There is a quiet anxiety running through leadership teams right now. The pipeline reports still tick along. The forecast still closes at roughly the rate it always has. Yet something feels off. Deals are arriving more formed than they used to, with buyers who already know your name, your pricing logic, and often your weaknesses. By the time your commercial team gets involved, the decision has effectively already been made.
This is not a failure of sales or marketing individually. It is the visible edge of a much larger shift in how B2B buyers actually decide — and it is a question for the leadership team, not a single function.
The buyer journey has moved out of sight
The numbers behind this shift are striking, and they should change how every commercial leader thinks about the architecture of their go-to-market. 6sense’s 2025 Buyer Experience Report found that 95 percent of B2B buyers now purchase from their Day One shortlist, up from 85 percent the year before. 94 percent of those buyers rank vendors in preference order before contacting any of them. The pre-contact favourite wins 80 percent of the time.
Sit with that for a moment. The shortlist is formed before you can see the buyer. The winner is largely chosen before any commercial conversation takes place. Everything your CRM records is the visible aftermath of a decision that was effectively made elsewhere.
Where is elsewhere? In the channels that have always been invisible to your systems, and which are now carrying more commercial weight than ever: Slack groups, WhatsApp threads, LinkedIn DMs, internal procurement debates, peer conversations over a coffee or a pint. The likes, comments and impressions your business can measure are the small visible tip. The shares, recommendations, internal forwards and private endorsements are the mass beneath the waterline — and that is where most purchase decisions are actually made.
Add to this a genuinely new variable: generative AI. 6sense found that 94 percent of B2B buyers now use large language models during their purchasing process. Magenta Associates’ 2025 research showed 66 percent of UK decision-makers use tools like ChatGPT, Copilot and Perplexity for vendor research, with 85 percent saying they have discovered a new vendor through AI-assisted research. That interaction happens in a closed environment. No click. No fingerprint. No record in any system the business owns. The most informed, highest-intent moment in the buyer’s journey is now routinely invisible to the leadership team accountable for the number.
Vercel’s CEO offered a useful data point on the scale of this shift: ChatGPT referrals grew from under 1 percent to 10 percent of new sign-ups in six months. In a less mature intelligence environment, all of that would show up as unexplained, unattributed growth — credited to nothing, understood by nobody.
Recommendation is the real commercial engine
Strip the modern B2B buying decision back to its mechanics and you find something quite old-fashioned at the centre of it: people ask other people. TrustRadius’s 2024 B2B Buying Disconnect Report found that 78 percent of buyers selected products they had heard of before they even started formal research. Gartner reports that 61 percent of B2B buyers now prefer a rep-free buying experience entirely.
This should reframe two things commercial leadership teams often get wrong.
The first is what positioning is for. Positioning is not a brand guideline, a tone of voice document or a set of messaging pillars — those are downstream outputs. Positioning is a leadership decision about what the business stands for and who it serves, tested against what customers and prospects actually believe, not what the business hopes they believe. Its job is to be recommended: to be the name a buyer’s peer types into the message when asked “who should we be looking at.” If the business is not on the Day One shortlist, it is very likely not winning the deal — and that is a question about how the position was set and tested, not how well it was communicated.
The second is what the business’s market intelligence is for. It does not exist to generate leads or downloads — those were always a proxy, and a weak one. Its purpose is to tell leadership what is actually being said about the business in the rooms it cannot see: the private conversations, the Slack channels it will never be invited into, the AI tools synthesising its reputation on someone else’s screen. If leadership cannot answer, with evidence, what its market believes about it right now, it is flying blind regardless of what the dashboard says.
Attribution logic has been quietly distorting commercial strategy
The reason so many commercial functions have drifted away from this is not incompetence. It is incentive design. Last-click attribution, and the broader culture of immediate measurable return, has trained a generation of commercial leaders to fund what they can prove rather than what actually works.
The logic is seductive. If an activity produces a tracked lead this quarter, it gets credit, more budget and a place on the dashboard. If an activity shapes a buyer’s perception over eighteen months and contributes to a deal that arrives looking unsourced, it gets nothing. Over time, investment migrates almost entirely towards the bottom of the funnel and the most immediate signals — demand capture that is, on closer inspection, just demand harvesting, with no plan for where next quarter’s demand comes from.
The LinkedIn B2B Institute, drawing on Professor John Dawes’s work at the Ehrenberg-Bass Institute, has been making the point persistently: at any given moment, only around 5 percent of a business’s potential market is actively in the buying window. The other 95 percent are out of market, but they are not inert. They are forming the impressions, opinions and shortlists that will decide who gets contacted when their buying window opens. Optimising relentlessly for the in-market 5 percent feels efficient. It is also why so many commercial functions hit a growth ceiling they cannot explain. You can only convert demand that already exists, and the work of creating tomorrow’s demand is happening somewhere the attribution model cannot see it.
The data makes this concrete. HubSpot, one of the most sophisticated organic-search operators in B2B, watched its measurable traffic fall from 13.5 million visits in November 2024 to under 7 million by December. Rankings up. AI citations growing. Traffic down 80 percent year on year. Any leadership team running the business against a last-click view of the world would read that as a collapse, and cut investment in exactly the channels now doing more of the influencing work, simply because the click trail has gone quiet.
Building the intelligence to see what attribution cannot
This is the point every leadership team eventually reaches, and it is an uncomfortable one. A significant share of what wins a business its next deal will never be cleanly attributable in a CRM: the peer recommendation, the half-remembered podcast, the analyst’s offhand mention, the LinkedIn post that landed in the right inbox at the right time, the ChatGPT session that put the business on the shortlist. These will continue to drive pipeline that arrives looking like it came from nowhere.
Trying to force visibility onto this through ever more elaborate attribution modelling is the wrong response. It produces false confidence and worse decisions. The right response is to build a different kind of commercial intelligence — one that does not depend on the click trail at all.
That means independent voice-of-customer research treated as a standing commercial asset, not a one-off project: structured, ongoing evidence of what the market actually believes about the business, tested against what leadership assumes it believes. It means win/loss analysis that goes beyond “we lost on price” to the commercial architecture question underneath it — was the business ever really on the shortlist, and if not, why not. It means treating share of voice relative to share of market as a leading indicator of the pipeline the business will be reporting in three or four quarters’ time, because the relationship between the two is one of the most reliable predictors of future growth. It means holding Net Revenue Retention and customer advocacy as evidence of the same dynamic playing out with existing customers, because retained and referring customers are the most efficient growth engine any business has.
This is not a softening of commercial accountability. It is a sharpening of it. These measures are harder to game and harder to flatter than a lead count, which is precisely why boards should trust them more.
What this means for the leadership team
For a leadership team sitting with this, the practical implication is not to abandon demand capture. It is to rebalance the business’s commercial architecture against the decisions that demand capture cannot make for it.
That means building the intelligence infrastructure to see the buyers a click can’t reach: independent VoC as a standing input to strategy, win/loss run as a commercial diagnostic rather than a sales post-mortem, and a genuine, evidenced position on what the business’s market believes about it — including what it believes from the AI tools now doing a meaningful share of the research on leadership’s behalf. It means accepting that the business’s most valuable commercial work may show up two, three or four quarters after it was done, and designing the board narrative, and the board’s patience, to match.
Commercial influence is going dark. It is not going out. The leadership teams that recognise the difference, and build the architecture to act on it, will be the ones whose pipeline keeps arriving pre-warmed. Their competitors will keep wondering why their attribution model has stopped explaining anything.
