6sense's research on B2B buying groups contains a number that should be more disruptive than it has been.
95% of winning vendors were already on the buyer's Day One shortlist. And 94% of buying groups had ranked that shortlist before speaking to a single seller.
That single number relocates the entire contest. The deal is not decided in the evaluation. It is decided in whatever happened before the evaluation started, in whether your name was one of the few written down at the beginning.
Now ask what your attribution model can see. It observes buyers who are on your properties, identifiable, and generating events. Every one of those conditions is met after the shortlist exists. Your dashboard is a detailed, accurate, real-time report on the part of the journey that comes after the outcome was mostly determined.
It is also, at best, watching one person. Forrester's State of Business Buying, 2026 puts a typical purchase decision at 13 internal stakeholders plus nine external influencers. Your analytics resolves to a session and a cookie: one device, one identity. The decision involves twenty-odd people, nine of whom do not work at the company and cannot appear in your data under any configuration you could build.
It helps to name what the missing part actually looks like. A Slack thread where someone asks the channel who they use for this, and three people name the same vendor. A podcast a buyer half-listened to on a commute that lodged a company name. A subreddit where a stranger takes your pricing apart against two alternatives, with receipts. A group chat between two people who worked together at a previous company. None of that is exotic or rare. It is where most of the shortlist gets built, and not one second of it will ever reach your reports.
The paradox
The better your attribution gets, the more precisely you starve the work that actually wins deals.
A crude model is bad at telling you what worked, but it is visibly crude, so nobody fully trusts it and judgment still gets a vote in the budget meeting. A sophisticated model is credible. It produces confident numbers you can defend to a CFO. And because those confident numbers still only describe the observable, post-shortlist window, credibility is exactly what converts a measurement limitation into budget policy.
The tool is not broken. It is working as designed, and the design assumes the decision happens where the tracking is.
This has already happened
Gartner's 2026 CMO Spend Survey (401 marketing leaders, fielded January through March) shows what it looks like once it reaches the money.
Awareness and conversion now take 62.6% of total media spend, up more than 10% since 2024. Loyalty and retention fell to under 15%, a 29% decline in two years.
Nothing in that research suggests marketers concluded that keeping customers stopped paying. Acquisition spend happens to sit at the two moments attribution reads most cleanly: the click that opens a session and the form that closes it. Retention spend does not. Budget moved toward where evidence was easiest to produce, not toward where returns were highest.
That is the argument with a dollar figure attached.
And the observable window keeps shrinking
Gartner's Sales practice surveyed 646 B2B buyers in August and September 2025 and found 67% prefer a rep-free experience for at least part of their purchase, up from 61% the year before. More of the decision is happening in places that generate no interaction to measure, and the trend is moving in one direction.
Improvado's analysis puts roughly 38% of B2B pipeline as arriving with no attributable touchpoint at all. That figure comes from an attribution vendor's own research, so take the direction rather than the decimal. The direction is not seriously disputed.
The gap between what your dashboard reports and what your market is doing is widening on its own, without anyone changing anything.
What to do instead
None of the fixes need new software.
The cheapest one is to ask the buyer directly. A single open field on your demo or contact form, "How did you first hear about us?", captures more of the dark funnel than any tracking script, because it queries the only system that watched the whole journey.
The next is to track shortlist presence rather than pipeline alone. If 95% of wins are set at shortlist formation, the honest leading indicator is whether buyers in your category can name you unprompted. That is a survey question. It costs less than most martech line items, and it measures the thing that actually decides the outcome.
The hardest is to ring-fence a budget you have agreed not to attribute. Set aside a fixed share of spend, exempt it from attribution review, and judge it once a year against category awareness instead. Left under attribution review, that budget gets cut on schedule, regardless of what it returns.
Keep the dashboard. It is genuinely useful for what it covers. But say out loud what it covers: the late, visible, identifiable portion of a decision that was substantially made before any of it began. Optimize against it without acknowledging that, and it will teach you, one defensible, well-evidenced decision at a time, to defund everything that works.