Turn NPS comments into a ranked driver list: read every verbatim on arrival, cited, tied to the score and the person, not a word cloud.
NPS verbatim analysis is the practice of reading the open-ended comment each respondent leaves beside their 0-to-10 rating, coding it into themes, and tying every theme back to the score and the person who wrote it. Done well, it turns a single Net Promoter Score into a ranked list of the reasons behind it — the drivers a team can actually fix. The score says how many; the verbatim says why, and only if someone reads it.
The complaint practitioners voice is almost always the same: “we have thousands of comments and nobody has time to read them, so we run a word cloud and move on.” A word cloud counts words; it does not tell you that price is driving detractors in one segment while onboarding is driving them in another. The reading is the work, and it is exactly the work that gets skipped.
Key takeaways
A Net Promoter Score compresses a person’s whole experience into one digit, and then the comment beside it explains the digit. Most NPS tools store both but treat them differently: the score is charted, trended, and benchmarked, while the comment sits in a column nobody opens. So the part of the survey that could tell you what to change is the part that goes unread.
The limit is architectural. A form-centric feedback tool stores a score in one bucket and a comment as loose text, with no mechanism to read the comment against a codebook or to keep it attached to the respondent across the next survey. Sopact calls the alternative the Verbatim Thread: every NPS comment read against your codebook the moment it arrives, themed and cited to the sentence, and kept on the same persistent contact record as the score and every later wave. A detractor stops being a number in a bucket and becomes a person with a stated reason you can follow up. The same record-centric model runs the rest of Sopact’s feedback tools.
NPS analysis moved through three eras. First, the emailed survey and a spreadsheet: scores tallied by hand, comments skimmed if at all. Then the NPS dashboard — Delighted, AskNicely, SurveyMonkey and their peers — which automated the score, the trend line, and a word cloud, making the number effortless and the reasons still shallow. The current era reads the verbatim itself: each comment coded into themes and drivers on arrival, tied to the score and the respondent.
The one test that separates the eras: ask a tool to show you, on one screen, a detractor’s score, the exact sentence that explains it, the theme it belongs to, and whether the fix reached that person by the next wave. A dashboard can show the score and a word cloud; it cannot show the sentence, the theme, and the follow-through on one record. If the answer is a word cloud, you are looking at score infrastructure, not reading infrastructure.
Detractors (0–6), passives (7–8), and promoters (9–10) are not three intensities of the same feedback; they answer different questions. Detractor comments name what is broken and who is leaving. Passive comments name the one thing standing between indifference and loyalty, which is usually the cheapest win available. Promoter comments name what to protect and the exact language your best customers use to describe the value, which marketing can borrow verbatim.
Reading them as one pile averages those signals into mush. A theme like “support” means “support failed me” in a detractor and “support saved me” in a promoter, and a word cloud shows only that “support” is frequent. Segmenting the read before theming is the difference between a chart and a decision, and it is the same discipline that detractor analysis applies to the bottom of the scale.
The workable method is four steps: clean the comments, split by segment, theme each comment against a fixed codebook with the quoting sentence kept, then count themes as drivers and rank them by frequency and by score impact. Cleaning removes blanks, gibberish, and duplicates so the counts are honest. Splitting by segment keeps detractor and promoter meaning apart. Theming against a fixed codebook — rather than an ad-hoc read each quarter — is what makes wave two comparable to wave one.
The output is not a word cloud but a ranked table: driver, how many detractors named it, an example quote, and the average score of the respondents who raised it. That last column is what turns theming into prioritization, because a driver raised by twenty detractors at an average score of 2 is a bigger fire than one raised by forty passives at a 7. Sopact runs these four steps on arrival, so the ranked list exists the day responses land rather than a quarter later.
A word cloud and a themed read start from the same comments and end in very different places: one shows which words are frequent, the other shows which reasons drive the score and who raised them. The contrast is the whole case for reading.
| The question | Word cloud | Themed read (Verbatim Thread) |
|---|---|---|
| What do you get? | The most frequent words, sized by count | Ranked drivers with a quote and a score behind each |
| Can it separate detractors from promoters? | No: all comments pooled into one image | Yes: each segment themed for its own question |
| Does a theme trace to a person? | No: words are detached from respondents | Yes: every theme cited to the comment and the contact |
| Can you act on it? | Rarely: a word is not a reason | Yes: a ranked, quoted driver list routes to owners |
The themed read is what survey analysis does for every open-ended question, not just NPS; the mechanics are walked through on how to analyze survey data.
An NPS number is a lagging summary; the reason a customer gave it is in the comment, and only if someone reads the comment while it still matters. That is the premise of the Loop, Sopact’s method for continuous intelligence: collect clean at the source, analyze the moment a response arrives, improve while there is still time to act. On a feedback program that means a detractor’s reason is themed the day it lands and routed to someone who can close the loop, not filed for a quarterly readout.
The Loop is also what makes an NPS finding defensible: every driver and theme traces to the exact comment it came from, the standard detailed in Loop traceability, so “price is the top detractor driver” is backed by the sentences, not a hunch.
One method, three moves that never stop
Then the cycle runs again, a little sharper each time. Read the method: the Loop methodology →
The fastest way to feel the difference is to run last quarter’s comments through it. Export your NPS scores and verbatims, then paste the prompts below into Sopact Sense’s Assistant, or reason through them with your team. The arrow above each links the Academy walkthrough with the expected output and tips.
Academy walkthrough → Analyze open-ended survey responses
Here are our NPS comments with each respondent's score: [ATTACH]. Theme every comment against our codebook, cite the sentence behind each theme, and show the top five themes among detractors, passives, and promoters separately.
Academy walkthrough → Analyze sentiment and its drivers
For these NPS verbatims: [ATTACH], score sentiment per comment, then show the specific drivers behind the detractor comments — the recurring reasons, each with a quoted example — so I know what to fix, not just the score.
Academy walkthrough → Connect quant and qual data
Here are our NPS scores and open-ended comments on the same respondent IDs: [ATTACH]. Show which themes explain the detractor scores, quote a comment for each, and list the specific respondents worth a personal follow-up.
Academy walkthrough → Analyze NPS across waves
Here are two waves of NPS from the same respondents on the same IDs: [ATTACH]. Track each respondent's movement between waves, show who moved from detractor to promoter and who slipped, and surface the comments that explain the biggest moves.
Each walkthrough is short and practical: what to do, the prompt to run, the output to expect, and the tips that keep it reliable.
Watch: reading feedback on arrival and tying every comment to the person who gave it.
NPS verbatim analysis is reading the open-ended comment each respondent leaves beside their 0-to-10 rating, coding it into themes, and tying each theme to the score and the person. It converts a Net Promoter Score into a ranked list of drivers. Sopact reads every comment on arrival and keeps it on the Verbatim Thread, so the reason and the score never separate.
Clean the comments, split them by segment, theme each against a fixed codebook with the quoting sentence kept, then rank the themes by frequency and by the average score of the respondents who raised them. Sopact runs these steps on arrival, so a ranked driver list exists the day responses land rather than after a manual reading marathon.
A word cloud counts words, not reasons, and pools detractors with promoters so the same word carries opposite meanings. It cannot trace a theme to a person or a score. For deciding what to fix, a themed read that ranks drivers and cites quotes is the difference between a chart and a decision.
Because they answer different questions: detractors name what is broken, passives name the cheapest path to loyalty, and promoters name what to protect and how to describe the value. Sopact themes each NPS segment for its own question, so the drivers are not averaged into mush.
By theming every comment against a fixed codebook and keeping it on a persistent contact record, the Verbatim Thread. Because wave two is coded the same way as wave one and lands on the same respondents, you can see who moved from detractor to promoter and read the comment that explains why.
Yes. Sopact reads comments in multiple languages against the same codebook on arrival, so a program running in several languages is themed as one dataset rather than separate manual backlogs. The multilingual mechanics are covered on the multilingual survey analysis page.
Not a word cloud but a ranked driver table: the driver, how many respondents in a segment named it, an example quote, and the average score of those respondents. Sopact produces that table on arrival, with each row traceable to the exact comments behind it.
No. The score stays the headline metric; verbatim analysis explains it. Sopact keeps the score and the themed comment on the same record, so leadership sees the number and the team sees the ranked reasons behind it, both traceable to source.
Next: work the bottom of the scale on NPS detractors, or read every comment on one record with feedback tools.
