In 2022 Meta’s share price was down 35%, and I put the number at the top of an article for business owners with a question under it: do you know what they’re trying to do? Yep, make more money. Marketers have a dichotomous relationship with digital advertising channels like Linkedin or Google. We want to use their platform as cheaply as possible. They want us to use their platform as expensively as possible. I’d been buying traffic from those machines for ten years by then, and I still recommended them, with one caution, which I still mean: we never know what Facebook or Google will decide to do tomorrow.
With the invention of the platforms there were very real opportunities, and over the long term I suppose they were arbitrage opportunities, but in the moment they felt like permanent structural shifts. When Facebook first came out there was a lag, and any company that went on and used it properly grew their brand exposure massively. Someone extremely talented at Facebook got a job at a traditional shoe manufacturing company to set up a direct e-commerce channel, built it on Shopify and ran Facebook ads, and because no other shoe manufacturers were going direct, those ads were causing the only brand awareness in the market. Everything goes well, sales take off, blahdy blahdy blahdy blah. Let’s call them Alex. Alex is a composite, not a client, and I’ve built them out of more than one person. Alex gets promoted and promoted and promoted, from their tactic, which is Facebook, up to their channel, which is e-commerce, and then further up within the organisation, into marketing manager or CRO or CEO or whatever it is that strokes the ego of the moment, and then the lag closed, because everyone joined, and it all averaged back out, and the platform lessons and tactics don’t really hold up to the lens of time. So there’s a lot of people like this who are really struggling, because they lack a more formal marketing education.
Because these platforms were such multipliers there was a new breed of marketing position, and it’s led to a whole lot of new words that mean the same thing as everything else. Growth hacking, to advertising, to media buying, to ad strategy, all of these things, they’re all very very similar. The advantage was an arbitrage. It got mistaken for a strategy, the strategy got a job title, and the job title gets a new name every few years. Everything is the same thing reinvented and rebranded.
Retargeting advertising audiences work with retargeting buckets, and there are many, many different ways to get put into a bucket. There are static indicators like demographics, and I would primarily use those to exclude people. There are behaviour indicators, owned data: if you’ve gone on someone’s website and they’ve got a tracking cookie, we could retarget you. And there are what I call in-market indicators, where you’ll get put into a bucket and made available to me to target you based on what you do on other websites. And that’s the role of the ad platform to aggregate that information. If Google owns a search platform, it can then sell your search data directly to me, and if it doesn’t own the platform it can’t really sell that data, so how do the people who own other platforms make money? They get ads. You install Google tech onto your website, then Google can put ads there and they can sell the data from that website. And so the network grows.
On Facebook it’s very similar, because a page gets classified into a topical entity. Not that many people on Facebook are actually going to go and like financial services pages, but what there is a lot of on Facebook is financial services ads. So whenever someone begins to interact with a financial services ad, that’s what they’re doing on one layer, but in the information layer, what they’re actually doing is interacting with a financial services page. And that interaction then will place them into all of the audience targeting buckets. So when one company spends a lot or does well or kind of breaks through, all of the other companies’ audience targeting in that market can become more profitable. The funniest one is a little checkbox on Facebook, which is I’m happy to display my ad alongside other advertisers, similar advertisers, which is just kind of humorously transparent. Ads are auctions originally, and now as it’s got more algorithm based it’s hidden and obscured between more layers, but there is a human auction, and if people bid less, the price goes down. I never see who I’m bidding against. I see a cost per click in a dashboard and decide whether I can afford it, and the platform sees every bid on every bucket, and it filled the buckets. Nobody who bids in that auction is standing anywhere they could watch it.
The classification isn’t even good. Let’s say you look at a page about paint stripper, and then if the classification is poor, it could put you into two different buckets, both paint and strippers, and then you would start to see ads relating to those interests, if someone is targeting people who are interested in strippers or in paint or in both. It has a word off a page and two buckets the word could fall into, and it fills both.
On 17 April 2025 I recorded the introduction to a marketing book into my phone, and the middle of it is an exercise. Close your eyes. This is a horrifying exercise. Mentally scroll through your browser history. And count how many of those websites you pay to use. All of the other creators are incentivised by money. They are advertisers. Advertising runs the internet. Every free site on that list has the tech installed, and every page I’ve read on any of them is a bucket I can be rented out of. I run the campaigns and I’m in the buckets.
Each advertising channel has an audience, and within each audience there are a variety of people. In 2022 I drew them as a pyramid with four layers, Cherries at the top, then Solution Aware, then Problem Aware, then Not Problem Aware at the bottom, and I put a percentage beside each layer with a note that the percentages are 100% bullshit, but make the concept easier to understand. The bullshit number for the cherries was 15% of the total audience. When I set up targeting for a new ad the platform’s algorithm goes to work finding the best people to show it to, and if it’s working properly it starts with the cherries, and after it has exhausted that pool it moves, theoretically, to the solution aware, and so on down. As your ad is shown to people further down the audience pyramid your conversion rate will decrease, your CPL/CPA will increase, and your budget will provide less for your business, month after month. The other potential outcome is that the platform never moves down at all and repeatedly shows your ad to the same cherries. This is especially easy to do in New Zealand as we are so small.
The 95-5 rule is that every interaction, the vast majority of it, will be with people who aren’t your customers, so it would make sense to invest in that interaction, because that’s actually what you’re buying 95% of the time, and if you’re investing only in the 5% that’s a ridiculous concept. The machinery for moving the other 95 up the pyramid has names. Funnels! User Journey Mapping! Retargeting! Marketing Automation! I hate funnels, it’s an overly fashionable word, and it’s doing the job growth hacking did, which is making a retargeting sequence sound like a strategy.
When you compete on SEO, there is a public scorecard. If you spend $10,000 and get no leads, everybody in your organisation can tell. If you accidentally email the entire database, everybody knows. So the operator watching cost per lead climb on a scorecard the whole building can read does the two things the platform sells, spend more and add a layer, and I’ve done both. Platform operators know how to do everything, and they know that they can spend more money on their platform, and there’s a lack of people saying oh this is the right amount to spend on this platform. How, how, how, never why or when or how much. That’s how an Alex is made.
Everyone in my market rents the same buckets, filled by each other’s spend, bids for them in an auction none of us can watch, descends the same pyramid, and hires the same person under this year’s name. The machine that was supposed to make my business visible is the same machine for everyone I compete with. If everyone rents the same auction, what happens to being different?
