Every CMO eventually faces the same meeting. The CFO pulls up a slide showing the marketing budget as a percentage of revenue and asks what the return is. The CMO shows last-touch attribution data that credits the majority of pipeline to branded paid search and the SDR sequence that ran the month before close. The CFO is not satisfied. Neither, if they are being honest, is the CMO.
B2B marketing attribution is one of the most consequential and most poorly solved problems in business operations. The way companies measure marketing effectiveness determines where they invest — and most companies are measuring it wrong. They are systematically over-crediting the last touchpoint before conversion and ignoring the months of brand exposure, content engagement, community influence, and peer recommendation that created the intent in the first place.
The Dark Funnel: What Attribution Models Cannot See
The term dark funnel was popularized by 6sense to describe the portion of the B2B buyer's journey that happens outside the visibility of marketing analytics systems. It is larger than most marketing leaders want to admit.
A buyer evaluating data integration platforms might spend four months in the dark funnel before ever clicking on a vendor's website. They listen to a podcast where a practitioner mentions a specific tool. They see a LinkedIn post from a former colleague. They read a thread in a Slack community where five people discuss their experience with competing products. They sit through a conference session where a speaker uses a case study that names a specific vendor.
None of these touchpoints appear in a UTM-tracked analytics dashboard. All of them may have been decisive.
Gartner's B2B buying research has found that buyers spend only 17 percent of their total purchase decision time meeting with potential suppliers — spread across multiple vendors. The other 83 percent is spent in independent research, peer consultation, and internal deliberation. The majority of that 83 percent is invisible to standard attribution systems.
Why Last-Touch Attribution Fails B2B
Last-touch attribution assigns 100 percent of the credit for a conversion to the final touchpoint before a contact enters the CRM or submits a form. In most B2B companies, this means branded paid search and outbound SDR sequences capture the majority of attributed pipeline.
The problem is not that these activities don't contribute. They do. The problem is that they contribute to capturing demand that already existed — demand that was created by something else. Crediting branded paid search for a deal where the buyer had been following the company's podcast for six months and read three research reports is like crediting a taxi driver for getting you to the airport when the entire reason you're traveling is a decision made months ago.
Multi-touch attribution models — linear, time-decay, W-shaped, U-shaped — improve on last-touch by distributing credit across multiple recorded touchpoints. But they only credit recorded touchpoints. The dark funnel activities that drove awareness, consideration, and intent remain invisible regardless of the model used.
Self-Reported Attribution: The Simple Fix That Works
The most underused tool in B2B attribution is a single open-text field on the demo request form: How did you first hear about us?
This question, consistently asked and consistently analyzed, reveals the actual top-of-funnel drivers with a clarity that no technical attribution model can match. Companies that have implemented self-reported attribution consistently find that the answers are dominated by channels that technical models undercount: podcast mentions, LinkedIn content, peer recommendations, conference presentations, and content read months before the conversion event.
Metadata.io has published data showing that when they cross-referenced self-reported attribution data against their technical attribution data, the two disagreed on the primary source for roughly 40 percent of conversions. The self-reported data consistently credited brand and content touchpoints that the technical data missed entirely.
Self-reported attribution is not a replacement for technical attribution. It is a calibration tool. The combination of a CRM-based multi-touch model and a consistent self-reported question gives revenue operations teams a far more accurate picture of what is actually driving pipeline than either method alone.
Pipeline Attribution Metrics That Hold Up to CFO Scrutiny
Marketing-sourced pipeline — the dollar value of new sales opportunities where the first engagement was a marketing touchpoint — is the most credible metric in a CFO conversation because it connects directly to the sales forecast. If the sales team is using the same CRM data, there is no dispute about what counts.
Marketing-influenced pipeline is harder to defend precisely because it requires some judgment about what constitutes meaningful influence. The best practice is to define influence rules in advance — with sales leadership sign-off — rather than retroactively. A common standard: any opportunity where a contact engaged with marketing content or responded to a marketing-attributed outreach within 90 days of the opportunity creation date counts as marketing-influenced.
Win rate by content engagement is one of the most underused metrics in B2B marketing. Pulling the data on whether opportunities where a prospect engaged with a case study close at a higher rate than those that did not requires an integration between content analytics and CRM, but the analysis is straightforward and the results are almost always compelling. This is the kind of evidence that changes internal investment conversations.
The AI Attribution Gap
AI answer engines have introduced a new attribution problem that builds on the dark funnel challenge. When a buyer asks Perplexity which data integration platforms are worth evaluating and receives a synthesized answer that names three vendors, the buyer may never click through to any of their websites. They take the answer, add the named vendors to their evaluation list, and proceed. There is no UTM parameter. There is no form fill. There is no way to track this in a standard analytics system.
A meaningful and growing percentage of brand awareness and consideration is happening in AI engines with zero visibility for the brands involved. Companies that are not monitoring their citation share in AI engines are operating with a significant blind spot in their attribution picture. The measurement infrastructure for this channel is early — but the channel is not.
