Last-Click Attribution Is Dead — Your Dark Funnel Is 70% of Your Pipeline
January 2026 brought a wave of 'attribution is dead' manifestos from B2B leaders. The headline was overdramatic; the underlying problem isn't. Roughly 70–80% of the B2B buying journey now happens in channels your analytics platform cannot see.
A CMO ran a dashboard demo at her board meeting last quarter. Pipeline by source: paid search, organic, content syndication, events. The CFO asked the obvious question. "If we cut paid search, the pipeline numbers would drop, right?" The CMO hesitated. The honest answer was: she didn't know. Her dashboard showed paid search as the source of last touch on a third of pipeline. It didn't show that the buyers had already decided before they clicked the ad.
This is the attribution problem that broke loose in the first quarter of 2026. The dashboards still work. The numbers still tally. The numbers are just quietly fictional. Roughly 70–80% of the B2B buying journey now happens in dark social and dark funnel channels — DMs, Slack groups, podcasts, LinkedIn comments, peer recommendations — that no analytics platform can see. The last click your attribution model awards credit to is the moment a decision became visible, not the moment it was made.
January 2026 saw a wave of B2B leaders publishing "attribution is dead" pieces — Jeff Pedowitz, Chris Walker, the analyst team at 6sense. They weren't being theatrical. They were watching the gap between what the dashboards said and what the deals actually told them widen past the point of plausible deniability.
The Dark Funnel Is Where the Decision Happens
The dark funnel is everything that happens between "buyer becomes aware of the problem" and "buyer fills out a form." For most of B2B history, that gap was narrow. The buyer searched. They clicked. They filled out the form. The path was traceable.
The path stopped being traceable somewhere around 2020 and stayed that way. Buyers self-educate now. They watch podcast episodes about your category. They lurk in Slack communities where peers share product recommendations. They DM each other on LinkedIn asking which vendor to evaluate. They scroll through the comments on your competitor's launch post. None of this is captured in your analytics platform. All of it shapes who they call.
By the time a buyer fills out your demo form, they have already decided. Survey data from 2025–2026 shows 83% of B2B buyers fully or mostly define their purchase requirements before they ever speak to a sales rep. The form fill isn't the start of the funnel. It's a request for a price quote.
Self-reported attribution surfaces what dashboards hide. Companies that started asking "How did you hear about us?" on demo forms found 30–50% of pipeline came from channels their digital attribution couldn't see. Not 5%. Not 10%. Half. The dashboard had been wrong by half for years.
The Single-Source Model Is Where the Lie Lives
Last-click attribution is the most extreme example, but every single-source attribution model has the same structural problem. It assumes the touchpoint it can see is the touchpoint that mattered. It can't see the dark funnel, so by construction it under-credits the channels driving the actual decision and over-credits the channels that show up at the bottom of the funnel.
Multi-touch attribution is better but still wrong. It distributes credit across the touchpoints it can see, which means it distributes credit across a subset of reality. A model that can see only 25% of the buyer journey isn't more accurate because it averages across that 25%. It's confidently wrong instead of obviously wrong.
The honest answer is hybrid measurement. Multi-touch where you can see, self-reported where you can't, signal-based intent data to triangulate. None of the three alone is sufficient. Together, they produce a picture that's defensible without being precise.
Where This Shows Up in Practice
Budget allocation conversations. The CFO asks which channels drive pipeline. The dashboard says paid search and content syndication. Self-reported attribution says podcasts, peer referrals, and the founder's LinkedIn. The right answer is the second one. The conversation that gets had is usually about the first.
Brand and demand budgets. Brand investments — podcasts, conference keynotes, founder content, community sponsorship — show up nowhere in last-click attribution. They show up everywhere in self-reported attribution. Companies that cut brand budgets to chase the metrics they can see are usually cutting the channels that drive the deals.
Marketing-sales alignment. Sales says half the pipeline comes from "word of mouth" and "founder network." Marketing says the dashboard shows paid and SEO. They're both right, and they're both incomplete. The disconnect is structural, not personal.
Forecasting accuracy. When the dashboard is the source of truth and the dashboard is structurally incomplete, the forecast is too. RevOps teams that have started layering self-reported and intent data on top of the digital dashboard report tighter forecasts. The gain comes from acknowledging the gap, not from closing it.
What to Actually Do About It
Add a "How did you hear about us?" question to every demo form. Free-text answers are gold; required dropdowns are noise. Free-text takes longer to analyze but produces signal you can't get any other way. Run the analysis monthly. The patterns emerge in the second or third month.
Run win/loss interviews on a real cadence. A 30-minute call with five recent buyers asking "What did you read, watch, or hear before you booked the demo?" produces dark funnel data that no software vendor will sell you. Most sales orgs run win/loss interviews ad hoc; the teams getting attribution right run them on a regular schedule and synthesize across cohorts.
Triangulate, don't pick a single source of truth. Multi-touch attribution for the digital surface, self-reported attribution for the dark funnel, intent data for the in-market signal. Three imperfect sources, weighted, beat one perfect-looking source.
Stop killing channels because the dashboard says they don't work. If the dashboard shows zero attributed pipeline from your podcast, the dashboard is wrong before the podcast is. Validate with self-reported data before you cut. The channels you can't measure cleanly are usually the ones that are working.
Make peace with measurement uncertainty at the executive table. Boards that demand precision will get fiction. Boards that accept the dark funnel exists will get directionally correct allocation. The mature framing is: "we don't know exactly which channel drove this deal, but we know which mix of channels drives deals like it." That's the honest answer. Practice saying it.
The Stakes
The companies that doubled down on last-click attribution through 2024–2025 are the same companies whose growth charts flattened in 2026. They optimized the metrics they could see and starved the channels they couldn't. The metrics improved. The pipeline didn't.
The companies that accepted the dark funnel as a structural reality — and built measurement systems that triangulate around it — are the ones whose budget allocation tracks the deals they actually close. Their dashboards look messier. Their forecasts hit.
Attribution isn't dead. The single-source dashboard is. What replaces it is hybrid, slightly fuzzy, and harder to defend in a CFO meeting that wants a single number. It also tells the truth, which is a higher bar than any dashboard cleared in the previous decade.