NPS survey software that keeps every 0–10 score tied to its reason and respondent, and themes verbatims into ranked drivers that actually move the number.
A Net Promoter Score is one number standing on a thousand reasons. The number is not the work; the reasons are.
Most NPS tools report a single gauge. This page shows what to collect, how to keep every 0 to 10 rating tied to the respondent and the verbatim behind it, and how those comments become ranked drivers you can act on — not a score you can only watch move.
By Unmesh Sheth · Founder & CEO, Sopact · Updated June 2026
The short answer
NPS survey software is a tool for running Net Promoter Score surveys: it asks how likely a respondent is to recommend a company, product, or service on a 0 to 10 scale, collects the open reason behind the rating, and reports the score along with the drivers. Promoters rate 9 to 10, passives 7 to 8, detractors 0 to 6, and NPS is the share of promoters minus the share of detractors.
The useful versions keep each verbatim tied to its score and respondent. The number always opens to the reasons underneath, and a detractor theme traces to the segment it concentrates in.
Good NPS software is built for the leader who needs the number, the manager who needs the drivers, and the team that has to win the relationship back.
30 seconds · leadership
The score and the move
The exec reads one line: NPS is 34, up two, and the top detractor driver is onboarding speed. The rest of the report exists so that line holds.
Needsone number, one driver, one direction.
3 minutes · CX or growth
The drivers, by segment
The manager scans drivers ranked by frequency, then filters to the cohort where detractors cluster. This is where the report's argument sits.
Needsranked drivers, plan and tenure cuts.
This week · the team
The reasons, in their words
The success or product team opens the verbatims behind detractors and slipping promoters and reads what people wrote — the input to a win-back, not a slide.
Needsthe verbatims behind a score band.
The recommend question is one tap, which is why most NPS tools optimize the gauge and treat the follow-up as optional. A detractor who rates "3" has a precise reason, and a promoter who rates "10" is telling you what to protect. That is the part you can act on. The fix is keeping the verbatim, the score, and the respondent bound together.
Generic NPS tool
One gauge and a trend
~5% of context reaches the analysis
Primary-data approach
Every score keeps its reason
context lifts toward most of what was said
The scale is not the problem. An NPS survey built only on the scale discards the part of the answer that names what to do. Pairing the rating with the open reason, both tied to the respondent and the segment, is the whole move.
Decide what each response captures and how it binds back to the respondent before you chart anything. The form tells you what input to use. The binding is where "segment it and track it" is built in or lost — wire it at send time and the analysis follows.
Core fields, with form and binding
The failure modes are concrete. No respondent_id and you cannot tell whether NPS rose because the same people warmed up or because detractors stopped answering. No reason_text bound to the score and a 34 has no driver. No segment and a problem in one plan reads as a base-wide trend.
Start at the smallest unit. A respondent writes a sentence under their 0 to 10 rating. A rule reads it and returns one driver category, with the original sentence kept underneath as evidence. Run it on every verbatim and a flat score opens to the reasons moving underneath it.
The rule runs on each verbatim as it arrives — no spreadsheet nobody finishes, no word cloud. The output is a count of how many respondents raised each driver, with their quotes one layer beneath.
Stage 01 · Raw input
The response, as it arrives
RSP-3380 nps: 7 (passive)
SEGMENT growth plan · 11mo
REASON "Love the product, but onboarding took three weeks and support was slow."
Stage 02 · Theming rule
One verbatim, one or more drivers
Stage 03 · Report fragment
Drivers across 980 responses
The driver attaches to the respondent record the moment the verbatim arrives, so the 30 percent on onboarding speed is not a label — it opens to the sentences that produced it. Filtered to the first-90-days cohort, that driver climbs past 45 percent, which is the cut that turns a flat NPS into a specific fix for new relationships.
Decision this enables
Which driver to move first, and for whom. The top bar is the lever on NPS; the segment filter names the cohort that owns it.
Because every answer is bound to respondent_id, one record assembles itself across waves. The current score, the reason, the bucket they fall in, and the themes sit on one record — the view a CX or growth lead reads when deciding who to win back and who to learn from.
The respondent ID joins every wave. When a new survey arrives, the record updates and the slip from promoter to passive is visible without anyone reconciling exports by email.
Stage 01 · Raw input
Two waves, same person
RSP-3380 prior wave: 9 (promoter)
RSP-3380 latest: 7 (passive)
SEGMENT growth plan · 11mo
REASON "…onboarding took three weeks…"
Stage 02 · Dictionary rule
Join and flag on the ID
Stage 03 · Report fragment
One respondent record
The movement is the signal. This respondent slipped from promoter to passive, and the reason rides on the same record. Because score and verbatim attach to one ID, a growth lead sees a churn risk to act on — a saved promoter is worth more than a blended score that ticked down a point.
Decision this enables
Who to win back this week, and what to fix for them. The slipping-promoter flag surfaces the records; the attached drivers give the team the specific reason.
The survey report examples page takes four of these build fragments apart — raw responses, the rule, the finished chart.
Relationship, transactional, product — different NPS surveys, one architecture underneath. Get these three in place at send time and the driver list, the segment cut, and the slipping promoters fall out of the data in minutes instead of weeks.
Technique 01
Persistent respondent ID
Every wave uses the same pseudonymous identifier. The ID lets you see whether NPS rose because the same people warmed up or because detractors stopped answering. Without it, every wave is a fresh aggregate. With it, movement is measurable.
Technique 02
Drivers coded at collection
Verbatims get coded into drivers the moment they arrive, not in a quarterly cleanup. The driver attaches to the same record as the score, so a chart of detractor reasons can sit next to the promoter and detractor mix from the same people.
Technique 03
Every number traces back
Every point of NPS and every percentage in a driver clicks back to a respondent, a score, and a verbatim. The citation chain is what turns a gauge into evidence a board or growth review can defend.
01What is NPS survey software?
NPS survey software runs Net Promoter Score surveys: it asks how likely a respondent is to recommend a company, product, or service on a 0 to 10 scale, collects the open reason, and reports the score with the drivers. The useful versions keep each verbatim tied to its score and respondent.
02How is NPS calculated?
Respondents who rate 9 to 10 are promoters, 7 to 8 are passives, and 0 to 6 are detractors; NPS is the percentage of promoters minus the percentage of detractors. The result runs from -100 to +100. Passives count toward the total but not the score, which is why the open reason matters for moving it.
03What is a good NPS score?
Context beats the absolute number, since NPS varies widely by industry; many treat above 0 as net positive, 30 and up as good, 50 and up as strong. The trend and the reasons behind it matter more: a flat NPS hiding a growing detractor theme is a problem the single number conceals.
04What is the difference between NPS, CSAT, and CES?
NPS measures overall likelihood to recommend; CSAT measures satisfaction with a specific interaction; CES measures the effort an interaction required. Teams often run NPS on a relationship cadence, CSAT after support tickets, and CES on self-service flows.
05Why does the open comment matter more than the score?
The score tells you the balance of promoters and detractors; it does not tell you what to change. The open follow-up holds the driver. Themed across all responses, those comments become the ranked list of reasons that actually moves NPS.
06How often should you run NPS surveys?
Relationship NPS commonly runs quarterly or twice a year; transactional NPS fires after a key event. The cadence matters less than the linkage: with the same respondent IDs you can track whether the same people moved from detractor to promoter, not just whether the aggregate shifted.
07How do you segment NPS by respondent?
Capture a respondent_id and segment fields such as plan, region, and cohort with every response. Those keys let you compute NPS per segment and trace a detractor theme to the segment it concentrates in, rather than reading one blended score for the whole base.
08Is Sopact NPS survey software?
Sopact is a primary-data platform used to run NPS and other feedback surveys where the verbatim matters as much as the rating. It captures the 0 to 10 score and the reason together, keeps them tied to a clean respondent ID across waves, and themes the comments on arrival so the drivers behind promoters and detractors surface without manual coding.
The walkthroughs above show the shape. These pages show the neighbouring tools and the open-ended method underneath.
Bring a recent NPS wave and watch the verbatims theme on arrival, bucketed by promoter and detractor and tied to clean respondent IDs you can follow across waves. Thirty minutes, your questions, your data.
