How is win/loss research different from NPS?
NPS measures how likely an existing customer is to recommend a company, expressed as a single relationship-level score. Win/loss research investigates why one specific deal was won or lost, drawing on the buyer’s full account of the evaluation. NPS tracks sentiment over time across a customer base; win/loss research explains the outcome of an individual decision.
The two instruments answer fundamentally different questions, even though both involve asking customers or buyers about their experience. NPS asks an existing customer one question, on a recurring cadence, and produces a single number that’s meant to be tracked as a trend. A drop in NPS tells a team that something has shifted across the relationship, but it doesn’t say what shifted, when, or for whom specifically. The score is a signal that something is worth investigating, not the investigation itself.
Win/loss research starts from the opposite direction. It begins with a specific, already-closed opportunity, won or lost, and works backward through a structured interview to understand the full sequence of decisions, conversations, and perceptions that produced that outcome. The output isn’t a number. It’s a detailed account of why this particular buyer chose what they chose, including details that never reached the sales team.
There’s also a population difference worth naming. NPS is collected from existing customers, people who already made a purchase decision and are reflecting on the ongoing relationship. Win/loss research is conducted with buyers regardless of outcome, including the buyers who chose a competitor and have no ongoing relationship to protect or reflect on. That population includes the perspective NPS structurally can’t reach: the people who didn’t buy.
A mature GTM organization typically runs both. NPS functions as an early warning system for relationship health across the existing customer base. Win/loss research functions as the diagnostic tool for understanding what actually drives individual purchase decisions, won and lost alike. Neither replaces the other, because neither was built to answer the other’s question.