The Price of Loyalty

One of my very first “proper” jobs involved managing projects that took enormous data sets from blue chip clients, and “mined” them for insights that their marketing teams could put to good use.

It was the late 1990s and digital data collection was fairly rudimentary compared to the tools and systems marketers have access to today. We ran one “black box” analysis that took two days to churn through the customer data set, thanks to the slow processing power of PCs back then, which was just a fraction of the millions of customers the company in question had. I was project managing a team of incredibly intelligent statisticians who were running queries on the meta data. The analytical output was fascinating.

For one financial services client we were able to correlate really random purchase patterns, like how people with certain star signs were more likely to apply for bigger mortgages.

The gold standard in the data analysis market was Tesco Clubcard.

Long before Meta became a data overlord, Tesco was perfecting the metadata analysis and putting it to good use to sell more products in their stores.

Tesco was able to predict from their data that placing condoms next to the white wine on a Friday night would directly increase sales of both items. Similarly, putting socks next to the black pudding on a Sunday morning was also a winning adjacent shelf combination. Aren’t customers weird?

Of course, in the early 2000s Meta came along and flipped that whole data mining model on its head. Meta decided that if it could extract enough data from human beings, it could cleverly make businesses pay them lots of moola so that they could sell those humans practically anything. An ambitious mission. A clever mission. And it worked.

But only one of these data loyalty models is sustainable. The other is built on fickle foundations.

Tesco has real stores, selling real products, to real people. They can shuffle their stock around the shelves, and optimise stock volumes. The fundamental truth is: People need to eat. Every day.

Meta, on the other hand, can only shuffle the algorithms. It relies heavily on the advertisers getting value for money to make its money. And advertisers are fickle, they only spend money if the advertising works. Advertising does work, but only if it reaches eyeballs first. If the humans don’t scroll, they have nothing.

All the anecdotal evidence I’m hearing from fellow planet dwellers is that they are done scrolling. No scrollers, no people. No people, no advertisers.

Tesco, on the other hand, is still opening its stores every day, and still selling food, and still collecting data.

Meta, in contrast, has problems. It has admitted daily users are in decline. Its answer is to invest more in AI, not to address why 20 million people deserted the platform last quarter.

Twenty million humans. That’s four times the population of Scotland deserting the platform in just 3 months. Ouch.

So now what for Meta?

Should they take all our data points and use them to train AI models? They are already doing that, and losing the AI race by the look of it.

And what happens when that data they harvested dates? And it inevitably will. Data ages quickly and every data manager knows: Garbage In, Garbage Out (GIGO). If the input data isn’t up to scratch, then the output data will be next to useless.

It’s fair to say that Tesco won’t be using their Clubcard data from 1999, although it might be interesting reading. Meta’s AI, on the other hand, could well be using profile data from 2015. It’s so last decade.

In fact, most of the AI models are full of useless data. Claude recommended that we collaborate with a celebrity today, despite the fact the celebrity had been dead for seven years. This isn’t a data hallucination, this is GIGO in real life. Their data is out of date.

Could this spell the end of the big global data experiment?

Probably not.

Have data, will manipulate.

But while Meta scrabbles around its AI replaced team for a totally new business model on which to base its future, us humans will still be trotting down to the supermarket to grab a bottle of white wine on a Friday night and, I bet, only Tesco knows what else.

There’s more where this came from. I share more of my thinking on Cunningly Good Marketer on Substack. Subscribe to get my opinions, experience and real-world lessons straight to your inbox.