen they are possibly very very bad for stocks and bonds. Gold, since it is limited in supply, tends not to devalue when there is inflation, so I have a small amount of gold (via an index fund) – roughly 5% of assets. The reason for just holding a small amount is that the risk of needing the gold is low, but it’s good to have a bit.

Side note: I also have about 5% bitcoin. This is maybe crazy. However, I see some analogy to gold in that the supply of bitcoin is limited and it takes work/money to increase that supply.

Other side note: I have considered investing in a real estate (REIT) index fund. I don’t consider that I understand them much and I don’t have any at the moment.

Loose 50-50

With non-retirement savings, I am a bit loose on the 50-50 rule. When I feel that equities are expensive I move towards holding more cash or bonds. When I feel equities are cheaper, I move more towards equities. This is subjective and I need a few more decades of experience. Interested readers may wish to Google the “Buffett Indicator” and also “Tobin’s Q” to read about some other indicators of whether stocks are expensive. One indicator that I look at is how much profit US corporations are making relative to their value. For example, in 2019, US non-financial corporate business made about $1 trillion dollars of profits after tax. At the end of 2019, those corporations were valued at about $34 trillion dollars. That return is about 3% when I look at profits divided by value. I don’t like 3% per year. I didn’t think that was a lot of compensation considering that the stock market can fall by a lot more than 3% (especially back in mid 2019 when I could get ~2% return on short term bonds).

As a side note, the market has fallen now since the onset of COVID-19. That has reduced the market value somewhat, but corporate profits are going to fall as well. Discussion question: Is the stock market really much better value right now than at the end of 2019?


Simple Ben is 50% stocks, 50% bonds.

A little more complicated is adding in a little gold, (or bitcoin because I’m a little crazy).

A little more complicated again is increasing stocks versus bonds when I feel stocks are cheap and reducing stocks when I feel they are expensive.



Some things I’m working on – Monday April 13th 2020

Howdy folks, about time for me to knock the cobwebs off of this blog. Here are a few things I’ve been working on lately that you may be interested in. Or, you may just want to find out why the featured image is an aardvark.

1: Non-alcoholic Beer – Growing Point 5’s online sales.

Some background: As you may know, Point 5 is a non-alcoholic beer brand that the Sandymount team launched recently (I’m just finishing off my second bottle for tonight right now). While it’s been a hard time with COVID for most businesses, beverage sales is one area that has been holding up strong – with people keeping well hydrated or entertained at home. A lot of big companies are reducing their marketing spending to save on cash, so the price of online advertising has gone down (apparently, I’m just learning about this). This means it’s a good time to get some online advertising going for Point 5 on Facebook and Instagram – so I’ve been learning how that all works.

Where I need help: I’m an engineer and so are the folks at Sandymount, so we’re fairly grand in terms of figuring out production. We’ve also now got a handle on distribution and fulfillment. With a bit of help from some friends and connections, I think we’re starting to figure out the digital marketing as well. The main challenge for me right now is finding someone that will – over time – run the Point 5 business as a GM. It’s a new and growing market – the non-alcoholic beer – so if you’re entrepreneurial (and especially if I already know you!) give me a shout to chat some more.

2. COVID-19 Stuff:

As you might have seen on social media I’ve been in and out (and in again) on pulling together some hand sanitizer production. Just as the project got shut down, Anton Hunt – a friend of mine from days at MIT – pulled it out of the grave by finding a distribution partner. We (a few engineering friends) are just producing the first bottles right now, so stay tuned for some updates (website here). In some more detail, here are a few things I’ve been thinking about on this front:

1. Moisturizer hand sanitizer. If anyone has good ideas for simple recipes that sanitize and also moisturize (I’m getting soft), that would be of interest. It’s easier said than done because when you change the recipe then you have to get approval from the FDA. Still, worth thinking about because the hand sanitizer market has probably increased not just in the short term but also as a medium term trend that will continue.

2. Other stuff to start supplying. One area to think about is simple masks because people are probably going to wear masks a lot more in the near and medium-term future. When I needed one, I looked on Amazon but it’s hard to find a comfortable one with good ratings. My aunt was telling me yesterday that Etsy was a decent source for handmade ones. Here is a simple open source design, albeit one that makes you look like an aardvark (shared with me by Anton – the design, not the aardvark). Let me know if there are other designs to consider (I think the key things to optimise for near term are comfort and high throughput manufacturing).

3. Random – Desalination on a Big Massive Barge:

One of the biggest costs of doing seawater desalination in countries like the US is planning permission. Projects can take ten years to get approval and can take many turns and redesigns along the way – all adding to cost. Why? To desalinate water you need to build large pipes to pull water in from the sea to your system. The system then turns that water into roughly half pure water and roughly half concentrated salt water (at roughly twice the concentration of the sea). That concentrated salt water then needs to be piped back out to sea and dispersed/diluted. All of this pipework, and also care to avoid sucking in fish, or gushing out the salt water on sensitive habitats, takes time and planning. Time is money and this tends to be a huge driver in the final cost of desalinated water.

Anyway, one option here is to build a barge at sea that has a desalination plant on it. You can read more about a concept from Mobile Offshore Desalination here. Good idea, I think! You put the barge a few kilometers out at sea and you pipe the water in – and hopefully save spondulix on planning permission and save a lot of time on getting a new water supply operational.

I’ve heard of a few companies working to put this kind of project together in the past. So, what’s the problem? The problem is not technical, I think – with the right people, you can build the barge with pretty low risk. The problem is that you have to find a buyer – either someone to commit to buy the water, or someone to pay for the whole thing. It’s the perfect kind of thing people/governments would like to have right when they need it, but generally won’t want to or be able to commit to in advance (without some long tendering process).

I’ve only started learning about this, so I might be missing something major. Still, it seems worth it for me to explore a bit more because offshore desalination does seem to have environmental and cost benefits over the usual gig on land.

Areas where you could help:

If you’re in government/organisation and are looking at a long term water supply option bigger than 100,000 m3/day.

If you are a big company with coastal operations that needs to secure a future water supply.

If you bought bitcoin ten years ago and want to turn it into a physical asset that can water the earth.

Let me know and we’ll see where it goes.


Sapiens: My review of a book about humans

The cover of Harari’s book reads “Sapiens: A Brief History of Humankind”. In fact, the book is more a philosophical meditation on the past, present and future of humans than a historical account. The author takes questions including: what is a human?, and, what makes humans happy?, and explores them through the lens of hunter-gatherer, agricultural, empire and modern/scientific societies. If there is an answer, it is perhaps that these questions will remain just as difficult, if not more difficult, for homo sapien societies of the future than those of the present or past.

1. Humankind has made incredible progress, but, progress in the name of what?

Just because we are technologically advanced doesn’t mean we are happier. People today may not be happy living in a hunter gatherer society, but that doesn’t mean that hunter gatherers were less happy or satisfied than we are today. Indeed, objective happiness – measured via surveys – seems to suggest that a state of happiness is driven by genetics and by environment, but a good environment doesn’t seem to require technology. Perhaps, Harari reluctantly suggests, “happiness is synchronising one’s personal delusions of meaning with the prevailing collective delusions”. Yes, sounds cynical, but interesting. Maybe this is why we don’t put the meaning of life into words…

Also interesting is Harari’s perspective on how societal or technical progress has not always been good for well being. Looking at the agricultural revolution where people moved from diverse hunter gatherer tasks to monotonous specialisation as peasants (where conditions were often worse), Harari argues the agriculturalisation of society is perhaps one of history’s greatest frauds. It is a lesson worth keeping in mind as we push for the vague notions of “science”, “technology” and “progress”. Today, our world has more technological capability and we have greater ability, through communications, to be aware of that capability. Today we benchmark ourselves not against our communities but against our world. We have greater visibility into inequality than ever before, and we have tools to create greater inequality than before. As society advances, forgetting that human happiness is relative, will perhaps be to our peril.

2. Are animal rights and human rights two sides of the same coin?

Harari begins the book highlighting how homo sapiens may well have established ourselves by eliminating neanderthals. The very question of who is considered to be human has been and will be tenuous. In the past, homo sapiens have eliminated one another on the basis of much smaller genetic differences (e.g. skin colour, gender, language etc.) than those between neanderthals and homo sapiens. In the future, genetic technology may bring us back to tensions between human species, perhaps homo-sapiens versus homo-futurus (that is made up, I can’t speak Latin, much less predict the future).

3. What is the future of mankind?

Harari finishes the book on a topic that is to the fore as of 2016, with much news around the development of artificial intelligence. Harari, at least in my opinion, breaks down, in a nice way, how humans might now evolve via technology: a) we modify our genes with biology b) we build robots that think for themselves, or c) we combine smart robots/computers with humans to increase our powers.

In some ways, as many (like Elon Musk) have pointed out, we already have AI – it’s in our mobile phones and computers. With our phones (notwithstanding the productivity sink of Candy Crush) we are already superhumans. Elon Musk seems to think that the best outcome is for some kind of option c) where humans are part of the AI rather than replaced by it. Seems like a good option to me (if this is a choice).

4. Some caveats

Covering the history of humankind in one book is definitely an endeavour that requires the sparing of some detail. That being said, I have to say my faith in the historical accuracy was occasionally shaken. One chapter recounts how St. Brigid is the most revered saint in Ireland, when it seems to me that would be St. Patrick. Along similar lines, Harari’s chapter “On the Scent of Money” recounts how money arose as a replacement for barter. I’m not an anthropologist, but I did read David Graeber’s book “Debt: The First 5,000 Years“, and it seems there is a strong case against barter having being the precursor of money [Graeber argues money (currency) arose as a means of kings to collect taxes or pay soldiers and civil servants]. These aren’t major points, but I am left wondering to what extent some other historical claims might need to be double checked.

Sapiens, it’s a good read. On a scale of yes or no, I would give it a yes.



Death by bcc

(Metaphorical) Death by bcc is the ultimate of all e-mail deaths. It ranks at the very top, above the basic e-mail blunder, and above the dreaded “reply-all”.

For those that are too sensible to have fallen foul of these errors, let’s kick off with a baseline example of death by “reply-all”: You get an e-mail, from a friend, to a group of your friends, inviting you to go go-kart racing. You want to go, but you’re worried one of your friends (cc’d on the list) may not be able to join because their head is too big to wear a helmet [Note: I may or may not have been that friend]. Moments later, after you have pressed the send button (and after the 10 seconds to “undo send” have elapsed on gmail) you realise you pressed “reply-all”…..

Death by “bcc” is sometimes just a compounded version of death by “reply-all”. Imagine the same scenario as described above, but where you, initially, were bcc’d while everyone else was cc’d. Now, not only are you replying all, telling them someone has a big head; you are hurling an insult that is coming completely out of nowhere…..

For the record, who ever the person is who used bcc, should be responsible for at least 100% of the blame if the situation escalates into the dreaded “bcc-followed-by-the-dreaded-reply-all”.

I don’t usually have morals of the story, but here is the moral of this story: Don’t use bcc. Ever. Either leave the person off, or else do them a favour and upgrade them (or downgrade them, depending on the case) to a standard, straightforward, tried and tested “carbon copy” or “cc”.

I don’t advocate having the death penalty, but if one were to optimise for sadism, I would recommend death by bcc.


Pure Decent Books About How the World Works – May 2016

1. Debt: The First 5,000 Years, By David Graeber.

Great book because it gets us out of our very modern and restricted understanding of how money and debt work. Perhaps, most interestingly, it debunks the idea that financial crises are a modern phenomenon – rather, they have existed for millennia. The book also suggests that money didn’t come about as a result of barter. The title sounds like an economics or finance book but is really more of a history or anthropology book.

Caveats: Definitely written in a pretty non-sequitur style, with webs of ideas rather than a linear train of thought. Sometimes this is frustrating, but more often entertaining – such as when recounting ancient trade customs involving elaborate sequences of feasting, trading and copulation. On a more serous note, the book does draw attention to the tight relationship between money, debt, enslavement and brutality throughout history, arguing that this is often overlooked in modern financial discourse.

2. Thinking Fast and Slow, By Daniel Kahneman

Clearly I’m one of many recommending this book, which sheds light on how we humans aren’t always logical when it comes to decision making. This book exposes the biases to which we humans are vulnerable and offers us a language to describe these biases. The book is incredibly non-anecdotal, very concise and well written.

Caveats: The first few chapters drag quite a bit, you’ve got to persevere until when the gold comes from chapters four and onwards.

3. Anti-fragility, By Nassim Taleb

Taleb, in this book, provides observations and advice on how we should handle uncertainties in this world, protect ourselves from risks and even leverage uncertainty to our advantage. Anti-fragility is about things that get stronger with uncertainty, volatility or in environments with fluctuating stresses. A good example is the human body, which strengthens when subjected to moderate exertion. Optionality – the possession of multiple choices – is also anti-fragile because it’s good to have options when the future is uncertain. The way I say it, it sounds trivial, but this book is deep. You may not agree with Taleb and you’ll certainly find him to be eccentric (which, in my opinion, adds to the entertainment factor). However, the way Taleb sees the world so differently really opens your eyes and makes you question how much of what we accept in politics, in finance, in business, in education and beyond, is shallow group-think.

Caveat: Taleb’s book is like an incredible piece of abstract art; it is beautiful but there is no continuous narrative and with every answer provided, there are at least two new questions.

P.S. I’d highly recommend this book above Black Swan, another book by Taleb, which emphasises how our world is shaped by rare and unpredictable, but highly impactful events. Anti-fragility has more concepts and teaches much more and is probably a bit better written.

4. What Money Can’t Buy: The Moral Limits of Markets, By Michael Sandel

While the world is seeing more and more markets emerge, not just online, but also in healthcare and education, this book takes a look at what that means in ethical terms. One simple conclusion emerges that the more markets there are, the more power is gained by those who have what it takes to engage in those markets (often money). This book argues that there are often complex and important questions that arise when markets are created and designed, but we often prefer to ignore that we are in fact taken ethics based decisions.

Caveats: Probably would be good to read this in book in tandem with “Justice” by Sandel. Also, heavy enough subject matter, but a very good read because it talks about markets in a way that is not seen in broad discourse.



Part 9: Entrepreneurship = Luck + Talent + Wisdom

Based on what I know today today, I estimate that entrepreneurship is one third talent, one third luck and one third wisdom. Talent is not teachable and, at present, is very hard to identify in first time entrepreneurs. Luck cannot be taught nor bought but plays an important role in success, and, unfortunately, is often misinterpreted as talent or wisdom. Finally, wisdom is teachable but we must be incredibly careful about following only guidelines that are borne out by data and not just a series of anecdotes.clover

Entrepreneurship is one third luck

We all love to believe in superstars, and superstars there are, but there is strong evidence to suggest that luck plays a big role in startup success. The best evidence I see is that entrepreneurs who are successful with their first venture have only a 30% chance of success on their subsequent venture. If luck played a weak role, then there’s no reason why previously successful entrepreneurs should continue to be successful from there-on*.

Entrepreneurship is one third talent

Data show that talent has a big influence on success. I believe this because the success rate of previously successful entrepreneurs is much higher than average. If talent wasn’t a thing then it wouldn’t matter whether your founding team had persistently been successful in the past – this simply isn’t borne out in the data. You might argue that having past-entrepreneurs on your team helps simply because they have experience. However, I don’t think this is the case, because then entrepreneurs who failed in the past would also be useful to have on your team, which isn’t true when you look at the data [in fact, their performance is very close to average].

Entrepreneurship is one third wisdom

There is also one important aspect of entrepreneurship that can be thought, which might be thought of as wisdom, rules of thumb, or, more formally, heuristics.

One way to develop rules of thumb is by carefully thinking through the fundamentals of a successful startup. Peter Thiel’s rules provide one such example. Sustainable profits are achieved by finding a way to achieve a durable competitive advantage, which, in its strongest form, is, by definition, a monopoly. A monopoly can come about as a result of enforceable patents, trade secrets, economies of scale or network effects; by optimizing for such characteristics, it seems one could improve the startups chances of success.

A second way to develop rules of thumb is to look at broad datasets but, unfortunately, there are few. Furthermore, amongst the few studies that exist, few, if any, characteristics of a startup that can be measured early in its life have been proven to lead to greater success in its later stages. Setting team composition aside, the only real correlator with success seems to be revenue, but there aren’t even enough data to prove this for startups at an early stage, and, what’s more, we know that many early stage startups simply don’t have any revenue – meaning we could really do with other new metrics.

Putting it all together

Taking the insights I have gleaned from the research and thinking of others, I have compiled a model to estimate the future success of early stage startups. For sure this model is at best weak and at worst wrong. However, I include it because I think the approach to building the model is a useful tool for others.

The model I propose has three pieces: team, traction and fundamentals – each piece worth one point, giving a maximum score of three. Importantly, the model is quantitative and relies only on information that is measurable:

1st Piece: Team (Maximum points = 1 pt.)

Raw talent – If at least founder has previously founded a startup that was acquired or went public, assign 1/2 point. Failing the above test, if at least one founder has founded a startup with current annual revenue of above $10 million, assign 1/4 a point.

Founding team size – Assign 1/6 of a point for each of the following roles on the starting team – up to a maximum of 1/2 point in total: a CEO type role; a lead technical role; a sales and marketing type role; some other distinct role.

2nd Piece: Traction (Maximum points = 1 pt.)

Sum up the total revenue earned by the startup to date. Include only revenue that has been received. Assign 1/6 pt for $0-100, 1/3 pt for $100-1,000, 1/2 pt for $1,000-10,000, 2/3 pt for $10,000-100,000, 5/6 pt for $100,000-$100,0000, and, 1 pt for above $1,000,000.

3rd Piece: Fundamentals (Maximum points = 1 pt.)

Intellectual property – Companies with chemistry or pharma based patents or trade secrets – ones that are either very hard to copy or else are really obvious if someone copies the patent so you can sue them; 1/3 pt. Companies with patents that are easy to copy but somewhat possible to track if other people copy; 1/6 pt. Other companies fitting neither previous description; 0 pts.

Economy of scale – Software; 1/3 pt. Hardware or infrastructure; 1/6 pt. Companies fitting neither of the two previous descriptions; 0 pts.

Network effects. 1/3 pt for any marketplace or communications company and 0 pts for all other companies.

Parting words – Can entrepreneurship be taught?

Before concluding, I want to point out that I’ve taken for granted the assumption that every entrepreneur wishes to optimize for success. Indeed, as an entrepreneur, you really only get to make one bet; you’re betting on the success of your own company, not a portfolio of companies. In this regard it makes sense to optimise for success. However, I really think that, while we should largely optimize for success, there must be limits. If we were all to follow the framework I have outlined above everyone would end up building a communications or marketplace based software startup. Indeed, half joking, that’s more or less what has happened. Importantly, though, not all problems we face in this world are addressed via marketplace based software startups and that, I think, is where the assumption that we should solely optimize for success might just fall down. Mission plays an important role.

Now, to conclude on whether entrepreneurship can be taught. Luck and talent each play a huge role in startup success and, by definition almost, neither is easy to teach. However, we can teach aspiring entrepreneurs to be smarter about how they build their founding team. By following the right rules of thumb – for example, focusing on building a startup with strong intellectual property, economies of scale and network effects – entrepreneurs can likely improve their chances of success. Unfortunately, as things stand today, there are relatively few good rules for entrepreneurs to follow that have been rigorously established. In my view, this stems from a lack of publicly available datasets on startups, insufficient focus on relating measurable characteristics of early stage companies (such as revenue or number of patent applications) to success, and an over-emphasis on hard to measure characteristics (like market size). With a focus on relating measurable startup characteristics to their future success, I believe there are many more rules of thumb that can be established to the benefit of future entrepreneurial generations.

*This isn’t strictly true. For example, it is possible that entrepreneurs who are successful first time round aren’t necessarily successful the second time because they fail to follow certain rules of thumb – e.g. they didn’t opt for a startup that had economies of scale and strong intellectual property – indeed there are many alternate ways of thinking about this. However, to me, it seems that a big reason is simply that luck plays a strong role, so if you’re successful first time around, there’s no guarantee of being successful the next time.


Can you teach entrepreneurship? Part 8: When to sell?

  1. Sell when your company has achieved its mission.
  2. Sell when you have run out of ideas.
  3. Sell when the buyer offers a price that you feel reflects the company’s full potential.

Theories on when to sell vary from the philosophical to the economic. These three mental models, gleaned from the work or writings of Elon Musk, Peter Thiel and Ben Horowitz, seem to cover the full spectrum. In this blog I will provide a summary of each.2000px-For_Sale_by_Owner_Sign.svg

Elon Musk’s mission based sale

Perhaps my impression of Musk is a function of the journalistic narrative of a recent biography, which casts Musk as a man on a mission to bring electric cars, renewable energy and the colonisation of mars to reality. However, even beyond this biography, I do think there is strong evidence of Musk being driven by mission. Firstly, Musk holds nearly all of his wealth in the handful of companies that he founded. From an economic perspective, this lack of diversification is highly sub-optimal – his eggs are in just a few baskets. Secondly, when you read Musk’s own words, such as his 2013 e-mail to SpaceX employees, it becomes clear that he sees companies as a vehicle for achieving a mission, not the other way around. His mission for SpaceX is to bring humans to mars. This mission is more important than near term financial success and he refuses to allow SpaceX to go public because he believes that public markets won’t respect and prioritise this mission as he wants.

Clearly I don’t know what Musk is thinking, but, if I were to guess, I would say that he would advise to sell your company only if you have achieved your mission, or, if you have lost control to the extent that your mission can no longer be fulfilled as you want it to be.

Peter Thiel’s pragmatic sale

As written in Zero to One, when a company is sold, Thiel feels the price is always to high. Conversely, when an acquisition offer is turned down, Thiel feels the price is too low. Thiel’s logic is that companies are sold when their leaders run out of ideas. When this happens, the company is therefore overpriced. When an acquisition offer is turned down, then the company’s leaders still have great ideas they want to implement and, therefore, the price offered must be too low.

I haven’t seen the data on this but, no doubt, Thiel very well may have. Regardless, I think the point is philosophically a good one. The corollary, I feel, for when one should sell is therefore simple. Sell when you have run out of ideas.

Ben Horowitz’s pragmatic sale

In The Hard Thing about Hard Things, Horowitz provides a mental model for determining when to sell that is based on economic principles.  As I understand, Horowitz feels that, as an insider, you are often the only one to understand the full economic potential of your business. For example, the market might currently recognise your potential profitability from current sales – say of product A – but may have little insight on the potential for profit from a future product – say product B. For Horowitz, the right time to sell is therefore when the market recognises the potential for profit from all of your products. If you don’t get until this point, you will be selling the company too cheaply.

In conclusion, as I have presented the mental models above, I make it seem like you have the choice as to when you can sell. Indeed, as the three entrepreneurs above also attest, the decision of when to sell is often strongly influenced or even fully controlled by others, whether investors, colleagues or someone else. Nonetheless, for everyone involved, I think it is well worth reflecting on the rational for why and when the time would be right to sell.


Can you teach entrepreneurship? Part 7: Be a master of big sales

Success in a startup requires skill in big sales. It is not widely known or accepted but big sales (e.g. which include hiring employees, getting investors and selling expensive products), require a set of skills that are very distinct from selling low price products (say, for less than $100). Sales is often something that we think of as an art and not easy to teach. However, I think that Neil Rackham, though his very long and careful studies of large sales in SPIN Selling, is changing this and his book is a huge resource for startup founders who want to learn sales techniques that are proven to be effective. In this blog, I will describe some of my key learnings.sale

Even if you sell a low priced product, I think that most of the selling that a startup founder does involves big sales. You know something is a big sale if the buyer has to do a lot of thinking before deciding to buy. Selling a newspaper is not a big sale. Selling video game often involves the buyer doing some research, but I still wouldn’t call that a big sale. Big sales require lots of thought and analysis on the buyer’s side: selling an enterprise software solution for $100k; or, convincing early employees to join or convincing investors to invest. These are all examples of big sales, and they require some specific techniques that are not widely known or even accepted.


The best way to ask is not to ask, the best way to sell is not to sell

If you google “closing techniques” you can find long lists of (sometimes hilarious) techniques to try and force a customer to buy. One example, “the assumed close”, might go something like this “So I’ll have the delivery man get the washing machine to you as soon as possible. When would suit for delivery, Wednesday or Thursday?” Such, techniques – which essentially involve putting the customer on the spot and forcing them into a decision – may have some success for small sales (e.g. a pencil, newspaper or the like) according to Rackham. However, for big sales, the technique of putting the customer on the spot very negatively affects the likelihood of closing the sale. Rather, to close a sale, it is much more effective to ask the potential customer about what the payoff would be for them if they adopt your solution. If you do this well, it is the potential customer who should ask to buy rather than you who should ask to sell.

Conflict avoidance

Apparently, traditional sales training often includes guidance on how to handle concerns that are raised by potential customers. However, in Rackham’s studies, he finds that these handling techniques don’t improve sales performance. Rather, Rackham finds that it is much more effective to focus on minimising the likelihood of the customer raising concerns. Rackham contends that, in big sales, rather than suggest what features might be of help to the customer, better instead to ask the customer about the implications of their current problems and then about the payoff that your solution could offer. In this way, the customer isn’t focusing solely on the drawbacks of your solution, which certainly exist. Rather, the customer is taking a holistic view of the net benefit for them.

The SPIN selling technique – the best way to convince someone is to help them convince themselves

Based on his findings from large numbers of studies, Rackham has developed heuristics for selling, one of which includes the SPIN – situation, problem, implications and need payoff – steps to guide a salesperson’s questions during selling. First you establish the customer’s position in their company, their role and their background – this is the situation part. Then you establish the problem faced by the customer, followed by a deep dive into the implications of that problem for the customer (lost productivity, wasted raw materials, product defects, employee safety, etc.) . Finally, having established the implications, you help your customer to establish the payoff for them if they adopt your solution (termed the need payoff). One of the main findings of Rackham’s work is that more implication and need payoff questions have a very strong positive effect on the likelihood of reaching a sale. Establishing the situation and problem are important for the salesperson (and to inform a startup’s strategy) but ultimately, it is in the implications and need payoff that the value for the customer becomes most clear. Here’s the rationale – in a big sale implications and need payoff are important because your solution is expensive and prospective clients need to convince themselves that the value for them is greater than the price as which the solution is being offered. The higher the price, the more time will need to be spent for them to convince themselves of the value.

After reading SPIN Selling, I feel that sales techniques for startups absolutely must be taught and big sales should be an emphasis. I think that the status quo is for entrepreneurs to hope that sales can be learned by experience. At worst, I think that the learning by doing approach – without the right prior teaching – is ineffective. More likely, I think it is damaging. We now know that traditional training techniques have been shown to be counter-productive in large corporation. That, for me, is a sobering thought and should set startup alarm bells ringing. Startups are uncertain environments. We desperately need rigorous studies that are backed up by large amounts of data. We need to teach the heuristics that arise from these studies and move away from a “learning on the job” or anecdotal approach to sales that I fear can be counterproductive.


Can you teach entrepreneurship – Part 6: Mastering Market Size

A startup’s market size must be big enough to justify any upfront investment. Applying this logic suggests $5 million is a very minimum empirical market size for software startups. The same logic suggests very minimum market sizes for hardware should be larger, and, for pharma, much larger. Rigorously validating market size requires time and money and this imposes an upper limit on market size. Perhaps the only justifiable support for market size is sales (or pre-sales) because of the uncertainty of sales estimations for very early stage startup products. If sales (or pre-sales) are the only legitimate way of usefully calculating market size, then market sizes quoted by startups should be very small indeed, perhaps $100k to $1 million in size – much smaller than the $100 million or billion dollar market sizes often touted today.

To calculate market size, start with a figure of $1 billion per year. Then, project a price (let’s say $1, for this fictitious app). Then, divide the market size by the projected price to project the number of customers (1 billion, in this case). Then, if you feel the number of customers is too high, reduce the projected number of customers (say, to 100 million) and compensate by increasing the projected price (say, to $10). Iterate until you are happy. You now have a projected market size that you can back up with a projected price and a projected number of customers.Moore_Street_market,_Dublin

For the avoidance of any confusion, the above paragraph is a joke. However, as with many jokes, there is perhaps an element of truth behind it. Market size is a much talked about concept in entrepreneurship but also one that I feel is incredibly murky.

Anecdotally, there are many market size heuristics; for example, some investors want to see a billion dollar market and some some want $300 million in year 5 revenue. Some investors want the opposite; Peter Thiel, in Zero to One, for example, is scared by founders who project large markets and prefers to see companies start by building small monopolies. Bill Aulet, in Disciplined Entrepreneurship,  also favours a small beach-head market and suggests initially aiming for markets that are between $5 million to $100 million in size (based on the startup’s maximum potential revenue). So, the question is, what is the ideal sized market? and why?


On a simple level, you could say that you define a market size to see if you could make a boat-load of money – the more the better. That could be true, but I think there’s a more refined reason. Ultimately, if you’re a founder, you’re probably passionate about something and you want to bring that something to life. For it to come to life, one requirement is that the business is a sustainable one – market size, in my opinion is part of that consideration.


To get to a sustainable business there are some barriers to entry that need to be surpassed. For software, it might be a few years of salary for founders and employees, e.g. a million dollars. For hardware, there might be additional R&D costs, e.g. a few million more dollars. For pharma, there might be also be the cost of clinical trials, e.g. hundreds of millions of dollars. If your business is going to be successful, you probably want your eventual annual revenue (one way to define market size) to be significantly larger than this barrier to entry. This sets a minimum market size.

One way to think about this is to consider the final worth of your company as being the maximum achievable revenue (which we’ll consider to be the market size) multiplied by the net profit margin multiplied by a price to earnings ratio:


Consider that the final company worth will need to be large enough to justify the capital required multiplied by the return on investment that investors require:



This means that, the minimum market size you require for a sustainable business is:


To put this in context, we can consider a price-to-earnings-ratio of 6X (a possible value for mature private companies), a return-on-capital for investors of 20X and a profit margin of 50%. This would give the following minimum annual earnings:

  • Software example: Minimum Capital Required = $1 million; Minimum Market Size = $6.5 million
  • Hardware example: Minimum Capital Required = $3 million; Minimum Market Size = $20 million
  • Pharma example: Minimum Capital Required = $200 million; Minimum Market Size = $1.3 billion

Obviously, the minimum capital required depends on the startup in question, as well as other assumed values above. However, I think the above calculation provides some good heuristics:

  1. Markets smaller than roughly $5 million (measured in achievable revenue) probably don’t make much sense to pursue.
  2. The more capital you think you need to bring your business to life, the larger the market size you will require.


In my opinion, the goal of calculating market size is to demonstrate that sufficient revenue can be generated to justify the upfront investment that is required to bring the startup to life. There are two ways you could do this: extrapolation or validation.

Extrapolation – here, you take whatever insights you have from a first sale (hopefully you have one) and try to project future revenues (preferably using outside data sets of relevance). It isn’t often stated explicitly but this kind of projection is very difficult because there are many reasons why current and future customers might be different. For example, the problem you solve now may not be a priority for new customers, new customers might already have an alternative solution, or, it may be more difficult to reach new customers. As an early stage startup I think the ability to extrapolate to new customers is very limited. Before a product is fully defined, the business model repeatedly tested and the sales cycle fully validated, I would ask whether projecting revenues is a waste of time.

Validation – here, you sell (or pre-sell) your product to a customer (or group of customers) and get an idea of your revenue and perhaps also the rough cost of acquiring the customer (i.e. marketing + sales costs). Getting validation like this isn’t easy but it does allow you to establish a minimum market size from the revenue. From this starting point, you can maybe also include in the market size any customers are extremely similar (if not identical). However, from here, you hit a trade-off. By doing more validation you can increase your market size. However, validation work costs time and/or money and a point is reached where you stop, look at the current market size, and see if justifies an investment that will accelerate market validation.

Ultimately, market size validation is, by nature, an inefficient process and its expense imposes a practical limit on market size. First of all, founders can’t outsource or easily automate the validation process because it is crucial to build relationships with customers and better understand their needs for themselves. This means that the cost of acquiring customers is the founder’s opportunity cost (i.e. the salary foregone in choosing not to work for another company). This could be well above $100k per founder per year. The second reason why validation is expensive is because it is essentially a low efficiency direct sales process. Mature startups have a more clearly defined product, a tested sales funnel and low customer acquisition costs. Early stage startups are still figuring things out and this makes sales/validation less efficient and more expensive. For example, while mature startups might pay $1 to acquire every $5 of future customer value, during validation, startups might effectively spend $1 on direct sales to acquire only $1, or even less, in future customer value. With this, admittedly blackbox, assumption for validation, a founder whose opportunity cost is $100k per year would validate a market size of $100k, which doesn’t even meet our minimum market size requirement of $5 million.

My model of market validation is almost surely wrong. However, I think it does illustrate how expensive it is for founders to properly validate market size. Clearly the idea that it is possible to validate, or even estimate with a useful degree of confidence, a market that is $1 billion in size is absolutely ludicrous. Validating even a few million dollars of revenue, I think, is something that requires extreme founder sales talent and skill.


I think that market size is a concept that needs serious reconsideration. I worry that founders and entrepreneurs are wasting significant time in performing and analysing market size estimations. I wonder whether, and hypothesise that, current methods are more likely to lead everyone astray than in the right direction.

Clearly it is important to know the market size because the business can’t be sustainable if the market isn’t bigger than the investment that is required upfront. I think it is important to question whether, for startups, it is possible to make projections of market size that include customers to whom the product has not yet been sold. Sales (or pre-sales), I think, are a much more sure way to assess market size. Even if the market size is then only hundreds of thousands or millions, I wonder whether having that small quantity of certain information is better than having a market size of billions based on useless information.





Part 5: The Dream Team

Raw entrepreneurial talent and a diverse founding team correlate strongly with improved startup success rates. In this blog, I’ll provide evidence for this claim and propose a way to measure the quality of a startup founding team by looking at prior startup successes and also the number of roles within the founding team.dream team

Data show that raw entrepreneurial talent has a big effect on startup success. For repeat entrepreneurs, one good way to identify talent is to look at prior entrepreneurial success – those who have already succeeded are likely to be among the most talented. For first time entrepreneurs, identifying raw talent is harder. For example, academic degrees don’t seem to correlate strongly with success or raw talent. Unfortunately, I don’t see a good indicator of raw talent for first time entrepreneurs and this question needs more research.

Data show that having multiple founders significantly increases the chance of success. One reason is that a larger team also makes it more likely that the team is diverse, and data shows that a diverse team (i.e. one with multiple roles such as technical, sales and finance) is also more likely to be one with greater success. There are probably many other underlying reasons for this such as greater moral support or a larger network with larger teams, also I’m not familiar with data showing the strength of these.


With the above in mind, here’s one suggested approach to evaluating a startup team:

1. Raw talent

If at least founder has previously founded a startup that was acquired or went public, assign 1 point.

Failing the above test, if at least one founder has founded a startup with current annual revenue of between $10 million and $100 million, assign 1/2 a point.

Otherwise, assign 0 points.

2. Founding team size

Assign 1/3 of a point for each of the following roles on the starting team – up to a maximum of 1 point in total: a CEO type role; a lead technical role; a sales and marketing type role; some other distinct role.

3. Combining these factors

Take the (unweighted) average of the scores for raw talent and founding team size to come up with a final team score (maximum final score of 1/1).


Many startup teams won’t get any points for raw talent because the entrepreneurs haven’t had a prior startup success as I’ve defined it – e.g. most MIT spin-outs will have score 0.5/1 at most, earning those points from “founding team size” alone. This is partly because 1) having founders with prior success (acquisition or IPO) greatly increases the odds of success for their future startups and the model needs to capture this strong benefit, and, 2) I don’t have a good way at present to measure raw talent for first time founders – this is the biggest shortcoming of the model. When it comes to team size, this model incentivises having multiple founders and also having those founders cover different roles. These aspects of founding teams are correlated with improved startup success.

In summary, raw talent (measured via prior startup success) and larger founding teams correlate with startup success. In building a startup team or evaluating an investment, I suggest looking for these key attributes.




The importance of raw talent in startups

Data in a 2012 paper by Eesley and Roberts’ indicates a strong correlation between the raw talent of the founding team and startup success (0.59 correlation between talent and the logarithm of revenues). The way they identify talent is by looking at the difference in revenues achieved by founders who have founded the same number of startups in the past. Unfortunately, this method of identifying talent isn’t something you can use for first time founders. I had thought that perhaps academic degrees would be correlated with success (and thus be an indicator of talent) but data from Eesley’s 2013 paper shows that the relationship isn’t very strong (correlation of less than 0.05 for master’s and doctoral degrees with a successful exit for the startup). Right now I don’t see a proven way to identify talent up front and I think this is an important question.

The benefits of having founders with prior startup success

A HBS study from 2008 (discussed in Part 4 of this series) shows that founders who have had previous startup success have almost double the chance of success on subsequent ventures compared to success rates for average first time founders. In short, it seems that prior success is one way to pick out the talented entrepreneurs from the crowd.

The benefits of multiple founders

Anecdotally, Y Combinator (the California based startup incubator) recommend more than a single founder as they feel a startup is too much work for one person. This seems to be supported by data. Data from a 2013 study from MIT shows that there is a a positive correlation (0.27) between founding team size and the startup having a favourable exit (IPO or acquisition). The same data shows that there is also a negative correlation (-0.19) between a startup having only one founder and reaching a favourable exit. [This latter result is not so surprising given, by definition, single founder teams are team of small founding team size.] A study from 1990, led by a Stanford author, also supports the finding of larger founding teams being correlated with success. It focuses on semi-conductor startups from 1978 to 1988 and reports a significant correlation (0.31) between fourth year sales and the size of the founding team.

Evidence for the benefits of having a diverse founding team

Data from the MIT study mentioned above indicates that having a diverse team is related to higher chances of a favourable exit (0.16 correlation). They measure diversity on a scale of 1 to 4, by counting which of the following roles are included in the founding team: technical (chief technology officer or scientist); sales and marketing; finance; and, other. [As I understand, it doesn’t matter whether, for example, the founding team has a member who has finance experience, it just matters whether one founder had a title that related to finance in order to be counted.] Data from the 1990 study at Stanford further suggests benefits of a diverse team and shows a positive correlation (0.24) between having a team with differing levels of semi conductor industry experience and the year 4 sales of the startup. On top of this, data from the study reveal a positive correlation (0.29) between having founders who worked together in the past and year 4 sales.



Part 4: Prior startup experience may not help much

Over the last while, I’ve frequently heard that serial entrepreneurs have a higher probability of success. Mistakenly, though, I’ve taken this statement to mean that entrepreneurs improve by learning from past experiences. Serial entrepreneurs do have a higher rate of success, but their success may have more to do with them having the raw talent than learning while on the job. This is disappointing because, in a way, it could indicate that entrepreneurship is hard to learn. However, if startups are highly uncertain environments then it makes sense to me that learning in such an environment should be difficult (see Part 0).

Elon Musk, a serial entrepreneur known for X.com, Tesla, Solar City and SpaceX. Credit: Dublin, Wednesday 31th October 2013: Pictured at the The Web Summit 2013, RDS. Photo by Dan Taylor/Heisenberg Media

Truth be told, there are a limited number of studies on the success rate of serial entrepreneurs. The most relevant in my opinion, in the startup context, is a 2008 study from the Harvard Business School. That study considers thousands of entrepreneurs who received at least one round of venture capital funding. Defining success as an initial public offering (IPO), the study examines the success rates of one-off and serial entrepreneurs who raised venture capital.

A sub-set of the results may be summarised as follows:

  • First time entrepreneurs have a success rate of 20.9%.
  • Serial entrepreneurs who failed in their prior venture had a 22.1% rate of success.
  • Serial entrepreneurs who had a track record of prior success had a 30.6% rate of success in their next venture.

The first thing to say here is that, for sure, serial entrepreneurs have higher success rates (roughly 50% higher). The second thing to say is that it’s very hard to know if these higher success rates are because serial entrepreneurs are naturally talented or because they have learned something from prior success. If talent didn’t matter, and only learning mattered, then the success rate for serial entrepreneurs should be higher and it should be the same regardless of whether their previous venture failed or succeeded. Clearly, this is not the case, so we can say that raw talent, for sure, has an important role. By contrast, if only talent had a role and learning were unimportant, then the success rate should be higher for entrepreneurs with prior success then those with prior failure. This fits the data. Unfortunately, the case where both learning and talent are important could also lead to a scenario where the success rate is higher for entrepreneurs with prior success than those with prior failure. So, it therefore seems that talent is definitely important and learning is possibly important.

If I had to draw a conclusion, however, I would lean towards thinking that talent is dominant over learning in determining success. I say this because I would hypothesise that those entrepreneurs who fail and then go back out again (which are few and far between when it comes to raising venture capital again after a first failure) are perhaps amongst the most talented of the entrepreneurs in the initial batch of first time entrepreneurs. Therefore, if their success rate second time round (22.1%) is only slightly above that of all first time entrpreneurs (20.9%), I would question whether they really are learning much from prior experience.*


The question I am posing here is whether serial entrepreneurs have higher success rates because they have learned from prior experience or because they are fundamentally talented. My fear is that educators, mentors and advisors might wish to believe the former, while the latter may be more true. If talent is more important than learning in determining success I think we should consider the implications:

  • For entrepreneurs, I think this would mean reduced focus on past experience and more focus on heuristics (rules of thumb).
  • For investors, this means there is potential edge for those who can recognise talent at an early age, before there has been a public signal of a prior success.

All in all, I hope this gets you to think again about whether startups are a learning-by-doing experience. At least when it comes to venture backed startups, I think there is no evidence that entrepreneurs learn by doing but at the same time, there is no clear evidence against.

*On top of the issues I’ve raised here, the authors consider, in detail, a number of other issues with interpreting the raw data. To some extent, success breeds success in that top VC firms are more likely to back serial entrepreneurs and serial entrepreneurs may be sought out by better employees. These effects can serve to amplify the success rate of entrepreneurs in subsequent ventures and result in an overstatement of the extent to which entrepreneurs are learning from past ventures.


Can you teach entrepreneurship? Part 3: How to make better decisions

Have you ever tried to make a decision by writing down, weighting and then rating a few different factors you felt were important? Mquestionaybe it was while picking a job, maybe it was picking a university course or maybe it was deciding on an employee to hire. If so, I think there’s a better way to make decisions that works particularly well when there are many uncertain and qualitative factors to consider – which is often the case for a startup.

The approach I’m going to give is not new – it goes back to a paper in 1971 by Dawes. Dawes took a look at graduate admissions and investigated whether the admissions committee was any good at predicting how students would be ranked, by professors, at the end of their graduate studies. The answer is that the admissions committee was very poor at this and the correlation between their assessment and the final ratings of students was just 0.1 . So Dawes built a simple predictive model that weighted, equally, incoming students’ GRE scores, grade point averages and the quality of their undergraduate institutions. This simple model, it turns out, had a correlation coefficient with the final ratings of students (after graduation) of 0.51, much higher than the admissions committee’s assessment! In other words, a simple model was much more accurate at predicting success than the admitting committee*.

To apply this model (called an “improper linear model”) I suggest the following steps:

  1. Decide on what the important factors are in making your decision and include only the most important ones (you should probably aim for 3-5 factors, no more).
  2. Weight all of these factors equally.
  3. For each factor, decide on what a score of 0/1 should look like and decide on what a score of 1/1 should look like. Thinking of what an intermediate score (0.5/1) should look like can be helpful too. Ideally these scores are derived from data (like GRE data) but this isn’t always possible.
  4. Rate each of your options on a scale of zero to one on each of the factors.
  5. Add together and rank the scores for each option to arrive at a decision.

To see an application of this approach, you can take a look at the model I developed to predict the success of startups in Part 2 of this series (where I use pretty qualitative data), or, even better, take a look at the work by Dawes (which uses qualitative data for factors).

The reason this approach works well, compared to the usual weighted approach, is that it turns out humans are inconsistent in how we assign weightings. In uncertain environments, the risk of getting the ratings wrong very often outweighs the benefits of trying to optimise the weightings. With equal weights assigned to each factor, you are essentially minimising the risk of getting the weightings wrong; for uncertain environments, as pointed out by Dawes 1979 paper, this is a clever thing to do. The second thing this approach tries to do (in point 3 above) is to shift the focus onto factors that are more measurable. The logic here is that often, in uncertain environments, it’s better to use a measurable parameter that isn’t quite right (for example, using the weight of a bull when trying to order to rank bulls by age) rather than to revert to something less quantitative (for example, how old his eyes look). In other words, the risks of making a bad qualitative judgement are very often greater than the loss of accuracy in using a more quantitative factor, even if that quantitative factor seems somewhat tangential.

I think startups are environments with lots of incomplete and uncertain information. For me, this makes thinking about how to make decisions well worth the while. Improper linear models are one tool I think is worth considering.

*It turns out that humans are notoriously inaccurate at gauging quality in interviews. Nobel laureate Daniel Kahneman had the same problem when he worked to evaluate interviewing techniques in the Israeli army. He also found that a simple model was better at predicting success.



Part 2: Check-boxes for startup success (inspired by Peter Thiel)

In this part, I’m going to combine some rules of thumb to predict the success of a startup. The model will be based on three of the Checkbox_1.svgfour characteristics that Peter Thiel feels are important for a startup to achieve a monopoly (take a look at chapter 5). The model may give some insights into why it might be that venture capital has tended to be focused on internet, biotech and communications as opposed to cleantech, industrial or consumer goods.

My idea of a startup* is to sink in some money early on in order to reap a massive reward later – aka massive profits. In a perfectly competitive market there would be no massive profits, so, as a startup you have to find a way to be in an uncompetitive market – aka a sustainable monopoly. Thiel has come up with what, at least to my naive self, seem to be four good fundamental characteristics that lend themselves to ensuring a position of monopoly. Therefore, I’m going to take three of these four characteristics, define them on a scale of zero to one (no pun intended) and then combine them into a single predictor of startup success (scored out of three).

1. Intellectual Property

The idea here is that if you have a patent or a trade-secret, you, by definition, have something that no one else has and that you can charge a premium for. At the high end of the scale I would say are companies with chemistry or pharma based patents or trade secrets – ones that are either very hard to copy or else are really obvious if someone copies the patent so you can sue them; such startups I would score 1/1. In the middle I would put patents that are easy to copy but somewhat possible to track if other people copy (0.5/1). At the bottom (0/0) I would put software based companies whose functionality could be built from scratch if you had enough dollars.

2. Economies of Scale

If the marginal cost of what you make falls as you make more then you are in a position to undercut competitors who are smaller in size. This can prevent competition. At the top end of scale (1/1) I would put software, which has a marginal cost of about zero. In the middle (0.5/1) I would put hardware or infrastructure, which doesn’t have the economies of software but still generally falls in cost with volume. Then, at the bottom (0/0), I’d put a service company, whose costs basically scale linearly with volume. [I’d also consider information databases to be part of the economy of scale of software, although I could understand arguments that they should be considered intellectual property].

3. Network Effects

Facebook and Uber and eBay are all really useful because there are lots of users. The more users the more useful the startup is (I’ll talk about Metcalfe’s law in a later blog). Once lots of people sign up it doesn’t make much sense to switch to a competitor and this has the effect of warding off the competition. The scoring on network effects is fairly simple; 1/1 for any marketplace or communications company and 0/0 for all other companies.

There is a fourth characteristic that Thiel feels is important for monopoly – brand. I think (again, naively) that what Thiel says about brand enabling profits is good stuff (e.g. Coca Cola or Burger King). However, I won’t include brand here in my rules of thumb because I don’t think you can measure whether a startup will have a strong brand or not when it is at an early stage.


Now, let’s look at what this model says about different types of startup:

  • Internet/communications: I’m lumping internet and communications into one because they all score 2/3 (or debatably 2.5/3 if you include strong algorithms as intellectual property).
  • Biotech/Pharma: Most pharma startups will score 1.5/3 on this test. They have very strong intellectual property and some economy of scale but no real network effects.
  • Consumer goods: Here I’d say 0.5/1 on IP (not as good as pharma), 0.5/1 on economy of scale and 0/0 on network effects, for a total of 1/3.
  • Industrial goods: Here I’d say 0.5/1 on IP (perhaps 1/1 for some proprietary chemistries/materials), 0.5/1 on economy of scale and 0/0 on network effects, for a total of 1/3.
  • Cleantech: Here it’d be 0.5/1 on IP, 0.5/1 on economy of scale and 0/0 on network effects for a total of 1/3.

The message here is pretty simple; internet, communications and biotech tick more of the boxes for achieving a monopoly position, and hence, I hypothesise, for success as a startup. As a consumer goods, industrial goods or cleantech company, there just aren’t the same number of avenues to monopoly.

Perhaps this model helps to explain why VCs have historically focused their dollars on internet, communications and pharma; or, perhaps you think this model is just Peter Thiel (with or without my further damage) looking at the data on startups and then backing out these characteristics. Either way, there are massive differences in the successes of different types of startups and I absolutely think it’s worth the effort trying to better understand why that is the case.

*to be clear, when I say my idea of a startup, what I really mean here is  my idea of a startup as a (capitalist) tool to fulfill a mission.


In Part 1, I described how startups involve lots of diverse and incomplete information. For this reason, I argued that it’s important in startups to avoid opinions and gut feelings and focus on following simple rules of thumb (like I’ve tried to give in this part). I gave the example of  how Ashenfelter, a wine enthusiast and statistician, by understanding the simple relationship between weather and wine prices, was better able to predict the future price of fine Bordeaux wines than wine tasting experts. Here I’ve included a key table from his 2008 paper.

Ashenfelter Wine
From Ashenfelter’s 2008 paper. You can see here how the prices at which wines are traded very often converge towards the price Ashenfelter predicts using simple rules of thumb for age and weather during the year of harvest.


Part 1: Entrepreneurship is like evaluating a fancy bottle of wine

Based on taste tests, it turns out that experts are very bad at predicting the future price of fine Bordeaux wines – so realised Orley Ashenfelter, a Priceton economist. As told by Daniel Kahneman in Thinking Fast and Slow, when Ashenfelter built simple rules of thumb to predict prices he was able to achieve accuracy far beyond that of the experts – his rules of thumb being based just season average temperature, rainfall during harvest and rainfall during winter.

So, why were experts so bad at predicting each wine’s future quality (effectively) price? – the answer, at least accordwineing to Ashenfelter and Kahneman, is that wine tasting is an example of a “low validity environment” – an environment where there is a lack of frequent high quality feedback. When there is a lack of quality feedback, and an abundance of complex information to consider (lots of aromas/notes/tastes), humans tend to be very inconsistent in making judgements because every time around they tend to focus on different information. Such behaviour, in my opinion, is also often the case in entrepreneurship.

In the assessment of future wine prices from tasting, there is both infrequent and low quality feedback. Infrequent because you only find out if the wine is good years later and low quality because even if you knew which wines would eventually be good, the abundance of aromas experienced during tasting makes it hard to pick out patterns. I think something similar is true with startups. Feedback is infrequent because it takes years to know if you are successful, and, feedback is also very low in quality because there are too many input parameters to know which ones were truly decisive (founder quality, technology, market characteristics etc.).


I think that recognising how startups are a low validity environment is very important. Knowing that feedback is infrequent and of low quality implies two things, I think:

  1. When making individual decisions, we should be very careful not to give much weight to what happened in previous startup decisions.
  2. When receiving advice based on someone else’s startup experience, we should be very careful to to give much weight to that advice.

To summarise, the problem here is that, in startups, feedback is of infrequent and of low quality and that means we should be careful when making decisions. One suggested solution is therefore to give more frequent, higher quality feedback but I think that this approach is misguided. The fact that feedback is infrequent and low quality is not because we are bad at teaching, it is because startups are, by nature, low validity environments – just like wine price prediction through taste. The lesson, therefore, is not to give more feedback* but rather to get entrepreneurs to focus on simple rules of thumb that have been borne out by large bodies of evidence. The development of such rules of thumb by entrepreneurs like Peter Thiel and researchers like Richard Dawes shall be the subject of these next blogs.

Many thanks to my brother Fergal for interesting ideas for and discussions around this blog, in particular the idea to use wine as an example.

*in fact, I think that can be very dangerous [see Kahneman on wicked feedback]


Can you teach entrepreneurship, part 0: The importance of seeing things much more than once

Giving advice on startups is really hard and I think this difficulty is frequently overlooked. It’s hard to advise on startups because, in the startup environment, you rarely see the same thing much more than once. To give good advice you need accurately recognise patterns, and, to accurately recognise patterns I think you need to have seen things dozens of times. In startups, where you are aiming to do something differently, it takes a very long time to see things twice – the startup environment is one where patterns take a very long time to repeat.pattern blog 0

If advice is so hard to give, the obvious question to ask is whether entrepreneurship can be taught, and, if so, how best this could be done? For me, the first step is to focus on studying where things happen much more than once.

In this coming blog series, I’ll try to focus on startup learnings that are backed by large numbers of examples. Preferably I’ll focus on studies where large amounts of information have been carefully analysed (like Neil Rackham’s study of sales processes) and, secondly, I’ll focus on experienced entrepreneurs/investors (e.g. Thiel, Musk, Horowitz) who have a comparative advantage given the vast number of startups that they have overseen.

As you’ll see, I do think there are significant rules of thumb that can be taught to startups, but I also feel that, without further data, there are many other aspects of startups where there really is little that can be said with certainty.




Hashtag One Euro Carbon

Theuroe beauty of a referendum is that there are often only two answers, “yes” and “no”. What we have at the Paris Climate Change convention is the exact opposite of a referendum: there are far too many questions, far too many proposals and far too many possible answers. Rather than debate who should be allowed to emit what and why, I would like to propose a very simple but small step that I think takes us in the right direction. It’s called #OneEuroCarbon and here is what each state would agree to.

Number 1: States collect a tax of €1 per tonne of carbon on gasoline, diesel, coal, gas and kerosene sold in the state.

Number 2: States return the money they collect as an income tax credit to the 1% of people in their state with the lowest income.

Number 3: States give a soft commitment to reviewing this system and upping these numbers in a few years (e.g. to a €5 per tonne tax to be provided as an income tax credit to the 5% of people with the lowest income).

The first part is a tiny but important step in the direction of taxing carbon. Taxing carbon is good because, like tax on cigarettes it’s a simple way to encourage us to reduce doing what’s unhealthy.

The second part is good for a few reasons: 1) it reduces inequality, which is clearly a big issue we can rally around 2) it replaces income tax, which is great because income tax serves to reduce employment.

The third part is good because it gives us something realistic to aim for.

Look, there are far too many interests involved to do anything at the conference that will have a real economic effect – like a carbon tax of €50 per tonne might. Better to focus on taking an administrative step in the right direction. A €1 per tonne tax won’t have a noticeable on anyone, but this first administrative step, especially if widely agreed upon, could give huge hope for saving our climate and huge hope to the poorest 1% who are being taxed for something we really should encourage (i.e. good work).



Please stop giving me things for free

Dear Spirit Airlines,

Boeing 727. Source: Champion_Air_727-2S7_LAX_N686CA.jpg: Aero Icarus derivative work: Altair78

I recently booked my first flight with you Spirit and was pleasantly surprised. You are the first airline I have seen to charge separately for carrying on a small sized suitcase. As a passenger who wears only shorts, a light white vest and sandals – even to the winter olympics – I greatly appreciated the gesture. Gone are the days of me having to subsidise my lavish heavy-packing brethren. As a European familiar with Ryanair I appreciate your efforts to become a “no frills” airline but I must be frugal in my praise. I know your innovations in breaking out prices go beyond carry-on luggage but I would argue that you really are not going far enough with this strategy. The freebies you have eliminated are only the tip of the ice-berg in terms of what is ultimately possible. Please stop giving me things for free.

The opportunities I see for cutting down on freebies are endless. Take, for example, those passengers who prefer to travel wearing shorts – like myself. Surely the 200 grams of material weight saved – compared with trousers – are deserving of a specific discount on my fare. With global warming in the public eye there surely are benefits for the consequent carbon dioxide savings too. Imagine a full flight of people wearing shorts… that could be up to two hundred times 200 grams of weight saved. PLUS, every flight would look like a trip to the Canaries – surely that ambiance alone would be worth charging for too. I don’t want to elaborate too much on the knock-on benefits of this policy but let me not finish without noting that, with a plane full of shorted passengers, the requirements for air conditioning would fall down too.

The second topic I’d like to explore is the in-flight experience. You may feel that, given existing sales of sodas, snack-packs and multi-combined-special-offer-combo-deals, this is an area in which further breaking down prices will generate diminishing returns. I beg to differ and believe that a wealth of unexplored avenues are right outside the window. Literally, I think there are opportunities right outside the window. Personally, I rarely look out the window and I firmly believe that such frugal behaviour is deserving of a lavish discount. You may think that sitting in a window-seat and looking out the window should come together but I would disagree. It is perfectly possible to sit by the window and not look out, or to sit in the aisle and look out of the window almost all of the time. Without breaking pricing down further, such a system really cannot be economically just or fair. Where you sit and where you look can, and should, be priced separately. To go even further – and I see absolutely no reason not to – I feel the window looking experience can easily be split up. A two times premium for good weather, a three times premium for good scenery and a five-times premium for take-off and landing seem to me an excellent path to go. Moreover, with so many options, you could create some combo deals that mirror or even comingle with combos on snacks. A ginger beer, three peanuts and a look out the window for landing – who wouldn’t part with $10 for a combo like that?

My final point of note relates to billing. Clearly, if you haven’t already, it would be beneficial to create your own specific credit card, directly or via a deal with a third party, and then charge customers a bonus fee for the pleasure of using any card other than that. This is a bit obvious so I won’t make this my main point. I will note, however, the added benefit of passengers carrying one more credit card in their wallet and the consequent opportunity to charge for carrying their additional credit card weight. This credit card strategy is really only the first step on the way to being a price-separating black belt. To reach those heights, if you do so aspire, I would suggest continuous billing. At each moment in time, passengers can control; the moments at which they would like to yawn, the amount of air they instantaneously wish to breathe, and, whether they would prefer to maintain a happy, sad or neutral face. The instantaneous price of altering each of these three control settings – along with 747 other settings that I haven’t the time to describe – would all appear on an interactive screen that looks all but a Bloomberg terminal, complete with historical price tracking as well as options, forwards and futures. Decades ago buying a flight involved only one decision – either you bought the flight or you didn’t. We’re now so far beyond that; customers are constantly considering whether to shell out on luggage, insurance or snacks, and continuous billing is obvious next step.

All in all, I wrote this article to ask, in an absurd way, whether this is a trend that has gone too far. Being charged for services that were traditionally free is a painful experience (take a listen to this NPR episode about free) because, as Daniel Kahnemann’s book – Thinking Fast and Slow – points out, giving something up feels about twice as bad as gaining that same thing feels good. Constant decision making about buying things is also stressful for passengers (and more generally, consumers) because people really hate being peppered with decisions where they have to considering giving things up. It seems to me we should ask whether businesses are taking into account the drawbacks of taking away what was free, and the drawbacks of forcing consumers to constantly take decisions. Perhaps these drawbacks are hurting businesses in the short term and hurting their brands in the long term. Not everyone wants an all-inclusive package holiday – but perhaps it isn’t random that straightforward all-inclusive (or partially inclusive) deals exist or came about.


Marketing is the solution to our world’s water problems

It’s not technology! it’s not finance! Advertising, or perhaps more correctly, marketing is the solution to our water problems and here is why.water

There is a great solution out there to improving our water supplies. That solution is to reuse the waste-water we produce. As Paddy Padmanathan of ACWA Power pointed out today at the world congress in San Diego, only about 2.4% of our wastewater is currently reused. Much of our waste water (at least in the developed world) is already well treated, and, with just a little more effort it can easily be brought to drinking water standards. Clearly there’s a huge opportunity to reuse waste water in the way I’ve just described – known as direct potable reuse – so why aren’t we doing it?

One non-reason why we aren’t doing direct potable reuse is cost. When freshwater sources run dry and conservation measures come in, direct reuse is generally the next cheapest option. For example, desalination might come in at $1.60 per thousand litres of water while direct reuse might come in at $1.08 – these are the numbers Borja Blanco described today in his assessment of water solutions for La Serena in Chile. Wastewater has less salt than seawater and that makes it cheaper to purify. So, when water runs out economics say that we should be doing direct reuse rather than desalination – but we aren’t!

Another non-reason why we aren’t doing direct potable reuse is financing (i.e. finding the money to build the plant). Financing is a non-reason because society has become very good at figuring this out. In fact, it could be argued that water treatment companies now compete more based on their ability to do financial engineering (find the money) rather than technical engineering (find the right technology). At least that’s the impression I got from the CFO of IDE, Gal Zohar, who spoke today about project financing. Different governments have different requirements (e.g. being able to pay upfront vs pay annually) but Mr. Zohar seemed confident that, having experience with a wide variety of deal frameworks, there is usually a solution.

That leaves the main reason why we aren’t using the great solution of direct water reuse – marketing. Ultimately, in La Serena in Chile it seems (albeit anecdotally) that people are uncomfortable with drinking and using water that has been recycled from waste. This is true in the vast majority of countries it seems. In Singapore, by contrast, the approach has already been accepted. Why the difference? In my opinion it’s marketing. Singapore invested massively in a marketing campaign and created a brand of reused water called NEWater that gave comfort to people that water supplied was of excellent quality, in fact even better quality than what might otherwise be provided by the municipal systems.

For there to be a great step forward in society, there needs not only to be a new approach or technology but a clear articulation and justification to the public of why that change is fundamentally good for them and good for society. Right now some public infrastructure companies compete on financial engineering rather than technical engineering. The future, I believe, is one where companies will compete based on marketing – the leading company will be the one who can reassure us all, in a truly honest way, of a better future.


The Dirty Secret of Entrepreneurship

The dirty secret of entrepreneurship is that, to be successful as an (for-profit) entrepreneur, you need to excel at something that is useless in itself. In short, to be successful you need to excel at capturing the value you create. While, value creation can be productive for society, the efforts needed to capture value are, in themselves, unproductive. They are “the business model” or “a cost of doing business” – to use a few euphemisms. Here’s perhaps why.secret

The first thing I’ve recently learned is that, to be an excellent entrepreneur, you must excel not just at creating value you must also excel at capturing that value as a profit for your company. To be explicit, an excellent entrepreneur is someone who not just creates, say $1B in value, but who manages to capture a very large portion of that revenue as profits (say $999M, then you’d be a pure class entrepreneur). By contrast, one way to be an excellent scientist (just one example) is to create massive value; capturing that value for yourself is much less of a feature. Stop for a second and think about the difference… I take no credit for highlighting this difference, nor am I sure who should. Peter Thiel in Zero to One emphasises how entrepreneurs focus too much on the size of the opportunity and not enough on how much they will capture as profits. Bill Aulet, in Disciplined Entrepreneurship, makes the point that too many entrepreneurs spend lots of time on value creation and not enough on value capture. Overall, the emphasis on value capture makes sense to me too. To get massive funding upfront for a startup, it’s not enough to create value in general, you need to be sure that you can capture that value and do so in a durable way over time.

The second thing I’ve recently learned is that capturing value takes a lot of effort. Actually, I’m definitely spending more time on value capture than on value creation. Figuring out a strategy to capture value, writing patents, keeping trade secrets takes massive resources – I can only imagine that efforts to capture value will increase and not decrease in time. This fact surprised me – I thought startups were all about value creation. Value capture just seems strange, perhaps indirect?, perhaps uncomfortable?; and maybe that’s why not enough entrepreneurs do it effectively(?).

The last thing I’ve learned is the reason I wrote this. Focusing on value creation has made me realise one of the virtues of being a non-profit or being in education. For sure, non-profits, governments and educational institutions have massive challenges of their own that are frequently documented. However, one thing they don’t have, at least to the same extent as entrepreneurship, is the need to spend massive resources on value capture. As an entrepreneur, I think it’s important to respect that.


Top Two Reasons the Red Sox are Losing

According to a very nice Uber driver I met on Sunday the two reasons are simple: 1) The Red Sox are losing because that was the plan 2) The Red Sox are losing more than they should be because they aren’t sticking to the original plan.

via http://espn.go.com/mlb/standings
via http://espn.go.com/mlb/standings

Now, let me – in my absolute naivety – try to elaborate.

They are losing because that was the plan

Apparently there are basically two (at least two) strategies to running a baseball team. The first strategy is to buy players that are pure class – the problem there is that doing so is expensive. The other strategy is to try and develop young players – the problem there is that your team may not be great during development. In essence, apparently the Red Sox plan is the latter, i.e. to build up loads of quality young players, and that explains why they aren’t winning so much. While in transition the whole plan is roughly the following:

  • Let young players improve by giving them game time.
  • To carry the team over until the young players mature, buy some old (30 yrs +) players that aren’t absolutely epic but are very good value for how well they play (aka “value players” – I hope I didn’t just make that up).
  • Look forward to winning in a few years (which should be about 18 months from now, according to my Uber friend).

So, in summary the plan was to have a few years of mediocrity and win in the future.

They are losing even more because they didn’t follow the plan

A few years ago, the plan apparently started to go wrong. First of all, the Red Sox won out the whole league (World Series). Apparently, it’s possible for that to be a bad thing. According to my new found friend, the victory was completely unexpected. It was a fluke because the Sox had just bought in some “value players” that happened to play the best baseball of their lives on that year. The problem, he continued, was that victory set unreasonable expectations. In other words, even though the team was mediocre/decent, people would expect epicness. This put pressure on the strategy of letting young players grow. In fact, he went on, the Sox partially abandoned their initial strategy. They bought some pretty pricey players (Pablo Sandoval being one) and they let go of a “value player” Jon Lester [who was a top class pitcher but perhaps had only a few years left at his peak]. Now the Sox have less money in the bank but what’s more, their younger players aren’t getting the game time they need to improve.

Whew… pretty pleased to know that all of this sub-par performance is perfectly normal.


The most important consideration in an argument

I often like to ask myself whether I was arguing because there really wasn’t enough at stake. Many times I find this to be true and I’d like to try to explain why.birds fighting

The first time I started thinking this way was when I heard the following quote – “In academia, we fought so much because there was so little at stake” – a (probably modified) version of a quote attributed to Henry Kissinger. I feel this is true way beyond academia.

When things are going well, in a non-profit, in a startup, in a social club, etc., there is progress and there is growth. People find themselves with more responsibility and there are more spoils to share. It is easy to have new people join because there is so much to work/responsibility/money/etc. to go around. In many cases, when a tough decision needs to be made, it makes sense to take quick action because stalling slows progress and that just hurts everyone. There’s a big incentive to work together towards growth and holding things up with arguments is bad for everyone.

Now, think about the opposite scenario where things are heading south. There’s less work/responsibility/money/etc. and people find that they are beginning to overlap in what they are doing. There is overlap in responsibilities, there isn’t enough work to go around and, suddenly, there seem to be too many people to share the spoils*. Ultimately this can lead to the boredom, the firing of employees or other specific things. What’s probably sure though is that it leads to unhappiness.

So, what can we learn from this. Well, I think it’s really useful to recognise, after an argument, whether it is a situation of growth or decline. If it’s a situation of growth then you should ask whether you are holding up progress and should find a quick compromise. If it’s a situation of decline, then maybe you should ask whether it’s best to part ways and look in some other direction.

Ok, so clearly I’m leaving out situations whether there are arguments in high stakes situations. Fortunately this is a blog; so I can just keep it short and be incomplete.

*Actually, this matter is probably compounded by the fact that people are much more averse (roughly twice it seems from reading Thinking Fast and Slow by Daniel Kahneman) to losses than to gains. So if your organisation is shrinking it probably feels twice as worse than if it grew at the same rate.


The best way to make a fortune investing in startups

The strategy of investing in hundreds of early stage startups per year appears, at least in one case, to allow a return of more than 80% year on year. This is the first of a series of blogs addressing this type of strategy, pioneered by Y-Combinator, Techstars, amongst others, and known as the “Accelerator” or “Incubator” model for startup investment. In this blog, I take a look at the annual performance of accelerators through the lens of Y-Combinator and show that its performance is absolutely outstanding compared to any other type of investment or portfolio.

What is Y Combinator?

Y Combinator, or YC for short, is a company that, in my understanding, specialises in making large numbers (~200 per year) of small early stage investments (as of late 2014 they typically invest roughly $120,000 for 7% of each company) in startups. YC then provides 3 months of intensive coaching, followed by an opportunity for the startups to pitch to other investors to raise a seed round of funding (typically low seven digit range in exchange for 10-40% of equity as I understand). In the end, the result is that YC ends up owning a small, but not insignificant, stake in a very large number of startups, some of which reach very large valuations.

How much money is YC making?

The key to understanding profitability rests in identifying in the average annual growth rate of a $1 dollar investment that an investor/owner might make in YC. To do this I have taken a Warren Buffett type approach and calculated the average per share return achieved by YC. The whole reason for using this per share approach is because new investments are made in startups by YC every year. This means that as time passes, more and more money is invested in YC (and by YC) and the profits/returns therefore have to be split amongst a greater number of initial dollars invested. The per share approach works so that, as more dollars are invested in YC, YC issues more shares as a way to figure out how to split the profits. Therefore, through time, both the number of shares issued and the book value of YC increases, and, what matters for investors is not just growth in book value but growth in book value per share**. To understand more about how the growth in shares is calculated you can read below***. To understand how the book value of YC increases you can consider the following:

  1. For every year I have calculated the book value of Y Combinator. There are two ways in which this can increase.
  2. The first way is that when YC invests in a new batch of startups then its book value increases by the very amount that it invests in these startups. This is straightforward addition.
  3. The second way that book value can increase is if the valuation of any of YC startups increases. To do this with complete accuracy we would have to track every single company. However, a shortcut is possible. Four YC startups (Zenefits, Stripe, Dropbox and Airbnb) account for the majority of YC’s value. By considering the increase in valuation of four companies alone, it is possible to get a good estimate of the increase in book value of Y Combinator through time*.
  4. Of course book value could also decrease. Indeed, when startups fold then the money that was initially invested would have to be written off. However, this fact is of little importance for our calculation because the book value of YC is dominated by the value of the large successful startups. The amount invested upfront in failed startups pales by comparison.

You can download the spreadsheet here: Return on Book Value Calculations.

Why are the returns offered by YC impressive?

Annual returns above 80% are absolutely sensational. For comparison, a typical annual return for venture capital over a 10 or 15 year period is around 10% – eight times lower than what Y Combinator is currently doing! Of course, the big question here is whether there is enough data to determine whether what Y Combinator has done is more than a stroke of luck. That question will form the subject of the next blog.

For now, here are some questions to think about (and comment about below):

  1. Do you think there is enough data to conclude that Y Combinator is truly beating other asset classes by a large margin? Why or why not?
  2. Is there a way to tell whether returns will fall as the program becomes larger? Why or why not?
  3. In 2014, Y Combinator increased the amount invested in each company (and also the valuation at which money is invested) – how might that affect returns moving forward?

*If you look at all of YC’s large companies (above $100 million) you find that considering the top four (in fact really the top two) accounts for most of its valuation.

**It is also worth noting that, in practise, investment funds may be set up so that a certain amount of money is committed over 5 years and then profits are returned to investors in the next 5-10 years. This might initially seem inconsistent with the method of considering book value per share where money is effectively raised by the investing company as investments are required. However, this is approximately the case because monies that are committed to funds are often not transferred until they are immediately required and so I would argue that the book value per share method is a decent one to use.

***The calculation of the number of (fictitious) shares is best described by an example. Say, in year 1, COMPANY X raises $10 from its owner and invested $1 in 10 startups. The book value of COMPANY X would therefore be $10. Furthermore, let us assume that COMPANY X issued 1 share to its owner (i.e. that is, the owner of COMPANY X holds one share worth $10). Say, now, that by year 2, each of the stakes held in year 1 startups had risen in value to $2. The book value of COMPANY X would therefore have risen to (10 X $2 =) $20 and the price per share would have risen to $20 per share (since there is just one share). Now, let us say that COMPANY X wishes to raise an additional $20 so it can invest $1 in each of 20 different startups. In that case, COMPANY X would issue one new share in return for $20 of investment. The total book value would then be $20 (from the ten year 1 startups) and $20 (from the twenty year two startups). If you think through this carefully, you’ll see that if there is no appreciation in the value of startups that are held by COMPANY X then the price per share will stay constant. If the startups appreciate in value the price per share will increase. As previously mentioned, if startups fail they will reduce the book value of COMPANY X because the investments need to be written off. However, this tends to be negligible if there are one or more startups that appreciate to very large valuations (as is the case for YC) that dictate YC’s overall valuation.

This blog should not be interpreted as financial advice. The calculated book values are estimates and the numbers of shares are fictitious. The level of uncertainty involved in dealing with data sets of this size and volatility is high and any conclusions must be subject to careful skepticism.


Taking out the Salt

The large amounts of water required in the hydraulic fracturing of shales for oil or gas is a pressing concern. In hydraulic fracturing, water is sent  down underground at high pressures to fracture the rocks below, allowing gas or oil to be released and captured. When water returns to the surface it often contains large amounts of salt that were dissolved from the rocks down below. These large quantities of salt have made it difficult to re-use the water for further fracturing processes – making it difficult to reduce water use by adopting recycling strategies.

Water produced from an operating oil or gas well, usually very salty after contacting underground rocks, can be cleaned of its salts and other contaminants using electrodialysis, and then reused to reduce the amount of freshwater needed. This diagram illustrates the process, with salty water in dark blue and fresh water in light blue. The electrodialysis process, using membranes and electric charges, is illustrated inside the circle. Illustration: Jose-Luis Olivares/MIT
Water produced from an operating oil or gas well, usually very salty after contacting underground rocks, can be cleaned of its salts and other contaminants using electrodialysis, and then reused to reduce the amount of freshwater needed. This diagram illustrates the process, with salty water in dark blue and fresh water in light blue. The electrodialysis process, using membranes and electric charges, is illustrated inside the circle.
Illustration: Jose-Luis Olivares/MIT

Earlier this month, my co-authors, Adam Weiner, Lige Sun, Chester Chambers, Prof. Syed Zubair and Prof. John Lienhard and I published an article describing how electrodialysis can be used to remove salt from these waters to facilitate greater reuse. We were able to show that, compared to the evaporators are currently available to purify these waters, the system we proposed was of similar energy efficiency but more cost effective. We feel that electrodialysis, whereby salt is removed from water by means of an electrical current, is promising for this application – the next steps are to see how the system will fair in real world rather than lab conditions.

The author’s copy of the manuscript may be downloaded here. The article published in Applied Energy is available here and the MIT news office article appears here.


The benefits of hybridising electrodialysis with reverse osmosis

In a recently published paper in the Journal of Membrane Science, Prof. Syed Zubair, Prof. Lienhard and I reported on how the cost of achieving high purity water with electrodialysis, an electrically driven technology, can be reduced through hybridisation with reverse osmosis, a pressure driven technology.

In reverse osmosis, multiple stages are required to recover a significant portion of the feed water. However, the membranes block salt very well and the product water is very pure.
In reverse osmosis, multiple stages are required to recover a significant portion of the feed water. However, the membranes block salt very well and the product water is very pure.

In reverse osmosis, water is pressurised and forced through a membrane that is capable of blocking upwards of 99.5% of salts. In electrodialysis, an electrical current is passed through a stack of membranes and pulls salt ions from a feed stream (known as the diluate) into a more concentrated stream (known as the concentrate). Reverse osmosis can provide very high purity water, but the feed water must pass through multiple stages before a large fraction is recovered. Meanwhile, in electrodialysis, because salt, rather than water is removed, almost all of the water can be recovered as a final product, but high product purity requires multiple stages to ensure all of the salt has been removed. The idea of hybridisation is to leverage the ability of reverse osmosis to achieve very high purity and the ability of electrodialysis to recovery a large portion of the feed water as product.

In electrodialysis salt, rather than water, is removed from the feed water. This means that the vast majority of the feed water is recovered as product water. However, to achieve excellent purity multiple stages are required.
In electrodialysis salt, rather than water, is removed from the feed water. This means that the vast majority of the feed water is recovered as product water. However, to achieve excellent purity, multiple stages are required.

We analysed different ways of combining reverse osmosis and electrodialysis systems and compared the cost of water from these systems to the cost of water from standalone electrodialysis systems. We found that the hybrid systems allowed for the greatest cost reductions when high product purity was required. This is because the number of stages of electrodialysis required rises rapidly as higher purity is needed. The inclusion of reverse osmosis can reduce this requirement.

Ultimately, this methodology can be used by engineers to guide their choice between hybrid and standalone systems. The operation of hybrid systems is still unproven but this study suggests there is good potential for these systems to reduce water costs when high purity is required.

This blog is based upon “The benefits of hybridising electrodialysis with reverse osmosis” by Ronan K. McGovern, Syed M. Zubair and John H. Lienhard V, published in the Journal of Membrane Science.

The author’s version of the manuscript may be downloaded here.