Chapter 5

The choice starts without you

Being seen feels like winning. Your brand appears in the search results, in the AI answer, on the shortlist. The report turns green.

But by the time a buyer speaks to you, the choice is already under way. The enquiry that marketing celebrates and sales chases comes late in the decision, not at its start.

Go back to the operations director choosing scheduling software. Before she emails a single supplier she has four names in her head. Two she has used, one she evaluated years ago, and one a colleague swears by. Her team spends months reading and comparing. Then she calls the supplier they already prefer.

She's an illustration, but the research below measures the pattern. The choice starts without you, and it isn't finished when you arrive.

The shortlist is written on day one

In 2025, 6sense surveyed about 4,000 people who buy for their organisations (6sense). Every purchase was worth at least US$25,000. It found:

• Buyers evaluated about five suppliers each time, and chose one.
• The winning supplier was on the buyer's first-day shortlist 95% of the time.
• 94% of buying groups had ranked that shortlist before speaking to any seller.
• The supplier they favoured before first contact won about four deals in five.
• Buyers made the first approach themselves in 79% of cases, roughly 60% of the way through.

Buyers had dealt before with almost four of the five suppliers they evaluated. In this kind of purchase the shortlist is remembered, not discovered.

One vendor's survey is thin ground, so here's a second. Bain and Google asked 1,208 US business buyers in 2022. Some 86% began with a list in mind, and 92% bought from it (Google, via Marketing Dive).

But the list is not locked

That evidence describes one kind of purchase: large, experienced and decided by a group. Most buying isn't like that.

Where one or two people make a considered purchase, the early list has a strong head start but a real second window. McKinsey's consumer decision journey research, published in 2009, studied almost 20,000 consumers. A brand in the buyer's initial set was up to three times more likely to be bought. Yet car shoppers in that study started with 3.8 brands and added 2.2 while evaluating (McKinsey). The data is old, but it's still the best-known measure of that second window.

Google's "messy middle" research, in a simulated shopping test, found a fictional challenger brand took 28% of preference from an established favourite when it carried strong reviews and benefits (Think with Google).

In habitual buying there's barely a shortlist at all. The choice is near-automatic, made from whatever comes to mind first and is easiest to buy.

Small purchases follow the same shape, only faster. Take a café owner looking for a bookkeeper, an invented example this book comes back to. He asks two other owners and an assistant, has three names by lunchtime, and calls one that afternoon. His list still forms before any supplier knows he's looking. It just forms in hours, and one good answer can join it. I found no study that measures this for small services, so treat it as illustration.

Even the largest purchases stay open. In the 6sense study, about one buyer in five changed their mind after talking to sellers. So the early list matters everywhere, and nowhere is it final.

What AI has changed

The shortlist habit is older than AI. What AI is changing is who writes the list.

The list is moving from the buyer's memory to the machine. G2 surveyed 1,038 business software buyers in June 2026. More than 80% had taken software recommendations from an AI chatbot in the past two years (G2).

The list still decides. Among buyers who took recommendations from AI, 80% bought from their first shortlist in at least three of their last five purchases. Among those who didn't, 65% did.

That's an association, not proof, but it suggests AI hasn't loosened the day-one list, and may have started to write it. The largest deals lag: in 6sense's study buyers used AI mostly mid-journey, to compare suppliers already listed.

A list written by a machine can also be rewritten by evidence. In one experiment, AI models dropped a famous brand when a rival offered better specifications (Chapter 13).

AI has moved the choosing further out of sight, and made counting who saw you an even weaker guide to who chose you.

What that adds up to

You're invited in late. By then you're confirming a preference or overturning one. The choice can still be won, but mostly before you're in the room. That raises the question of where. No buyer comes to know you, believe you or prefer you in general. They do it at a particular moment, with a particular need.

Whose evidence this is. 6sense sells intent-data and account-based marketing software. G2 runs a software review marketplace. Google sells advertising; it partnered on the Bain study and wrote the messy-middle research, and Marketing Dive ran the Bain figures as a Google-sponsored post. Bain and McKinsey sell consulting. There's no peer-reviewed source in this chapter.