Marketing measures what it can see, and what it can see most clearly is an individual: a named contact who opened the email, took the call, attended the demo, replied with a question. This individual becomes, almost by default, the unit through which progress gets tracked. Their engagement deepens, their title gets noted, their responsiveness improves, and somewhere along the way their behavior starts to stand in for the health of the deal itself.

The trouble with this default is straightforward to state and easy to overlook in practice: in a meaningful share of consequential B2B purchases, the individual being tracked is not the entity that makes the decision. The decision belongs to a group — a set of stakeholders with different priorities, different veto power, and different reasons to say no, all of whom have to arrive at enough agreement to move forward together. The engaged individual may be an advocate inside that group, a champion, even the person who initiated the whole process. They are still one voice inside a decision unit that has to reach its own kind of consensus before anything moves, and their engagement, however strong, does not automatically produce that consensus.

The Unit Problem

This is worth naming precisely as a unit-of-analysis problem, because that is what it actually is. Marketing organizations build their measurement systems, their lead scoring, and their sense of pipeline health around the individual as the fundamental unit of progress. But the individual is not the decision-making entity in the purchases this measurement system is meant to track. The committee is. A measurement system built around the wrong unit will report real, accurate information about that unit — the individual genuinely is engaged, genuinely is progressing in their own understanding — while telling leadership almost nothing reliable about whether the actual decision-making entity is progressing at all.

This distinction is not the same as noticing that B2B purchases involve multiple stakeholders, which is now widely acknowledged and has produced a great deal of tactical response — stakeholder mapping, multi-threaded outreach, account-based targeting. Those responses treat the problem as one of coverage: reach more people inside the account, and the committee problem is solved. That misses the more fundamental issue. Coverage of more individuals is still measurement at the individual level, simply distributed across more individuals. It does not change the unit being measured; it multiplies it. The actual unit that decides — the group, functioning as a group, reaching or failing to reach internal agreement — remains invisible to a measurement approach built around tracking individuals, no matter how many individuals are being tracked simultaneously.

Why Individual Engagement Doesn’t Transfer

The reason strong individual engagement doesn’t reliably translate into committee movement is that the committee’s obstacles are frequently internal to the committee, not external to it. A buying group can be composed of members with genuinely different priorities — one weighing cost, another weighing risk, another weighing operational disruption — and reaching internal agreement among them is its own separate process, with its own separate timeline, that does not accelerate simply because one member became more convinced. An engaged champion inside the group may in fact be racing ahead of the group’s actual internal consensus, becoming more confident in the decision while the rest of the group has not moved at all.

Research into complex B2B buying consistently describes internal group dynamics — disagreement among stakeholders, conflicting priorities, and difficulty reaching consensus — as a significant and recurring source of stalled decisions. This is consistent with treating the committee, not the individual, as the site where most of the friction in a stalled deal actually lives. A vendor’s champion can be fully convinced and fully engaged while the deal stalls entirely inside a disagreement between two other stakeholders the champion doesn’t control and may not even be fully aware of.

The Misread in Reporting

This produces a specific and recurring misread inside pipeline reporting. A deal with a highly engaged, responsive individual contact gets reported as healthy, progressing, or “hot,” because the individual signals available to marketing and sales are all pointing in a positive direction. The deal can nonetheless be stalled — not despite the champion’s engagement, but somewhat independent of it, because the actual obstacle sits inside a part of the buying group the champion’s engagement was never going to resolve. When the deal eventually stalls or dies, the explanation offered is often vague — “went quiet,” “priorities changed,” “lost to no decision” — because the individual-level signals genuinely didn’t predict the outcome, and no one was tracking the unit where the real obstacle existed.

What Changes

Recognizing this distinction does not require abandoning attention to individual engagement, which remains a genuinely useful and necessary signal — a champion is not nothing. What changes is what that signal is allowed to represent. Individual engagement should be read as evidence about that individual: their interest, their conviction, their willingness to advocate. It should not be extended, without separate evidence, into a claim about the state of the group’s internal agreement, because the individual’s engagement and the group’s consensus are produced by different processes, on different timelines, sometimes influenced by entirely different considerations.

The more accurate posture for leadership reviewing pipeline health is to treat strong individual engagement as informative but insufficient — a necessary input to a decision, not a report on the decision’s status. The actual unit that will approve or decline the purchase is the group, and the group’s condition — whether it is aligned, divided, stalled, or still forming a view — is a separate question that individual-level engagement metrics were never built to answer, and generally don’t.