The economics of taste.

§I Taste is a claim about the future.

The word taste is misused so consistently inside brand work that it has become almost useless. Taste is not aesthetic preference. It is not “what I like.” Taste is the act of choosing, in advance of evidence, the detail that will matter later. It is a predictive faculty. The businesses with it see things roughly nine to eighteen months before anyone else, and they buy the space cheaply. The businesses without it pay list price for the insight after it has been retail-priced by a competitor.

Every taste judgement is really a claim: this detail will matter, and most of the people around me do not yet see that it will. When that claim is right, the cost of execution was low because no one else was bidding; when the claim is proven, the brand is already encoded in the culture at a price it could not pay again. This is what taste is on a balance sheet. It is a form of early capital allocation, at better multiples than a money manager can access.¹

The mistake most founders make is treating taste as the department of the creative director. They pay a salary, they approve a moodboard, they go back to looking at the dashboard. What they have bought is not taste; they have bought a hedge against their own reluctance to exercise it. Taste that matters on the balance sheet is always made by the person with profit-and-loss responsibility. The creative director proposes. The founder edits. If the editing function is absent — if the founder is too polite, too distracted, or too keen to be liked by the team — the taste fails at the final gate, because the hundred small decisions that matter are made by the person who least wants to be wrong.

A company with taste is a company whose founder is still willing to have an unpopular opinion in a thursday meeting, in year nine. — field note, client conversation

§II The four quadrants of taste.

We find it useful, in diagnosis, to split taste along two axes. The vertical axis is conviction: how willing is the team to back a judgement that has not yet been validated. The horizontal axis is specificity: how precise is the judgement when it is made. The two axes produce four quadrants, and each quadrant produces a different kind of brand.

Our argument is short: only the top-right quadrant produces a brand that compounds. Every other quadrant is an expense. The Vibes Brand pays for the rebrand. The Committee Brand pays for a decade of category share loss. The Researched Brand pays for research that it then overrides at the moment it matters. The Taste-Made Brand pays a specific, usually unpopular, price at one key moment and then benefits from that purchase, quietly, for twelve years.

§III Why the dashboard erodes it.

There is a modern problem that did not exist thirty years ago: the real-time performance dashboard. Every decision, in a dashboard-led business, asks to be justified by a number that can only be collected after the decision has been made. This sounds rigorous. It is in fact a near-complete inversion of how taste operates. Taste makes the judgement first and tolerates the period of uncertainty that precedes proof. The dashboard cannot.

What the dashboard does well: surface problems with existing products, diagnose velocity decline, track the performance of promotional activity, discipline marketing spend.² What the dashboard does poorly: decide what a bottle should feel like in the hand at the moment of repurchase; decide whether the word on the back label is the right word; decide whether the taste in the category has moved. These are all taste decisions. They cannot be pre-validated. If you wait for the dashboard to confirm them, they have already been made, badly, by someone braver in a competitor’s office.

We find, in our own practice, that the brands that compound are run by founders who hold a firm line about which decisions are dashboard decisions and which are taste decisions, and who refuse to mix them. The promotional calendar is a dashboard decision. The new pack is not. The gondola end is a dashboard decision. The voice on the label is not. A brand that treats everything as a taste decision becomes self-indulgent. A brand that treats everything as a dashboard decision becomes a commodity. The judgment sits in knowing which is which.

The dashboard is a useful instrument and a terrible boss. — Lena Osei, internal memo

§IV The founder as final editor.

The analogy that keeps returning in our work is to a magazine. A magazine has a publisher, a commercial director, a managing editor, and columns of specialist editors. What gives the magazine its character is none of those — it is the final editor, the single person who says this belongs in the magazine and this does not, issue after issue, year after year. Remove the final editor and the magazine is technically still produced, but it stops having a point of view. Within six issues the reader feels the absence, even though they cannot name it.

A consumer brand is the same shape. Its final editor is the founder. If the founder delegates the final editing function — to a CMO, to a brand director, to a committee — the brand continues to ship, but it stops having a point of view. Within six quarters the shopper feels the absence, even though they cannot name it. Six quarters is when the repeat rate starts to quietly slide. Four years is when the category share numbers show it.

The single most common reason founders stop exercising the final editing function is not time. It is fatigue from being told, for years, that their taste is “subjective.” No — their taste is the asset. The shareholder case for the founder’s continued editing involvement is the shareholder case for the brand continuing to have a reason to exist. Everything else — the dashboard, the range, the team — is executed around that editing.

§V What disciplined taste costs and pays.

It is possible to approximate the price of taste, because over thirty years the taste-made brands have left financial records. Broadly, a brand run with high specificity and high conviction outperforms category on gross margin by 6–11 percentage points over a decade, on unit pricing by 15–28%, and on retailer-held shelf space by 1.4× to 1.8× fair share.³ These are not marketing numbers. They are balance-sheet numbers.

The visible cost of taste is small: the salary of a person or two who can see things before they arrive, the willingness to decline projects the committee likes, the patience to not promote a range extension that would add 2% to revenue and 9% to brand entropy. The invisible cost is larger: it is the weekly discomfort of disagreeing with likeable people about details the dashboard cannot arbitrate.

The payment, when the call is right, is a brand that does not need to be rebranded for a very long time. A logo that lasts nineteen years. A bottle shape the category copies. A word on a back label that a competitor spends three years trying to dislodge. A category that, eventually, orbits around you. The economic term for this is a moat. The accurate term is taste, held quietly, across many quiet years, by a founder who was willing to edit.