Attention is not engagement.

That was the challenge a very senior advertising leader gave me when I was building SmartAssets, a creative effectiveness solution.

His point was simple. People love a cute baby on a tyre. They will click. You don't need data to tell you that, and you certainly don't need a platform. It's the same logic as the oldest trick on the internet: "Free sex, click here." We are base animals driven by instinct, and getting a human to look at something is not rocket science.

To some extent, I agree. Getting any basic attention is easy, if you're prepared to lean into our base mammalian instincts. But what is it worth?

We used media data to work out not just which static images and videos engaged people, but which specific element inside the asset was driving that engagement. The promise was to move past "we think this worked" to "this is the thing that worked, and here is why."

We worked with one of the world's leading baby care brands. We assessed all of their social content over a period of time to find out what moved the needle. The finding ran directly against the baby-on-a-tyre thesis: babies do not sell nappies.

Babies are eye-catching. We are hardwired to notice them, and the content with babies in it got attention. But attention is not the same as engagement, and it is a long way from a purchase. The person buying nappies is not the baby. The parent or carer does not look at a cute baby and identify with it. They think: I need this nappy not to leak. I need it to be affordable. I need it to be safe for my baby's delicate skin. Those are the things that engage the buyer, and the data demonstrated that assets with no baby in them performed just as well when they spoke to what the audience cared about.

So attention and engagement are two different things. Attention is the cheap, instinctive bit, and it is hardwired, which is exactly why it is easy to trigger and easy to overvalue. Engagement is about what the actual decision-maker cares about.

This matters more now, because of AI.

When you can produce a great deal automatically and easily, the barriers to entry of making and distributing something fall lower and lower. The approach of just getting a campaign done as fast as possible, and iterating from the learnings, is available to everyone. The focus then has to shift from "how do I make this" to "why am I doing this, and why should anyone care?"

The ability to create content at pace and be present in the multi-screen, multi-channel world your audience inhabits still matters. But the thinking, the strategy, the analysis, the work of getting to the crux of why anyone would ever care about the thing you have produced, is critical to success. The nappy data is a small version of the same lesson: the easy, automatic answer (put a baby in it) was not the answer that worked. Someone had to ask what the buyer actually cared about.

I recently read Dr Sam Illingworth on Substack arguing that now is the time for all the humanities kids: the people who studied linguistics, philosophy, and the like, to come forward and help with AI. I can't help thinking that's true. When answering the "how" becomes this accessible, the "why" is what makes the difference.