The founder is the product.

§I A specific claim.

The claim in this essay is narrow and I want to state it precisely. The founder of a consumer brand is the only renewable source of taste the brand possesses for roughly its first four years. After year four, if the brand has been run correctly, the organisation begins to produce taste independently of the founder. Before year four, it does not. Every attempt to pretend otherwise, during that window, costs the brand expensively and silently.

This is not a statement about founder personality, founder celebrity, or founder presence in the brand's marketing. Those are separate and largely irrelevant questions. This is a statement about where the consistent, specific judgement that the brand ships on lives in the organisation. In year one through four it lives in the founder. There is no exception to this pattern in our file.

The piece is in two parts. Part one: what the claim means, operationally, day to day. Part two: how to run the edit function such that the claim does not collapse into founder-as-bottleneck or founder-as-brand-face.

§II Taste as renewable input.

Taste, in the sense we mean, is not a one-time decision. It is a continuous input. Every week, the brand makes fifty to a hundred taste decisions: the image on the homepage, the word in the subject line, the order of the range in a retailer's window, the material of the gift box, the paragraph in the team newsletter, the photograph on the back label. Each decision is small. Together they are the brand.

A brand's aggregate taste, per quarter, is the sum of these fifty-to-a-hundred weekly decisions. If each one is made by a different person, the aggregate is a committee-produced taste — which is to say, an averaged, inconsistent, drifting taste. If each one is informed by the same editing function, the aggregate is a coherent, specific, recognisable taste.

In years one through four, the only person capable of being the editing function consistently is the founder, because there is no accumulated institutional taste to refer back to. The organisation's taste is whatever the founder edited last week. In year five and beyond, the organisation begins to carry its own reference set, and the founder can delegate more. This is what we mean by the four-year window.

§III The error of delegating early.

The commonest error we see, between year two and year four, is the founder delegating the edit function. Usually to a CMO or a creative director. The delegation seems responsible. It frees the founder's time; it adds category experience; it signals a maturing company. It also stops the brand from producing taste.

What happens is predictable. The delegated edit function makes decisions against the taste reference set the delegatee brings with them — often a former employer's taste, often a 2016-era playbook. The brand begins to drift towards the reference set of someone who did not build it. The drift is imperceptible in any single decision. Over twelve months it is visible.

We have watched this cost repeat rates six to fourteen percentage points, cost retail performance one-to-two category positions, cost organic social performance twenty-five to forty percent. None of these losses were attributed to the delegation at the time. They were attributed to market conditions, to algorithm changes, to seasonality. They were the delegation.

You cannot delegate the editing function until the brand has a codified point of view. Until then, the founder is the point of view. — Saoirse Hale, in conversation

§IV How to run the edit without becoming a bottleneck.

The anxiety most founders have about staying in the edit function is that they will become the company's bottleneck. This is a legitimate concern and it has a practical answer. The founder does not approve work. The founder edits it. There is a meaningful operational difference between the two.

An approver sits at the end of a pipeline, saying yes or no. A queue builds behind them. An editor reads work in batches, with a notebook, and returns it with notes. The queue does not build because the editor never holds the work; the work flows back to the team, with the editor's thinking attached, and the team executes the next version without needing to book another approval window.

Operationally, we ask founders to commit to three editing windows a week, each of around ninety minutes, during which all brand-facing work is read in batch and returned in batch. Everything else is delegated. The delegation is to a brand director whose job is to ship on the founder's notes — not to generate taste of their own. This structure preserves the edit function and removes the bottleneck. We have watched it work in sixteen of the seventeen founder-led companies we have run it through.

§V The year-four transition.

In year four, if the structure above has been maintained, the organisation's accumulated taste reference set becomes large enough, and specific enough, that other people can begin making taste decisions that remain consistent with the founder's judgement. This is the point at which the founder can begin to delegate edit — not before.

The transition is gradual. First, colour decisions migrate. Then copy. Then visual identity execution. The founder remains the final editor on the annual plan, on the brand thesis, on the seasons, on the second product. Those four stay with the founder for another two to three years.

By year seven, a well-run consumer brand can survive its founder stepping back from day-to-day editing, because the organisation has produced its own editors, trained on the founder's notes, consistent with the founder's taste. Attempts to compress this timeline — to let the organisation edit itself in year three — have, in our file, failed reliably enough that we treat the four-year figure as a floor, not a target.