Michael Kaminsky explains that every marketing result compares what happened to an invisible counterfactual: what would have happened without the campaign. Geo tests, A/B tests, and statistical models all try to reconstruct that unseen world. Without a control group like in clinical trials, you cannot prove the campaign caused the result.
Chapters
Every marketing result you've ever seen is built on something that never happened. What do I mean? Imagine two identical versions of your business. The same day, the same customers, the same market. In one world, you ran a marketing campaign and sales went up. In another world, you didn't run the marketing campaign. That second world, the one without the campaign, it creates something called a counterfactual. And the gap between these two versions of reality is how we think about the causal effect of your marketing. The problem here is obvious. You only get to live in one of these worlds. The moment you ran your campaign, the other world, it disappears from existence. You can never observe what happened if you hadn't run that campaign. So every marketing measurement method, every geo test, every model, every A/B test, really implicitly you're doing the same thing. You're trying to reconstruct the world you didn't get to live in. What would've happened if not for the marketing campaign. To glean from another domain, this is exactly why clinical trials have a control group. Half the patients get the drug and the other half get a placebo. The placebo group is the counterfactual. It's the closest thing science has to the world where nobody took the drug. Without it, you just have people who got better and would be left with no idea if it was the drug that did anything. Marketing measurement is trying to solve the very same problem, but with messier data, fewer controls, and a lot of money on the line. Whether in marketing or medicine, the question is never just what happened, it's compared to what.
