Open any executive dashboard and you’ll find the same quiet act of violence. A thousand customer opinions compressed into a single number. NPS of 42. CSAT of 88%. Average satisfaction: 7.3. The chart looks calm, authoritative, decision-ready. And in producing that calm, it has deleted the one respondent who could have changed your year.
Averages are a summarizing instrument, and summarizing means discarding. That’s the whole job. The median tells you where the crowd clustered, which is genuinely useful when you need to decide whether to ship or hold. It is structurally incapable of telling you what the person at the edge of the distribution was trying to warn you about.
We can say this with some confidence, because in 2025 our team conducted nearly 1,000 stakeholder interviews across 33 research studies. When you aggregate that volume, you get clean rankings. The top reasons buyers chose a vendor were knowledgeable teams, user-friendly platforms, and strong service and support. Tidy. Defensible. The kind of finding that fits in a slide.
But the rankings are the map, not the territory. The number that actually reframed how clients thought about retention wasn’t in the top-line list at all. It was this: across those studies, 34% of projects explicitly named partnership and relationship-driven engagement as a primary driver of satisfaction, retention, or wins. Partnership doesn’t show up as a feature on anyone’s comparison grid. It isn’t a checkbox a survey thinks to ask about. It surfaced because analysts heard the same unprompted theme in interview after interview and followed it. A dashboard optimizing for the ranked feature list would have smoothed that signal into nothing.
What the number is actually for
Survey data is a map of the terrain. It tells you how many, how often, how concentrated. It tells you that buyers lose net-new deals over functionality gaps, unclear positioning, and pricing–hard numbers that justify a budget line or anchor a thought-leadership claim. A good quantitative study earns its keep by sizing the problem and ending the debate about whether the problem is real.
What it cannot do is tell you why a prospect found your positioning unclear, or what they were actually confused about before they gave up and defaulted to “pricing.” The number locates the wound. It cannot describe the pain.
What the outlier is actually for
The insight that reshapes a roadmap rarely comes from the median response. It comes from the customer who describes a problem in language your product team has never used. Consider what our 2025 work surfaced on dissatisfaction: roughly 30% of studies pointed to communication and coordination gaps–not product failures–as the major opportunity to strengthen relationships. When dashboards are tuned for product metrics, they are measuring the wrong axis. The thing quietly eroding trust wasn’t the feature set; it was the silence between updates. You only learn that by listening for insights beyond the satisfaction score.
This is why the two studies work as a sequence. The survey doesn’t replace the interview. It tells you which interview to go conduct. Run the quant study, find the segment whose answers cluster strangely, then send a senior analyst to sit with those people and ask the follow-up the survey could never ask. The number identifies the anomaly. The conversation explains it.
Even AI, the topic every dashboard now wants to quantify, proves the point. It appeared in 40% of our 2025 studies, but the finding that mattered wasn’t the percentage. It was the split underneath it: AI builds satisfaction when it’s practical and secure, and corrodes it when value is uncertain or adoption drags. Same technology, opposite outcomes, invisible to any single number.
The next time a chart hands you a confident figure, ask what it had to throw away to look that clean. The answer is usually the most interesting customer you have.
A dashboard tells you where everyone agreed. An interview tells you who was right.
James Rice – Chief Digital Officer
