Summary
- I have spent thousands of dollars on Facebook, Instagram and Snapchat ads. In some cases they resulted in zero sales, in other cases the cost per sale far outweighed my profit per sale.
- The effectiveness of digital ads is impossible to measure because there is no way to conduct a randomised trial. AB tests that platforms promote are not statistically valid.
- Digital ads are presented as having the superpower of “targeting”. It is hard to prove that targeting delivers a lower cost of customer acquisition. It is much more likely that targeting results in prices being bid upwards.
- The concept of customer life time value is being used to encourage startups to spend more on customer acquisition via ad spending. Credit Jaffer Ali.
- Where digital ads appear to generate a return, you have to ask the question of whether you would have had those same sales without doing any ads.
- Many of my ads have been trolled in the comments. Others have annoyed would-be customers who feel they are being spammed. This hurts sales long term.
- A healthier approach to advertising is a breakeven approach, whereby you breakeven on marketing costs on your first sale.
- With digital ads it is very hard to break even on your first sale. For now, I believe it is best to consider other forms of advertising such as e-mail advertising.
References
The following three references are central to my discussion below and I highly recommend each of them:
- Rory Sutherland’s Alchemy – An entertaining read covering the physchological aspects of marketing. To quote David Ogilvy – ” The trouble with market research is that people don’t think what they feel, they don’t say what they thing, and they don’t do what they say”.
- Jaffer Ali’s interview on Risky Conversations. Jaffer has an eluminating perspective on the drawbacks of digital ads and typical eCommerce businesses, as well as unique approaches of his own on marketing.
- Freakonomics multi part series on whether advertising works. Some may find portions of the series too academic, but the questions raised around the value of advertising are good ones.
I believe advertising works, but there may be a bubble in digital ads
As Rory Sutherland likes to say: “A flower is a weed with an advertising budget”. Our emotions can certainly be controlled by what we see and hear. Our emotions affect our purchasing patterns. I accept this and I accept that people would not be spending so much money on advertising if it was useless.
However, it is entirely possible that advertising works, but digital advertising is in a bubble. My direct experience with paid advertising, and my reading of the references above, have brought this to light for me, although I am far from the first.
My Experience with Digital Advertising
To date, I have spent a few thousand dollars, maybe even ten thousand dollars – through hand sanitizer and non-alcoholic beer businesses – on paid ads. I have:
- Tried Snapchat ads for non alcoholic beer – yielding zero sales for hundreds of dollars in spend.
- Tried Facebook ads – yielding some sales, but at a cost of over $30 per sale – way higher than the profit per sale.
- Tried Google ads – yielding no sales for non-alcoholic beer and yielding some hand sanitizer sales (back when sanitizer was scarce) at a huge cost.
Digital Advertising as a Black Box
First off, you may argue that I am not doing it right. I accept that argument, but I am not going to address my own incompetence in this article!
By trying paid ads, it has become clear to me that here are aspects of paid advertising that fundamentally make it hard for advertisers to generate a strong financial return:
- The advertiser only has the appearance of having control over how ads are shown. Actual control is in the hands of the platform.
The problem is – you have no visibility into exactly how those targets are being chosen behind the scenes. One week you may hit a certain 10,000 users. Another week you may hit another 10,000 users. I have never run the same campaign twice and achieve similar results. I was like a lab scientist running tests but with no control over the test parameters to make things a fair comparison. Even with the AB tests that digital platforms allow, you can run the same AB test multiple times with different results.
When I go into Facebook (or other platforms), I select a group of people to target. I have access to parameters such as geography or job title or age. Based on these settings, I can define a target group – let’s say of 100,000 people. I then set an advertising budget (say $100 per day). That advertising budget determines how many people – within my target group – the advertisements will hit. The problem is, I don’t decide exactly how these groups are defined, and who within the group will be hit with my ads. The system is a black box that gives the appareance of control but can easily be designed on the back end to optimise for the advertising platform’s revenue.
2. The complexity of the domain in which advertising occurs is vastly underappreciated and is misrepresented.
My second criticism is a statistical one. It is very difficult in a complex system to obtain meaningful comparisons in performance between ad sets. Measuring the performance of a single campaign is easy. Comparing one campaign to another is – in most cases – statistically impossible. Platforms like Facebook are accidentally or deliberately giving the appearance of a measurable process by offering AB testing systems that give the appearance of control over the advertising process. The reality is that there are far too many variables involved in ad performance that cannot be held constant in conducting a comparison.
Targeting does more to increase the price than the effectiveness of ads
Precise targeting of ads creates discrete markets that allow advertising platforms to have advertisers bid against each other and drive up process. This is a clear benefit to the platform
Much less clear is whether a high level of targeting can generate a strong financial return for advertisers. The more I target, the smaller my audience and the higher my advertising costs per sale (especially if I have competition). I need to think more about this but, as I do more targeting, I see myself going up a price curve that is progressively steeper. In other words, I’m getting less bang for my buck in terms of advertising dollars as I do more targetting.
Furthermore, I think targeting is over valued because it is easy to measure performance based on the acquisition costs of your final target audience, but forget about the advertising costs you had to pay for in order to test and find what your proper target audience should be.
Lastly, from the perspective of the advertising platform, they love that I am spending money to find out who my target audience should be – because – once I get to my “target”, they are still the ones in control and can still charge me up to my gross profit in order to give my access to that target! Any advertising monies spent on testing is a sunk cost to me and a sunk profit to them! It’s not like I can have a preagreed arrangement where I say, “I’ll pay $100 to find my target audience but then you’ll agree not to screw me over on pricing once I get to that target. Ok?”. “No”.
Life time customer value as a business school philosphy to drive advertising spending
As Jaffer Ali points out, the “razor – razorblade” narrative pervades business school, corporate and startup thinking. Every investor is looking for the recurring revenue model where – yes, you may have to sell a razor upfront for cheap – but you’ll make it back selling razorblades. Ironically, for the original razorblade company (e.g. Gillette) this hasn’t worked out so well recently given Dollar Shave Club and Harry’s getting on the scene with their cheaper blades.
Lifetime Value or LTV is where you take all of the future value of a customer and discount it back to a present day number. As modern textbook thinking goes, so long as the lifetime value of a customer is signficantly in excess of advertising (customer acquisition costs, CAC), then you are golden. The problem with this narrative, is risk. Summing up the future value of a customer often assumes that the future will look like the present, and it often does not.
For advertising platforms, it is convenient to set risk aside as a real parameter. All that is needed is to push the narrative that their ads provide you with customers with great future value. If you have a customer worth $100 over the next three years – why wouldn’t you spend $50 on Google ads to acquire them today? Easy! Why not even pay $99?
Sometimes no ads are just as good as ads
It’s a very hard argument for a marketing department to completely turn off their advertising. However, without doing this, it is tricky to know whether ads make a difference. In many cases (for example, buying google ads when you are already ranking highly organically in searches), it turns out that you are paying to advertise to customers that would anyway have found you. It’s like giving people a discount after they have already checked out!
Rory Sutherland likes to make a similar point in Alchemy.
Advertising can lose you customers
When I was doing Facebook ads, there was a certain percentage (maybe 20%) that would get trolled in the comments. For example, for Point 5 Non-alcoholic beer, there would be comments asking “What’s the point?” or emoji symbols of puking down a toilet. (The first one I found funny, the second I didn’t like for obvious reasons). At best I might say that “there is no such thing as bad publicity”. At worst, I worry this damages my brand.
Add to that the fact that many people hate ads and use ad blockers. I use mozilla firefox with privacy badger in order to block ads. I also tend to use DuckDuckGo, even though it’s not as good as google. People often don’t like digital ads and tracking, and I feel that some of that sentiment rubs off onto my brand if I use paid ads.
A better approach to advertising
I think it makes sense to think about customers in a long term kind of way when it comes to branding, mission and service. However, when it comes to advertising spend, I’m coming around to the view that it pays to be short sighted. In other words, if an advertising strategy doesn’t pay off very quickly – and it involves spending cash upfront – then I don’t do it. Here’s how I think about it:
- If I want a long term effect, then spend the money on making great content (videos, art, podcasts) where it isn’t possible to easily calculate a financial return.
- If I want a shorter term effect, only spend money on advertising to the extent it is breakeven within a short timeframe (30 days).
I take the view that paid digital advertising is overpriced at present and best avoided. I’m glad that I don’t own any Facebook or Snapchat stock any more. Unfortunately I don’t have any easy alternatives other than creating good content, cross promoting with others (businesses, influencers) that are drawn to my brands, and building up a database/sales channel that I can control fully (like an e-mail list).
If you liked this writing, consider subscribing to my monthly newsletter. Please also comment below with your own thoughts on digital advertising and if/where there are areas that you disagree.