Transactional NPS: how to measure satisfaction at the moment it matters

Transactional NPS (tNPS) is a Net Promoter Score survey triggered by a specific event in the customer journey — a completed purchase, a service appointment, an onboarding milestone, or a resolved support ticket. It asks the standard NPS question, "How likely are you to recommend us to a friend or colleague?", on a 0–10 scale, but it asks it about that interaction and right after it happens.

That timing is the whole point. A relationship survey sent quarterly asks customers to average months of experience into a single number. A transactional survey asks them about something that happened this morning, while the details are still sharp and while you can still act on what they tell you.

Both use the same 0–10 question and the same Promoter / Passive / Detractor buckets. The difference is scope and cadence:

• Relationship NPS (rNPS) is sent on a fixed schedule (quarterly or twice a year) and asks about the brand overall. It is a strategic, board-level metric — it tells you whether loyalty is trending up or down.
• Transactional NPS (tNPS) is sent after an event and asks about that event. It is an operational metric — it tells you which locations, agents, or processes are creating or destroying goodwill, right now.

You want both. rNPS is the thermometer; tNPS is the diagnosis. A flat relationship score can hide a service desk that is quietly burning customers — the kind of thing only a per-interaction signal surfaces in time to fix.

The math is identical to standard NPS. Group responses to the 0–10 question:

• Promoters score 9–10.
• Passives score 7–8.
• Detractors score 0–6.

Then: NPS = % Promoters − % Detractors. The result runs from −100 to +100. Calculate it per touchpoint — per location, per channel, per team — not just as one company-wide figure, or you will average away the very variation you ran the survey to find.

Resist the urge to obsess over the absolute number. A tNPS of +40 means little in isolation; the trend for this location versus last month, and versus your other locations, is what tells you where to act.

Latency is the silent killer of transactional surveys. Both recall and response rates decay within hours, so the trigger should fire as close to the event as the channel allows:

• In-person service (a visit, an appointment, a repair): within 1–2 hours of checkout or job completion.
• E-commerce: after delivery is confirmed, not at the time of order — the experience is not complete until the box arrives.
• Support tickets: immediately on resolution, while the help is still fresh.
• Onboarding: when a meaningful milestone is reached, not on a calendar date.

Match the channel to the moment, too. SMS and WhatsApp earn fast, high-rate responses for time-sensitive transactions; email suits longer questionnaires and B2B. And always respect quiet hours, frequency caps, and consent — a survey that annoys is worse than no survey at all.

The single biggest mistake teams make with tNPS is treating the number as the deliverable. The number is a routing signal. What you do in the seconds after it lands is the actual program:

• A detractor is an open service-recovery window. Route that comment to a human immediately — the research on recovery is consistent: a fast, personal save often produces more loyalty than if nothing had gone wrong.
• A promoter is social proof waiting to happen. This is the right moment — and only after they have told you they are happy — to invite a public review.
• A passive is a question: what one thing would have made this a 9 or 10? That free-text answer is your product roadmap in miniature.

Always pair the 0–10 question with one open-ended follow-up. The score tells you how many; the comment tells you why — and the why is what you can fix.

It is tempting to use the tNPS score to send only promoters to your public review profiles and quietly bury detractors. Don't. Selectively suppressing negative reviews — "review gating" — violates the FTC Fake Review Rule (16 CFR Part 465) and the terms of service of Google and most major platforms.

The compliant pattern is simple: survey everyone, and invite everyone to leave a public review. Use the tNPS score to decide follow-up, not eligibility — surface a detractor's comment to your team for a private save in addition to, not instead of, the public invitation. You get the operational signal and the review velocity without the legal exposure. See our review gating entry for the full picture.