With A Straight Bat

For years, marketers argued about fairness. Who deserved the credit for a sale … the ad you saw last, the email you opened first, or the campaign that quietly nudged you somewhere in between? Attribution models became a kind of courtroom drama. Multi-touch, first-click, last-click … each promised a truer picture of how dollars turned into decisions.

That argument already feels dated. Artificial intelligence is changing the premise. The question is no longer which channel caused the action, but whether the action can be foreseen at all. Machine learning does not care about tidy chains of custody. It looks at signals, some obvious, most invisible, and predicts intent before a customer lifts a finger.

This shift is subtle but seismic. Attribution is about history. Prediction is about probability. In one world, budgets chase what happened. In the other, they chase what is expected to happen next.

The temptation is obvious: money moves faster, with less waste. But so is the challenge. Prediction is not proof. Models trained on the past still carry its biases, and confidence can outpace accuracy. A misplaced decimal in forecasting is more dangerous than a misplaced credit in attribution.

Attribution looked backward. Prediction looks ahead. One told us who to thank; the other tells us where to move. In a trade built on timing, that feels less like a gamble and more like progress.

Photo by cottonbro studio on pexels.com