Marketing has never known more about people. Every click, scroll, pause and purchase is measured, modelled and now summarised by AI in seconds. And yet most advertising has never felt more generic.
That is not a paradox. Data is excellent at telling you what people do. It is silent about why. And the why is where every great idea lives.
A dashboard cannot be surprised
Data sees patterns in behaviour that already happened. An insight is the opposite: a truth people recognise the moment they hear it, but have never said out loud. You do not find it in a spreadsheet. You find it in a kitchen, a queue, a supermarket aisle, a conversation that goes off script.
There is also an old trap here, known as Goodhart's law: when a measure becomes a target, it stops being a good measure. Optimise for clicks and you get clickbait. Optimise for views and you get noise. The number goes up while the brand goes nowhere.
Data can tell you that people are leaving. Only people can tell you what they are walking towards.
Go and watch
The best insights usually come from looking at ordinary life with patience. When I worked on the launch of an online grocery service, the data said people shopped in supermarkets. Watching them said something sharper: the trip was full of small daily battles nobody enjoyed. The campaign did not sell the service. It showed the battles. People saw themselves, and did the rest.
Use AI for the what. Keep the why human.
AI is extraordinary at the what: sorting, clustering, spotting the pattern in a million comments. Use it. Then take what it finds into the real world and ask why. The machine finds the signal. A human has to feel it.
- Start with behaviour, then go and watch it happen.
- Listen for the sentence people say when they think nobody is selling to them.
- Treat every number as a question, not an answer.
- Test the insight on a stranger. If they laugh or nod before you finish, you have it.
Data makes you informed. Insight makes you understood. Only one of them makes people care.
