The winning vendor was already on the buyer's day-one shortlist 95% of the time.
That figure comes from 6sense's 2025 B2B Buyer Experience Report, a vendor-run survey of nearly 4,000 B2B buyers, which also found that four in five deals go to the vendor the buyer favoured before first contact, and that first contact with any seller now happens about 61% of the way through the journey. Read those three together and the shape of the problem is clear. The B2B shortlist is made before anyone visits your website, and by the time a buyer does arrive on your site, they are mostly confirming a choice.
That is why a consultancy that grew for years on content and inbound can see the requests stop. A founder of a small customer-data consultancy described exactly that on a forum this year: steady growth from inbound, then nothing notable for months, and the question of whether buyers now do the research themselves with AI. The top reply came from a corporate strategy lead who said they were cutting most of next year's consultant budget and would hire only for niche needs. Both halves of that exchange are the same event seen from either side of the table.
The website did not get worse. The place where the shortlist is made moved. Here is where it is made now, a worked example of one buyer and one seller, and a check you can run on your own firm this afternoon.
Where the shortlist forms now: peer asks, AI answers, the founder's feed
Before a buyer with a complex problem visits any vendor's site, three things usually happen.
They ask a peer. A message to someone in a similar role at another firm, a question in a private community, a conversation at an event. The answer is two or three names. Those names are the shortlist, and the website's job from that moment is to not lose the place.
They ask an AI assistant. The same question typed into whichever assistant the buyer uses, often several times in different forms. The answer is a synthesis of what is written and cited about the category, and it names firms. The 6sense survey found nearly 90% of buyers reporting that AI features were part of the solutions they bought, which tells you how normal the tools have become in the buying process itself. An in-house global account-based marketing director voiced scepticism on LinkedIn about how much visibility in these assistants matters; the honest position is that it is not yet measurable with any precision, and that a firm which is never named in the answers is relying entirely on the peer ask.
They read the founder's feed. When a name comes up, the buyer looks at the people, and for a services firm the person is the founder. Ten posts tell them whether the firm thinks the way they do. Edelman and LinkedIn's 2024 B2B Thought Leadership Impact Study of 3,500 decision-makers found 73% consider an organisation's thought leadership a more trustworthy basis for judging its competence than its marketing materials, and 75% say a piece of thought leadership led them to research something they were not considering.
None of the three is the website. All three decide who the website gets to talk to. Our map of the complex B2B buying journey shows where in the stages these moments sit; this article is about what to do at them.
Peers, assistants and the founder's feed hold the shortlist; the website holds the place.
Jenny wants to be found in AI search; James has to be found by Jenny
Take a worked example with two people in it. They are not clients. They are a buyer and a seller drawn so that the problem has faces.
Jenny is the CMO of a software company. Her board has noticed that buyers ask AI assistants about the category and her company is rarely in the answer. She wants to show up in AI search. She has a budget, a nine-month approval process, and a shortlist she has not written down yet but could name in ten seconds if asked.
James runs a growth marketing firm that builds the kind of SEO and content strategy Jenny needs. To win the work he has to do three things in order: get in front of Jenny before she writes the shortlist, earn enough trust to be the favourite on it, and convert her when she goes in market. He will be in a room with her for perhaps 61% of the way through, at best.
To get in front
James cannot buy his way onto Jenny's day-one shortlist with an advert she scrolls past. He can be seen building something she is interested in. So he posts about three things, and only these three:
- How the company is growing. What his firm tried this quarter, what it cost, what came of it. Jenny runs a growth function; a seller who shows his own growth working in public is a seller she can evaluate without a call.
- How they are doing marketing. The actual mechanics of his firm's own SEO and content: which pages they built, which queries they chased, how they decided. Jenny wants to know if James can do this; the most credible evidence is James doing it to his own firm where she can see.
- What James learns building the business. The mistakes, the thing that surprised him, the client conversation that changed his mind. This is what makes a stranger root for a firm.
To earn trust
Being seen is not being trusted. Jenny trusts two things: a framework that explains her problem better than she could, and a result that shows the framework worked somewhere.
So James publishes the framework he actually uses to decide what a software company should be found for in AI answers, with its steps and its limits. Then he publishes results, in the only form that survives scrutiny: period, what was done, what changed, measured where, and what cannot be claimed. On this site, the form of words is that Actualyse case studies report a range of outcomes, with the source of each stated; James uses the equivalent for his own firm.
To convert
Jenny goes in market in month seven. She asks two peers and gets three names. She asks an assistant and gets a list that overlaps. She reads James's last ten posts and recognises her own problem in four of them. She visits his site to confirm the shortlist, and the site is built for exactly that visit, with the framework, the results and a way to talk to James without a form that asks twelve questions. Our guide to designing a website for a long sales cycle covers that build.
The cadence underneath it all: James posts three times a week, and moves to five when the measures start moving. He does that for two quarters before Jenny's name appears in his CRM.
James earns the shortlist in public, months before Jenny writes it.
What James publishes, and what the website is for
Pull the example apart and it is a division of labour.
The founder's feed does the getting-in-front and most of the trust. The website does the confirming and the converting. Search, including AI search, is the connective tissue: the thing that lets Jenny's question find James's framework when she asks an assistant, and lets her find the case study when she asks a peer. Our article on B2B SEO when buyers research for months covers how the site's content has to be shaped for that long, intermittent research.
An illustrative count makes the point. Suppose James's account list holds 400 software companies with a CMO who could be Jenny.
In market this quarter (95:5 rule of thumb): 400 × 5% = 20 Already holding a day-one favourite (6sense): most of the 20 Reachable by cold outreach this quarter with a live need and no favourite: perhaps 1 or 2
That is why cold outreach to the 400 produces so little, and why two quarters of posting to the same 400 produces a different result. The posts reach the 380 who are not in market yet, which is the only time a stranger can become a favourite. The 95:5 figure, reported by Marketing Week in 2021 from Professor John Dawes of the Ehrenberg-Bass Institute and the LinkedIn B2B Institute, is a rule of thumb rather than a measurement; the shape is what matters.
The firm that wants help building that SEO and content strategy, so that its framework is what the assistant finds, is asking a commercial question that belongs on the SEO service page rather than here. The wider programme that runs the feed, the search layer and the capture together is the growth marketing service. The case studies show the kinds of firms each has suited.
The feed earns the place; search connects the question to it; the site confirms it.
If your website and campaign evidence cannot explain where committee-led deals slow down, map the buying journey with Actualyse — book a call
A three-step check to run on your own firm
You can test where your own firm stands on the shortlist in about an hour, with no tools.
- Ask an assistant your buyer's question. Not your firm's name. The question Jenny would type: "which firms help a software company get found in AI search" in your category's words. Run it three ways in two assistants. Write down every firm named. If yours is absent from all six, your shortlist presence depends entirely on peers.
- Ask your last three clients where they first heard of you. Not where they found the website. Where the name first came up. The answers are your real top-of-funnel channels, and they are almost never "searched and found your site".
- Read the founder's last ten posts as a stranger. Open the profile logged out, or ask someone outside the firm to do it. Would a Jenny recognise her problem in four of the ten? Would she learn how your firm thinks? If the ten posts are company news and reposts, there is no feed to earn the shortlist with.
Score it simply. Named in assistants, named by clients as first touch, and a feed a stranger would learn from: three out of three means the shortlist is being earned. One out of three, usually the clients, means the firm is living on referrals it does not control, which is the consultancy in the opening story a year before the requests stopped.
The honest limit of the check is that assistant answers vary by day, phrasing and user, so step one is a snapshot and never a measurement. Run it monthly and watch the direction, not the count.
The falsifiable claim: a firm that is absent from the assistant answers and whose founder's last ten posts contain no framework will see first-contact enquiries arrive with a shortlist already written, and will win them at a lower rate than the referral enquiries. If that firm's cold inbound converts as well as its referrals, the claim is wrong.
One hour, three questions, and you know where the shortlist is being made without you.
What does not put you on the shortlist
Three things get funded when inbound slows, and none of them reaches the moment the shortlist is written.
More website pages. The site is visited to confirm. Thirty new pages confirm harder. They do not get the firm asked about by a peer or named by an assistant unless they carry the framework and the results in a form other people cite.
Cold outreach volume. It reaches the 95% who are out of market and the few who are in market and already have a favourite. The arithmetic above is the whole argument.
A rebrand. A new identity on a firm nobody is naming is a new identity nobody is naming. Brand work earns its cost after the firm has something to be known for.
The pillar on why B2B sales cycles are getting longer covers the wider slowdown that a B2B marketer of about ten years described on a forum this year as leads falling at every company they knew. Some of that is the economy. Some of it is the shortlist moving to places most firms are not.
Pages, volume and logos arrive after the shortlist is written.
FAQ
How do B2B buyers build a shortlist before contacting sales?
Through peer recommendations, AI assistant answers and the people they can see at the firms named, usually the founder. The 6sense 2025 survey found the winning vendor on the day-one shortlist 95% of the time, so the work is to be named before the buyer starts.
Why has inbound dried up for my consultancy?
Most often because the shortlist is now made in places your content does not reach: peer asks and assistant answers. Run the three-step check. If you are absent from assistants and your clients name referrals as first touch, you are living on a channel you do not control.
Does AI search visibility matter for complex B2B?
It is one of three shortlist moments and the hardest to measure. Run your buyer's question monthly in two assistants and watch whether your firm's name appears; treat the result as a direction, never a count.
How often should a founder post to earn the shortlist?
Three times a week as the floor, five when it is working, for two quarters before judging. The posts cover how the firm is growing, how it markets itself and what the founder is learning, plus the framework and results that earn trust.
Is the website still worth investing in?
Yes, for the visit that confirms the shortlist. It needs the framework, the results and a frictionless way to start a conversation. It does not need to carry the whole top of the funnel, because it no longer can.
Summary
- Assume the shortlist is written before the first visit, and fund the three places it is written: peer conversations, assistant answers and the founder's feed.
- Publish three things to get in front: how the firm is growing, how it does its own marketing, what the founder is learning; publish frameworks and sourced results to earn trust.
- Post three times a week, five when working, to a defined account list, for two quarters before judging.
- Run the three-step check monthly: assistant answers to the buyer's question, where the last three clients first heard of you, and the founder's last ten posts read as a stranger.
- Build the website for the confirming visit, and stop funding more pages, cold volume or a rebrand to fix a shortlist problem.
Actualyse builds websites and campaigns designed for long, committee-driven B2B sales cycles. Book a call to talk through where yours stands.
