NPS feedback is the score, the verbatim, the prior history, and the attached context — read on one record. Most programs collect it and never close the loop.
NPS feedback is the full signal that arrives with a Net Promoter Score — the rating, the verbatim, the same customer's prior submissions, the context underneath each of those. Most NPS programs collect all of it and never close the loop. Sopact reads every response on arrival, routes the failure named in the verbatim to a named owner with the prior context attached, and tracks the loop until it actually closes.
By Unmesh Sheth · Founder & CEO, Sopact · Updated May 2026
NPS feedback is the full signal that arrives with a Net Promoter Score: the 0 to 10 rating, the open-ended verbatim that follows it ("What is the primary reason for your score?"), the same customer's prior responses, and any other record attached to that contact. The score alone is not the feedback. The feedback is the score plus the verbatim plus the context, read together on one record.
NPS is one of the most disciplined feedback instruments in customer experience: a single 0 to 10 question with a single open-ended follow-up. Two columns of data, comparable across customers, repeatable across waves. The narrowness is the point.
NPS as a metric is a number. NPS as a program is a loop — the verbatim is read, the failure is named, an owner takes the action, the loop closes. A program that reports the score and stops there is using the instrument but not the discipline.
An average score from a thousand customers is noise. One detractor's verbatim — named, dated, attached to the contact — is signal. The verbatim is where the next quarter's churn is named in plain English. The score only counts it after.
An NPS response is not one piece of data. It is four, each on the same persistent contact ID. Programs that treat the score as the response are working with the smallest strand and discarding the other three.
A number from 0 to 10. Quantitative. Comparable across customers and across waves. Useful for trends, useless for action on its own.
The customer's open-ended answer in their own words. Qualitative. Names the failure or the win. The only strand the team can act on directly.
The same customer's score and verbatim from prior waves. Without it, every response is a stranger. With it, every response is a trajectory.
Case notes, support tickets, documents, transcripts attached to the same contact. The verbatim reads against all of these. The context is where the action becomes specific.
All four strands live on one record per customer. The score reads against the prior score. The verbatim reads against the prior verbatim and the attached document. The trajectory replaces the snapshot.
The strand most programs work with is strand 01. The other three sit in exports nobody opens. That is the most common reason an NPS feedback program produces beautiful charts that fail in the moment of action.
For twenty years, "NPS feedback" meant a quarterly dashboard. The score moved, the chart updated, somebody wrote a memo. The verbatim was exported but rarely read. The customers kept naming what was wrong; the team kept reporting whether the number moved.
Two things changed. First, the analysis got easy — every modern AI model classifies ten thousand verbatims against any codebook a team can write in seconds. The reading is no longer the bottleneck.
Second, the standard of what counts as "closed-loop NPS" has changed with it. A quarterly summary memo no longer qualifies. The work moved to the response level: every verbatim read on arrival, routed to the right owner with the prior context attached, with the action and its result tracked on the customer's own record — so the loop closes around the customer, not around an average.
The rating answers what changed. The verbatim answers why. The contact ID answers since when. The attached context answers against what. The loop closes when the action named in the verbatim happens, gets logged, and is visible the next time the customer comes back.
This is the same locked argument that anchors /use-case/nps-analysis, the analytical pillar — expressed here through the closed-loop frame. The pillar covers the broader methodology; this page focuses on the action layer.
The phrase "closed-loop NPS" is used loosely. These are the five stages it has to pass through to earn the description. Programs that do stages one and two and call themselves closed-loop are doing something else.
Every NPS response carries the score and the open-ended comment. They live on a persistent contact ID, alongside any prior submission from the same customer. The collection is the easy part; tying the response to the same contact across waves is the load-bearing part.
Every verbatim is read the moment the response submits, classified against a versioned rubric the team controls (named failure modes, promoter triggers, new themes flagged for review). The original wording is preserved on the record. The label and the source comment both stay.
A detractor whose verbatim names a known failure mode routes to a specific person — the CS lead, the account director, the program coordinator — with the original wording attached, the prior-quarter verbatim attached, and any other record from the same contact pulled alongside. The owner arrives at the conversation already informed.
The owner has the conversation, makes the fix, escalates, or escalates back. Whatever happens is logged on the customer's same record — the same persistent ID that holds the score, the verbatim, the prior history. The action is part of the record from this point forward.
When the same customer answers the next quarterly NPS, the team can see the prior response, the verbatim, the action taken, and the resolution — on one record — before reading the new score. The next wave reads against a closed loop, not a stale one. That is what closed-loop means.
Both programs below have a quarterly NPS survey, a dashboard, an executive summary, and a Slack alert when scores cross a threshold. Both report themselves as closed-loop. Only one earns the description.
The dashboard says closed-loop. The same customer's Q4 verbatim is unanswered.
The dashboard looks the same. The customers' relationships are different.
"Closed-loop NPS" is almost always self-reported. The honest test is per-response, not per-program: for each detractor in the last wave, can the team show what the verbatim said, who called back, what was done, and whether the customer's score moved the next time? A closed-loop program can answer that for every detractor. Most cannot.
The instrument and the loop are the same across all three. Only the moment of collection changes — and that changes what the verbatim is going to be useful for.
Right after a support ticket closes, an onboarding session ends, a release ships, a renewal call wraps. The verbatim names what worked or what broke about that specific event. Most useful read against the same customer's relational NPS — one bad transaction is not a churn signal; one bad transaction from a customer whose relational NPS just dropped is.
A quarterly read on the customer's overall view of the relationship — not any single interaction. The verbatim names what is working overall, what is fraying, what is at risk. Run with a persistent contact ID across waves, it becomes a longitudinal signal — the same customer's relationship trajectory, wave by wave.
The same 0 to 10 question and verbatim, asked of employees. The verbatim is the part HR teams reliably underuse: the resignation that arrives six months later was usually named in the eNPS comment two quarters earlier. Same loop, same record-per-person discipline, different audience.
Sopact runs the same loop across all three. The contact ID is the customer's or the employee's; the codebook is the team's; the verbatim reads against the prior verbatim and any attached document. The instrument is the same. The audience is what changes.
This page covers the full feedback signal and the loop that has to close around it. Three adjacent reads in the NPS cluster handle the analytical methodology, the commercial buying decision, and the time axis.
What NPS analysis means in 2026, the methodology, the AI-era thesis, the longitudinal context. The broader treatment above this page.
Read the pillar →The full feedback signal — rating, verbatim, prior history, attached context — and the loop that has to close around it.
This pageThe buying decision — categories, criteria, comparison table for teams shopping for a tool that reads the open-ended comment specifically.
Read the sub-hub →If you came here to understand what NPS feedback is, stay on this page. If you came to compare tools for reading the verbatim, the sub-hub is the right next read. If you came for the broader methodology, the pillar is.
Same NPS instrument, three different buyers. The cost of an unclosed loop is different in each context — and the loop-closing workflow is what makes the program defensible to the people the program reports to.
A relational NPS at quarterly cadence, with the verbatim routed to the account owner the day it arrives. The CS lead walks into the renewal call already having read the detractor verbatim from two waves back, what was done about it, and whether the score moved the next time. The customer notices.
A training cohort with an end-of-program NPS plus a 90-day follow-up. The participant verbatim names what transferred to the job and what did not. The program team reads each response on arrival, routes the curriculum signals to the lead facilitator, and quotes participants verbatim in the next funder report. The loop closes around the cohort.
A scholarship or grant program with NPS-style feedback after each cycle plus a six-month follow-up. Each awardee's response lands on the same record as their original application. The board reads the headline number and the awardee verbatims, in their own words, attached to specific names. The case for next cycle writes itself.
Your scores, your verbatims, your contacts. Sixty minutes. No demo accounts.
NPS feedback is the full signal that arrives with a Net Promoter Score: the 0 to 10 rating, the open-ended verbatim ("What is the primary reason for your score?"), the same customer's prior responses, and any other record attached to that contact — case notes, prior surveys, documents. The score alone is not the feedback. The feedback is the score plus the verbatim plus the context, read together on one record.
The score is a number on a -100 to 100 scale, calculated as the percentage of promoters minus the percentage of detractors. The feedback is the entire signal the customer left: the rating they gave, the comment they wrote, the history they have with the program, and how all of that compares to their prior submission. A program that reports the score and discards the feedback is throwing away the part the team can actually act on.
The NPS feedback loop is the workflow that closes around a single response: collect the score and the comment, read both on arrival, route the result to the right owner with the customer's prior context attached, take the action the verbatim names, and confirm the loop closed before the next wave. Most NPS programs collect step one and call themselves closed-loop. The loop is only closed when the action is taken and recorded.
Three failure modes. First, the verbatim is summarized into a sentiment label and the original wording never reaches the person who could act. Second, the score routes by threshold but without the prior context, so the CS lead arrives blind. Third, the loop is reported quarterly instead of tracked per response, so individual save calls never happen and the metrics only check whether average sentiment moved. A closed-loop program closes around each customer, not around an average.
Transactional NPS feedback is collected right after a specific interaction — a support ticket, an onboarding session, a renewal call, a feature release. It measures the experience of that moment. The verbatim names what worked or what broke about that specific event. Transactional feedback is most useful when read against the same customer's relational NPS history, so the team can tell whether one bad experience moved the relationship.
Relational NPS feedback is collected on a fixed schedule (usually quarterly) and measures the customer's overall view of the relationship — not a single interaction. The verbatim names what is working overall, what is fraying, what is at risk. Run with a persistent contact ID across waves, relational NPS becomes a longitudinal signal — the same customer's relationship trajectory, wave by wave.
eNPS is the employee version of NPS — the same 0 to 10 question and open-ended follow-up, asked of employees instead of customers. The verbatim is the part HR teams reliably underuse: the resignation that arrives six months later was usually named in the eNPS comment two quarters earlier. eNPS feedback works the same way as customer NPS — the score is the easy part; the work is reading every verbatim on the same employee's record across quarters.
Customer feedback is any input from a customer — surveys, support tickets, reviews, conversations. NPS feedback is a specific instrument: one rating, one open-ended question, run on a defined cadence. Its discipline is its narrowness. The mistake is treating NPS as the only customer-feedback channel; the strength is that NPS produces a comparable signal across waves when run with a persistent contact ID, which most other feedback channels do not.
The score is an arithmetic compression — two very different customers can produce the same number. The verbatim is the customer's own description of what is working or what is broken. A score moves; a verbatim names the failure mode. A CS lead can make a save call from the verbatim. They cannot make a save call from the score.
Closing the loop means: the verbatim is read on arrival; a named owner takes the action the verbatim suggests; the conversation gets logged on the same customer's record; the action and its result are visible the next time anyone reads the customer's feedback. A "closed-loop NPS program" that produces a quarterly summary memo is not closing the loop — it is reporting that the loop existed. Real closure happens per response, on one record.
NPS feedback is the signal. NPS analysis (/use-case/nps-analysis) is the methodology that reads it — the analytical pillar of the cluster. NPS verbatim analysis (/use-case/nps-verbatim-analysis) is the commercial sub-hub for teams shopping for a tool that reads the open-ended comment specifically. Run with a persistent contact ID across waves, the whole thing becomes a longitudinal signal (/use-case/longitudinal-design) — the same customer's relationship trajectory over time.
Three things. Read every verbatim on arrival — the volume excuse is gone. Attach the response to a persistent contact ID so the same customer's trajectory is visible across waves. Route the failure named in the verbatim to a named owner with the prior context attached, and confirm the loop closed before the next wave. The score still has a place. The verbatim is where the work actually happens.
NPS feedback is the signal. The cluster covers the methodology, the buying decision, and the time axis. Pick the door that matches the question.
The pillar above this page — what NPS analysis means in the AI era, the workflow, the methodology, the locked thesis behind the cluster.
For teams shopping a tool that reads the open-ended comment specifically — six buyer-checklist criteria and a category-by-category comparison.
What it means to follow the same contact across waves — the missing axis on most NPS programs, and the one that turns wave-by-wave feedback into a relationship trajectory.
Your scores, your verbatims, your contacts. Sixty minutes. We read each response on arrival against the prior wave, route a detractor verbatim to a named owner with full prior context attached, and walk through what closing the loop on every response would have changed about the last quarter. No demo accounts. No slideware. Your own feedback, read live.
No slideware. No demo accounts. Your own feedback, read live.
