Why your next marketing leader will make the same call as the last one

[Read me if you've replaced a marketing leader and watched the new one make the old one's decisions.]

Walk far enough up a mountain in the Alps and the trees start doing something odd. Down in the valley a Norway spruce grows thirty metres tall and dead straight. Up near the treeline the same species grows sideways. It's flat and bent, hugging the rock, with its branches all streaming off the downwind side like a flag somebody nailed in place. Foresters call it krummholz, German for "crooked wood", which I mention mainly so you have something to say on your next hike.

Nobody looks at a krummholz spruce and decides it's a bad tree. It's the same species, grown from the same kind of seed as the ones in the valley. It just spent its whole life in wind that would snap anything growing upright, so it grew into the one shape that survives up there. Its seedlings in a sheltered valley would stand up straight. The wind did most of the work.

Most of it, though, not all of it. A spruce bent by the wind still looks like a spruce, and a different species on the same ridge bends its own way. The ridge shapes the tree, and the tree still matters.

Here's the marketing version. You hire a marketing leader who comes highly recommended. A year and a half in, brand is still an afterthought and the budget has drifted towards whatever shows up in pipeline by month three, so you part ways. You hire someone with a completely different CV. A year and a half later you're looking at the same budget shape. Same drift, different person. You replaced the tree and kept the ridge.

I've been the replacement tree. I once worked with a founder who had recently parted ways with their Head of Marketing. The reasoning made sense at the time. Activity was up and execution quality was high enough, but there was little to no impact on pipeline, and there were more things on the "planned" list than the "done" list. "Marketing strategy" was, unfortunately, on neither. It had a familiar shape, a junior with no formal marketing background tasked with delivering what took a senior a decade (and a degree) to learn. I wasn't replacing a bad marketer, just the wrong marketer for the role.

So I set about building the marketing strategy while reviewing the backlog. Good news, the backlog had the foundations of a decent marketing strategy. Great news, there was a common blocker for almost every item on it. Terrible news, the common blocker was the knowledge wrapped up inside the founder's head. I was more successful than the previous marketer, but that success came more from people dynamics than from marketing acumen. And the strategy that eventually led to significant growth looked strikingly similar to everything the former marketer had loosely collected on the backlog. The ridge, it turned out, was the founder's head. What I had that my predecessor didn't was a better working relationship with the person holding it. Same backlog, same strategy. The ridge decided the shape. How well it grew came down to the tree.

The careful version of the claim goes like this. The patterns we call "marketer bias" are not primarily cognitive. They're mostly produced by the decision environment the marketer sits in. That means the evaluation window, the runway, and who grades the work and when. The environment doesn't invent these tendencies. It amplifies ones that are already there. And the same tendency comes out looking different depending on where the leader learned the job. Put a different competent marketer on your ridge and they'll probably bend too. Just not identically.

Two leaders from my 13 thesis interviews show what that looks like in practice.

The first runs marketing as a team of two, on a budget closer to 10K a month than 100K. They told me about a round of competitor ads. The ads didn't bring in leads within about a week, so the team decided they weren't working and killed them. Looking back, they thought it had been called too fast, without checking whether the problem was the ad copy, the conversion or the application itself. On a team that small, I suspect, nobody gets a free afternoon to diagnose a failed test, so a week of silence gets read as a verdict. Later in our conversation they told me the company wasn't advertising on its own brand name and had stopped going to events, and reckoned they'd underestimated the value of brand. My read, not theirs, is that nobody chose to neglect brand. It just never had an owner. Call it the small-team wind.

The second leader is a corporate transplant, in my coding, whose most recent team was about ten people. They believe brand is key to a company's longevity, and they deliberately leave it until last. Two of their reasons stood out. One is that people outside marketing tend to switch off the moment you say the word (a reaction I suspect you've seen). The other is arithmetic. In their reckoning a CMO gets something like 18 months to prove their value, and a decent brand uplift can take something like 18 months to show. So they build the commercial engine first and bring brand in after. Different reason, same place in the queue for brand, which is last. This time the wind is the CMO's own clock.

The cohort counts point the same direction. Seven of the 13 built their careers predominantly in startups, five came in from corporates, and one worked as an advisor. When I counted how often each group talked about demand work compared with brand work, the startup-natives discussed demand 3.2 times as often as brand. The corporate transplants did it 1.8 times as often. (That's discussion frequency across 56 coded passages, which is a proxy for attention. It isn't a spending ratio, and I'll keep saying so until nobody quotes it as one.) Both groups lean towards demand, one of them noticeably harder.

Why read that as structure rather than character? Gerd Gigerenzer's work on ecological rationality argues that a rule of thumb isn't good or bad in the abstract, only good or bad for a particular environment. Killing a test after a week is a bad rule for brand. For a two-person team that, as far as I can tell, has nobody free to work out why a test failed, it's at least an understandable one. Daniel Kahneman and Gary Klein made a related argument in 2009 about when expert intuition deserves trust, and their answer was environments regular enough, with feedback quick enough, to learn from. My thesis argues that tactical marketing mostly has that kind of feedback. You run the paid social test and you know by Friday. Brand mostly doesn't, because its returns take six to eighteen months to show up. So the same leader can be well calibrated on Tuesday's test and badly calibrated on next year's positioning without being a different person. It depends on which decision is on the table.

None of this means working on the person is pointless. A 2025 review of organisational debiasing by Barbara Fasolo, Claire Heard and Irene Scopelliti sets out two routes. One tries to straighten the tree, by training the decision-maker. The other puts up a windbreak, by changing the environment the decision is made in. Which one works depends on the context, and their framework suggests the windbreak suits places where turnover is high. That makes sense to me, because a windbreak stays put when the people don't. (CMOs, by most counts, turn over faster than anyone else in the C-suite.)

In Edition 1 I described this as the part of the research I most expected to be wrong about, in part. The line I used there, that any qualified marketer in that seat would show the same patterns, works better as a question than as a claim, so here's the question. Take the last long-horizon call you deferred. The brand campaign, say, or the analyst briefing that keeps sliding into next quarter. Ask honestly whether a competent stranger in your chair, with your quota, your board date and your runway, would have made the same call. If the answer is no, that one's on you, and the good news is that you can fix it yourself. If the answer is yes, stop trying to fix the person. Write down which structural fact actually decided it. The likely suspects are the evaluation window, who grades the work and when they grade it. Change that one before the next planning cycle. If it was the window, for example, take the brand line out of the monthly pipeline review and put it in front of whoever does the annual one. My hypothesis, and it's a hypothesis rather than a finding, is that one windbreak does more than any amount of resolve.

The full pillar, including the cohort differences, what the structural claim doesn't say (it isn't absolution, and biases are real) and the theory underneath, lives at The Structural Causation Thesis on gossamergrowth.com. (About 4,100 words. It still argues with everyone, including me.)

Back on the ridge, the spruce is still growing sideways and doing a perfectly good job of it. You could fell it and plant a new one. Give it a year and a half and you'll recognise the shape. Then you get to explain krummholz to the board.