Sample frameDay 352·Growth · Compound

Morning. Your last six tests all won a little. That isn't the good news it looks like.

Growth.

The Local Maximum

5 min read·Apply by lunch

The question

Line up the last dozen experiments you've run, write the size of each result beside it in date order, and look at where the line is going. If the wins are getting smaller and the flat results more frequent, you're not running out of ideas — you're standing near the top of the hill you happen to be on, and nothing currently on the roadmap is capable of finding you a taller one.

The idea

James March had been at Stanford since 1970 and had spent his working life on how organisations decide and learn — Organizations with Herbert Simon in 1958, A Behavioral Theory of the Firm with Richard Cyert in 1963. In 1991 he published a short paper in Organization Science, Exploration and Exploitation in Organizational Learning, that names what your test roadmap is quietly doing. He split adaptive behaviour into two kinds. Exploitation is refinement, selection, efficiency, execution: improving what you already have. Exploration is search, variation, risk-taking, play, discovery. His point wasn't that one beats the other; it's that they compete for the same resources and the competition is rigged. Returns from exploitation are nearer in time, more certain, and land closer to whoever produced them, while returns from exploration are distant, unreliable and frequently collected by somebody else. So a system left to optimise itself drifts towards refinement and becomes, in March's phrase, effective in the short run and self-destructive in the long run. Day 85 made this a portfolio argument about capital across whole businesses; this morning it runs one level down, across the experiment slots on a single surface. The trap is that hill-climbing works. Every small step really does improve the number, right up until the only direction left is down.

The framework

01
Plot the win curve before you plan the next quarter.
Take every experiment from the past year, sort them by date and plot the size of each result. You're not after the average; you're after the trend and the spread. A healthy programme shows the occasional large win scattered among the small ones. A programme at its local maximum shows a tidy shrinking sequence — three per cent, two, one, flat, flat — alongside a team quietly redefining success downwards to match. Read the curve before you write next quarter's roadmap, because last quarter's roadmap is what produced it.
02
Is this test changing the page, or changing the premise?
Sort every test into one of two piles. Refinements change the execution of something you've already decided: the copy, the layout, the sequence, the packaging of the offer. Relocations change the decision itself — a different customer, a different pricing unit, a different acquisition motion, a different first session. Most roadmaps come out close to a hundred per cent refinement, and no team has ever discovered that without doing the sort, because each individual test looked entirely sensible on the day it was proposed.
03
Every step up the hill was a real improvement.
Nothing on that roadmap was a mistake. Each test raised the number it was pointed at, each was proposed for a defensible reason, and the programme did exactly what it was built to do, which was climb. That's why auditing the individual decisions finds nothing wrong. What repays auditing is the generator: who is allowed to propose an experiment, which surfaces they're briefed to touch, and which parts of the business have simply never appeared on the list. A roadmap made entirely of page-level tests wasn't chosen so much as commissioned.
04
The swings your traffic can settle.
Yesterday's arithmetic decides which relocations are worth booking at all. A change to a button has to be enormous to clear the noise on ordinary traffic, while a change to who you sell to, or to what you charge for, moves the number far enough that a modest sample can settle it. So the swings you keep deferring for being too disruptive are frequently the only ones your data could ever resolve, and the refinements you keep running are the ones it never could. Judge a relocation on whether it found a hill you weren't standing on; most won't, which is what makes the ones that do worth the slot.

Members only · 4 principles + template + AI mentor

364 more frameworks are waiting.

The Local Maximum is Day 1 of 365. One framework every morning for a year — across strategy, sales, negotiation, leadership, and more.

$1 /day

Billed $365/yr · cancel any time

365 frameworks

12 topics · templates · AI mentor

Reserve your place — $365/yr →

One payment a year · the whole library · no auto-renewal.

More Growth frameworks

All 24 growth frameworks →