Chapter 2

Marketing's one job

Marketing has one job: to get the brand selected. Not seen, not liked, not followed. Selected.

Everything else either makes that more likely or guesses at whether it happened. So ask one question of every tool, channel and campaign: will this get us selected?

The "so that" test

Take any marketing activity and keep asking "so that what?" We post on LinkedIn, so that people in our market see us. So that they remember us. So that we're on the list when the need arises. So that they choose us.

Try it on a trade stand, a price cut or a new logo. Every chain, long or short, should end on the same sentence. One that ends anywhere else has no job to do.

Most of the work happens before anyone is buying

The Ehrenberg-Bass Institute, whose body of research is the most tested in marketing, has long argued that brands grow by being easy to think of and easy to buy. Professor John Dawes estimates that up to 95% of business buyers aren't in the market for a given product at any one time (Ehrenberg-Bass). He's clear that the figure is a rule of thumb that varies by category, not a precise count.

So most marketing reaches people who won't buy for months or years. It works by leaving a memory that surfaces when they finally do. A 2025 study fits: in large business purchases, buyers wrote their shortlist on day one, mostly from memory.

The stand-ins

A choice made from memory, months later, leaves nothing to count at the time. So marketing counted what it could see.

Sales, market share and win rate are missing from that table on purpose. They aren't stand-ins. They're the outcome, counted after the fact.

What nobody could see was the odds beforehand: at this moment, with this buyer, how likely were we to be picked? Every row above is a "might", a way of estimating those odds. Being seen matters, because nobody chooses a brand they've never met. But it's the first step, and it has long been treated as the whole job.

What a choice is

A choice has four parts: one or more buyers, a moment of need, a set of alternatives and a pick. The alternatives always include doing nothing.

Sometimes the alternative isn't another firm at all. In 2023 a major airport asked us to quote for a digital gift finder, a short quiz to help shoppers choose a present in time for Christmas. It came through a warm referral, and we had a quote back within a fortnight of the brief. About a month later, after a wait on a key person's leave, the answer came: they'd bought an off-the-shelf product instead. We hadn't lost to another agency. We'd lost to a different kind of answer, one that was never on our list of competitors.

Two things follow. First, the job is bigger than promotion. Of the old marketing mix, the 4Ps of product, price, place and promotion, only one is promotion. What you sell, what it costs and where it can be bought decide choices as surely as any campaign.

Second, the job doesn't end at the first sale. A renewal is a choice. So is a recommendation to a friend.

Selectability

Selected is the outcome. Selectability is the odds of it.

I define Selectability as the odds that a buyer, at a given moment, picks you over the alternatives, doing nothing included. A moment is one kind of buyer at one step of their journey.

Its counterpart is visibility: the odds that a buyer, or a machine, notices you. A brand can be seen everywhere and selected almost nowhere.

Selectability is also specific. You can be highly selectable to one buyer at one moment, and barely selectable to another. One number for the whole brand hides that.

People have tried to estimate these odds for decades, through brand-equity models such as Kantar's Meaningful Different framework, which predicts a brand's share from surveys of how buyers see it (Kantar), conjoint studies and win-loss reviews. Those average a market, stage a choice or look back at one deal. What they lack is the moment. What's new is that AI assistants now advise buyers, and can be asked those buyers' questions directly, as often as you like. Part five turns that into the Selectability Score.

The Selectability Score is itself a stand-in. It samples what buyers and machines say; it doesn't observe the sale. It's better because it measures the choice at the moment, not exposure, and it comes with honesty rules. Your own sales data always outranks it.

Not only marketing's job

Read the first sentence of this chapter again with "sales" in place of "marketing". It still holds. So the job is simple, and the best research has said so for years. How, then, did we build an industry where every team hits its number while the brand loses the choice?

Whose evidence this is. The Ehrenberg-Bass Institute sells research to sponsor brands. Kantar sells brand tracking, including the framework named here. There's no vendor survey data in this chapter.