How do we get onto our ideal customers' shortlist?

How does a B2B shortlist actually form?

A B2B shortlist forms from memory first, then from colleagues and outside checks, and it is usually ranked before any vendor hears about it. An ideal customer profile tells you whom to want. It says nothing about why those people would think of you.

The sequence, as far as the evidence reaches:

1. A situation triggers the search. Something breaks, a contract ends, a new leader arrives. Romaniuk, writing for LinkedIn’s B2B Institute, argues that memory-generated brands are the starting point and other sources are usually consulted only when those fall short.
2. The group brings its history. 6sense reports buying groups averaging 11 members, with 7 to 9 earlier purchase journeys in the category. Buyers, it writes, “place about four out of the five vendors they evaluate on their shortlist from day one.”
3. Outside sources test the names. Peers, review sites and AI assistants come in here. Whether they add names or check them is the most contested point in this literature.
4. The list gets an order. In that research 94% ranked the list before contacting any seller.
5. Sellers validate, and sometimes change the outcome. 42% said their preference changed after those conversations; about one in four groups changed the final choice.

The most useful sentence in that research describes two gates: prior experience shapes “who makes it to the table rather than what breaks the tie.” The tie is broken by features, ease of implementation and price. Getting listed and getting chosen are separate problems with separate causes.

Illustration: a place on the list is offered between two people, and the buyer keeps the hand that holds the cards.

How reliable is the claim that 95% of winners were on the day-one shortlist?

The 95% figure is a real result from a vendor’s self-reported survey, but the same table gives 85% for prior years, and it proves less than its usual wording suggests. It shows that a late addition is at a disadvantage, not that the list causes the win or that the outcome is fixed on day one.

If four of five evaluated vendors are on the list from day one, as the table gives for both periods, a winner picked at random would be a day-one vendor 80% of the time. The informative gap is 80 to 95, or 80 to 85 on the earlier figure, not 0 to 95. Per seat, my arithmetic gives a day-one vendor about 24% against 5% for the late addition on the 2025 figure, roughly five times, and about 21% against 15% on the earlier one, roughly 1.4 times. The handicap is that range: real, wide, and not zero.

The answers are also retrospective, and buyers tend to remember the winner as there from the start. Confounding is likely, since vendors already known may also be stronger or safer. And the sample leans away from you: 78% of purchases reported for 2024 were renewals, replacements or enhancements. No page I opened reports results for post-seed vendors.

The numbers quoted as one finding measure different things:

Number	What it refers to	Publisher, and what it sells	Method on the page	What you may conclude
95% in 2025, 85% in prior years	Bought from one of the four day-one vendors	6sense, account and intent data	“nearly 10,000 B2B buyers” over three years. Years given, no field dates or weighting	Late entry is a handicap of 1.4 to five times per seat
94%, 84%	Ranked before contact; of those, bought from the first vendor spoken to	6sense	As above	Buyers arrive with an order in mind
77%	Bought the favorite from the end of selection	6sense	As above	Favorites usually hold, not always
54%, then 37%	AI chatbots influencing the shortlist	G2, reviews and buyer data: March, June	1,076, then 1,038 decision makers. March options sum to about 300%. Wording not printed	Assistants are present. No trend
50%	Shortlist places from peers or prior familiarity	The Reality Department, research agency	250 UK respondents, August 2026	Trust carries about half here
72%	Journey starts with peers in private groups	Wynter, B2B audience research	Marketing leaders. No sample size stated	True of marketers, at most

The 6sense figures also sit awkwardly together. 94% ranked before contact, yet the same program says buyers select a preferred vendor before engaging sales “roughly 80%” of the time, and a ranking without a favorite is hard to picture. Weighting the 84% with the roughly 50% for non-rankers gives about 82% buying from the first vendor spoken to, above the 77% who bought their favorite. Different questions or years may explain this; the pages do not say.

What holds across publishers is the direction: buyers commonly form an order before you can influence it. Most of these publishers also profit from the conclusion that you must be known early, so their agreement is weaker than it looks. The Reality Department is the most candid, describing its work as “original data built to earn coverage or generate leads.”

How do you tell if you are not invited or invited and losing?

Sort your last 20 or so decisions by where you dropped out, because being uninvited and losing an evaluation need opposite fixes. Use real cases in one segment, including deals you only heard about later. With numbers this small you are reading cases, not rates.

What you see	What it usually means	How to check	Next move
Fitting accounts buy a competitor and you hear afterwards	You do not come to mind in that situation	Ask recent buyers which names they started with, and why	Claim a buying situation
Invited late, asked only for a price, no access to anyone	You are filling a column	Ask to speak with one more person in the group	Apply the test under “not chase” below
You reach the final and lose to the known name on risk	Proof or risk gap	Ask what would have had to be true	References, security facts, a scoped pilot
You lose on features, implementation or price	Product or offer, not visibility	Compare against those reasons	Separate the GTM problem from the product problem
The evaluation ends with no purchase	The real competitor is the current way of working	Ask what they do now	Sharpen the trigger, or leave

How can a startup without a known brand get onto a shortlist?

A startup without a known brand gets onto shortlists by narrowing the field until the people who matter can know it, then accumulating evaluations and honest mentions among them.

1. Choose the accounts, and say who you are not for. Keep the list small enough to name the people. Being shortlisted by the wrong buyer costs both of you weeks.
2. Claim one buying situation. The next section shows how.
3. Earn evaluations, including ones you lose. In the 6sense data, 85% of buyers had previous experience with the vendor they bought from, and 85% of that experience was previous evaluations. Under 3% knew the vendor from brand or content alone. Among buyers of a new capability, 34% had evaluated that vendor before without buying. The source itself puts runners-up in a strong position “for future deals.”
4. Track people, not only accounts. The same page explains why a purchase is old news to the group: “people move around, but generally within an industry and function.” When you lose, ask whether you may check in when something changes, and note what would change it. My inference: this list of honest former evaluators is the cheapest shortlist asset a young company has.
5. Give fitting customers something true to repeat. Ask only those who got real value, and agree exactly what they are comfortable saying.
6. Publish a comparison you would defend on a call. Include doing nothing and building it internally, and say when a competitor is the better choice. Then give the champion an explanation a second stakeholder can follow.
7. Reach out only on evidence of a live situation. Verify the buying signal before you contact anyone, and ask permission before sending more than the first note.

Once this works in one segment, it belongs in a repeatable GTM motion.

How do you find the buying situations that put you on a list?

Buying situations, which researchers call category entry points, are the moments that make a buyer think of a vendor at all, and you collect them by asking recent buyers what happened. Romaniuk defines them as “the cues that category buyers use to access their memories when faced with a buying situation.”

A persona describes a buyer from outside: role, company size, goals. A place on a list comes from knowing the moment from inside, in the buyer’s words: what happened that week, who else was at the table, what they feared getting wrong. Romaniuk’s prompts help: when, where, while what else, with what, why, how feeling, with or for whom. Her B2B addition is to ask from two angles, the business’s objective and the person’s own stake, up to “even just not getting fired.”

She then filters on credibility, competitiveness and commonality, ending with five to eight situations and one per message. With a team of two, I would start with one and earn the second. One record per conversation:

The situation, in the buyer's words:
What happened the week the search began:
Who else was at the table, and what each feared:
Names that came to mind before any search:
Can we deliver this now (yes with proof / yes without proof / no):
Who already owns this situation in the buyer's mind:
How often it comes up in our segment:
Decision (claim / watch / leave):

The paper’s main test covers 17 US business insurance products and links more entry points to lower stated likelihood of switching in 16 of them. That measures keeping a customer by stated intent, not getting a stranger’s vendor onto a list; the paper cites older work for selection and for actual switching, which I did not read. The publisher, LinkedIn, sells reach. Use the method to know what to look for, not as proof that it lifts a startup.

Who in a buying group adds a vendor, and who removes one later?

No source I could open measures who adds a vendor to a B2B shortlist or who removes one, so the table below is a working model and the record under it tests it against your own losses.

Reality reports that 74% of purchases went through three or more rounds of internal scrutiny after the shortlist, and 57% had security, legal or procurement join after the demo or only at signing. G2 reports finance involvement rising from 31% to 46% in a year. My reading: you are often removed by someone who arrived after your champion made the case.

Role	How they shape the list	How they remove you	What they need from you
Champion or user	Names options from experience and peers	Cannot explain you to colleagues	A short explanation with limits stated
Decision-maker	Adds names trusted from earlier jobs	Sees career risk in an unknown name	A reference at their level
Finance	Rarely adds, questions the case	Assumptions do not hold	The calculation, assumptions visible
Security, legal, procurement	Rarely adds, can veto	A gap found at the contract	Those facts in writing, early

Fill in one record per lost evaluation:

Role that raised the objection:
Round they came in (list forming / demo / after the demo / signing):
What they asked, in their words:
What was missing:
Fixable before the next round (yes, how / no, why not):

An illustration, not a real case: security, came in after the demo, asked where customer data is stored, no written answer, fixable with one page sent earlier.

When the same role keeps arriving in the same round with an answerable question, offer that answer at the start. As Reality puts it: “Bringing them in earlier doesn’t add friction.” A gap you could not have closed is a product or fit problem.

How much do peers and private communities shape a shortlist?

Peers clearly shape B2B shortlists, but the size of the effect depends on who you ask, and nobody outside these private conversations can observe them. Three surveys give three answers that cannot be averaged.

Wynter says 72% of marketing leaders start by asking peers in private groups and only 9% by googling a software category. Reality attributes half of shortlist places to peers or familiarity combined. G2’s March chart puts peers and colleagues at 33%.

You cannot target private channels, so you can only give people something true to say. Wynter draws the right boundary: “Do not disguise promotion as peer advice.” Contribute under your own name in one community your buyers really use, ask customers before naming them, and stay quiet when you have nothing to add.

Do review sites and analyst lists matter for a young company?

Review sites and analyst lists matter once buyers in your segment use them, and the published evidence is too interested to settle it. G2 reports review sites at 38% as a shortlist influence, and that review citations top what makes buyers confident in an AI answer. DerivateX, an SEO and GEO agency, found review aggregators at 0.9% of ChatGPT citations, with G2 and Capterra at zero across 233 recommendations.

Analyst input ranked near the bottom of the reasons 6sense’s buyers gave for choosing a winner. That concerns breaking the tie, so it does not rule out a role in getting listed.

My decision rule: claim a review profile when buyers name that site unprompted and your category exists there. Read an analyst report’s inclusion criteria before spending any time, and only when enterprise buyers ask for it.

Do AI assistants such as ChatGPT build the shortlist now?

AI assistants are now part of many software buying journeys, but the evidence splits on whether they build the shortlist or check one that already exists.

• They build it. G2’s March survey has 51% starting with a chatbot more often than Google, and 33% buying from a vendor they had never heard of. That is hard to square with 6sense’s under 3%. G2 also discloses that generative models helped design the survey and analyze results.
• They barely matter. 6sense writes: “Our statistical models attribute less than 2% of variability in these outcomes to LLM usage.” The models are not published.
• They check. Reality found 77% used AI in vendor research, most often to fact-check a vendor already found, and calls AI “the checking layer on that trust, not a substitute for it.”

My reading: assistants are one more place where existing reputation and third-party text resurface, and for some buyers a place where a new name appears. DerivateX ran 40 buyer prompts ten times each in ChatGPT, and found the recommended tool’s own site cited only 11.6% of the time. It separates “Named from memory” from “Cited from search,” which helps a challenger: “A challenger brand cannot shortcut its way into the model’s memory,” but it can appear in what is retrieved at answer time.

Its finding that every cited page used lists, and 68% a comparison table, comes with no comparison group. “Best tools” prompts retrieve “best tools” pages, so this may describe the prompt more than a winning format. Copy the honesty and leave the formula. A comparison with named limits is something an assistant can read and a person can forward, which is why it helps to publish the decision facts an agent needs.

The cheap test: write ten prompts from real buying situations, in buyers’ words, and run them monthly in the assistants your buyers name.

How do you stay known when most accounts are not buying?

Most accounts are not buying at a given moment, so what reaches them has to be useful without urgency, and you can calculate the size of that waiting room.

The 95-5 rule is often quoted as a law. Its author, John Dawes, writes: “The 95% figure is not meant to be a precise rule.” It comes from an example: companies changing a principal bank or law firm around once every five years gives 20% a year and something like 5% a quarter. 6sense, by contrast, says roughly 60% of a well-defined ICP is not in-market at any given time. Dawes shows his time window; the 6sense page gives no method.

So compute yours. Dawes suggests asking buyers how often they purchase. Divide your named accounts by the cycle in quarters: 120 accounts on a three-year cycle means about ten per quarter, an illustration and not a benchmark. That is the ceiling on what any campaign can do this quarter. A new category has no replacement cycle, so watch your interview triggers instead.

Dawes also warns that mental availability is a multi-year task. His remedy is broad advertising, which a post-seed company usually cannot fund. My adaptation is to shrink the denominator: a useful note, a finding or a small tool, sent to the few hundred people who matter, and no chasing when they do not answer. More in building demand before the brand is known.

Which shortlists should you not chase?

Do not chase a shortlist when being on it costs both sides time. No study I found tests this for startups, so this is judgment, not measurement.

Four questions before you accept an invitation:

1. Can we talk to someone besides the person collecting quotes?
2. Do the requirements describe the buyer’s problem, or one competitor’s feature list?
3. Would this buyer be glad they chose us a year from now?
4. Can we meet the requirements today, without claiming what we plan to build?

Two or more honest noes, and I would decline. Do it in writing, say why, and name who fits better if you know. That costs you a slot you were unlikely to win and leaves a person with an accurate memory of you.

How do you measure whether you are on the shortlist?

You cannot see a shortlist directly, so measure it through what buyers tell you and the invitations you receive. 6sense claims vendors see only 10% to 15% of research activity and fewer than 30% of eventual buyers filled out a form on the winner’s site.

Measure	How to record it	Its limit
Invitations	Evaluations in your segment you were asked into, per quarter	Small numbers, long cycles
Named at the start	In win and loss calls: “Which names did you start with, and when did we come in?”	Memory bends toward the outcome
Origin in their words	Free text: “How did you first hear about us?”	People forget and simplify
Former evaluators	Who assessed you, when, and what would reopen it	Needs permission and upkeep
Assistant log	Your ten prompts, monthly: named, described, cited	Answers vary between runs

Wynter’s advice fits: ask in their own words and do not force a private recommendation into a last-touch model.

How I’d approach this

I would start with your last 20 decisions and five buyer conversations. A founder or first GTM hire can do most of this page without outside help.

I am Piet Baudoin (Poldermarketing, remote, Dutch and English). I work on positioning and messaging and findability in Google and AI answers, and I also build and run AI automations for clients, with care: lead follow-up, enrichment and qualification, CRM setup, reporting, content distribution and tool integrations. My focus is post-seed startups building go-to-market: the first customer, the second, then a repeatable process that brings the product to the customer who is genuinely happy with it. Google Ads and Meta Ads are relatively new to me. The free growth scan is a reasonable first look at what a buyer finds when they check your name.