You understand your marketing calls on the replay

[Read me if you've ever understood exactly why a marketing call went wrong, about a quarter after you made it.]

It's 22:40 on a Saturday and a former international rugby player is standing at a touchscreen, freezing the replay of a try that never happened. He draws a circle round the outside centre. He draws an arrow to the winger, unmarked, with twenty metres of grass in front of him. He rolls it back to the exact frame where the centre's eyes went to the defender instead of the space and explains, very calmly, why the pass never came. (Easy to be calm when two back-rowers aren't running at you.) He's right about all of it. The centre has probably watched it on the team bus and agrees with every word.

What the replay can't do is make the centre pass next Saturday. There's also a harder question the studio never quite gets round to asking, which is whether he should have passed at all. Maybe the drift defence was already coming across and the pass was an intercept waiting to happen. Maybe holding on was the right call and the pundit is explaining a mistake that wasn't one. The replay is great at explaining why he didn't pass. It's much less help on whether he should have, which is the only bit the centre actually needed, and it always shows up after the whistle.

You have a studio too. It's the quarterly review after you cut the brand budget to protect pipeline. Branded search has gone flat, and you can name what happened with perfect precision. Loss aversion. The euro you might lose on brand felt bigger than the euro you'd gain on demand, the quarter was tight, the CFO was watching, and you took the safe option, which it wasn't. You're right about all of it. You'd make a decent pundit.

If you winced at that, welcome to the studio. This is the Metacognitive Paradox, and in plain English it goes like this. Knowing the name of the thing doesn't stop the thing. You can spot a bias, name it, even explain it to your team, and still make the call it was pushing you towards.

When I went back through the 13 marketing leaders I interviewed for my thesis, 12 of them demonstrated awareness of at least some cognitive biases. That's a much lower bar than naming them explicitly (roughly one of the 13 did anything like that). Separately, 10 of the 13 showed the activation trap, with short-term performance winning out over long-term growth. Two different counts, measuring two different things.

The awareness was real, though, and it's worth looking at what it actually did.

One leader, on a marketing team of about three people with roughly 10K a month to spend, lived through the period when the marketers they followed on LinkedIn were broadly agreed that Google would punish anything an LLM had touched. They're a firm believer in generative AI. They held off anyway on AI-assisted writing and a lot of content expansion, because the people they took to be industry professionals kept saying the penalty was coming. (To be fair to them, it was a very loud consensus, and loud is easy to mistake for right.) Looking back, they described having been inside a bubble, only ever seeing the inside of it, and reckoned they could have pushed harder as long as the quality held up. Every part of that diagnosis is sharp. All of it arrived after the whistle.

Another leader, a solo founder with a marketing budget of one to two thousand euros a month, looked into listing on an affiliate platform. The listing alone came to about 12K a year, around a thousand a month. Budget, they told me, wasn't really what stopped them. What stopped them was a return they had no way to estimate, and they could say exactly why. Their experience was in paid lead gen and e-commerce, where they at least had a feel for whether a guess was in the right ballpark, and they had nothing comparable in affiliate to fall back on. So they passed. That's about as clean a piece of self-diagnosis as you'll get in a 24-minute interview. And here's the uncomfortable part. I still can't tell you whether passing was a mistake. It may well have been the sensible call. The analysis explained the hesitation perfectly and had nothing to say about whether the hesitation was wrong.

So that's two replays. One ran late, and the other ran without a verdict.

Now the part that's my opinion rather than what the research shows. A whole industry sells awareness as the cure. The unconscious-bias workshop, the forty-five-minute keynote on System 1 and System 2, the Cognitive Bias Codex poster with 188 biases on it (which, if you think about it, is a strange thing to hand someone who's about to make a decision). All of it is studio. It gives you better graphics, slower slow-motion and a sharper vocabulary for what went wrong last quarter, then sends you back out for the second half on the same legs, with the same board date. I don't think awareness is worthless. It's just a studio being sold as a training ground.

Here's what I'd do instead, and it's practice rather than a finding. Keep a replay log. At your next monthly review, write down three past calls you now understand differently from how you understood them at the time. Next to each one, write how long it took you to see it. A week? A quarter? A year? That lag is the number worth watching, because it tells you how late your awareness is arriving. If it's measured in months, then before you make the next long-horizon call, put a check-in in the calendar 30 days out with one question already written in it. What early signal would tell us this is drifting off course? The question is there so the call gets reviewed on a date you picked in advance, not whenever loss aversion gets nervous. It's a check, not a verdict (brand still gets its six to eighteen months, and killing it at day 30 is exactly the trap from Edition 2). You won't turn yourself into a better centre. You'll just stop waiting a quarter for the studio.

Which leaves the obvious question. If the people making these calls can see them this clearly, why do the same calls keep getting made? That's next week, and the answer has less to do with the leader than with the ground they're standing on.

The full pillar, including why the thing producing the bias is the same thing you'd need to override it, lives at The Metacognitive Paradox on gossamergrowth.com.

Somewhere this Saturday a centre will watch the replay on the team bus, nodding along. He'll be fine. He'll understand next Saturday's one perfectly too.