NPS & Survey Analysis

NPS verbatims clustered into themes and ranked by impact - not a top-line number on a dashboard. The next product conversation starts with the three drivers that move score, evidenced by raw voice and segmented by who said what.

What the managed workflow does

Tracks the score over time

12-month rolling NPS with monthly response volume, segment mix, and quarter-over-quarter delta - so movement is visible, not buried in a single 'Q2 average' card.

Clusters every verbatim

Reads every free-text response, groups by theme, and tags severity with frequency × segment × usage signal - so the readout names what to fix, not 1,420 quotes to skim.

Cross-tabs by segment

Theme × segment heatmap shows where each driver lives - promoter signal vs detractor signal vs passive - so investment leverage is visible per cohort.

Names the top drivers

Top 3 retention drivers and top 3 churn drivers per cycle - sourced to verbatims, with prior-cycle delta - so leadership has a continuous track, not a single-quarter snapshot.

NPS over 12 months · segment mix · drivers per theme

1,420 RESPONSES · 9 THEMES · NPS +52 (+14 QoQ)

Inputs synthesized

In-product NPS micro-surveyFree-text verbatimsCustomer cohort metadataProduct usage signal · per-respondent

What lands per cycle

Promoter signal · what's workingDetractor signal · what's bleedingPassive split · investment leverageTheme synthesis · top NPS drivers

Inputs synthesized

In-product NPS micro-survey. Sampled across logged-in users on a stratified cadence - not a once-a-quarter mass blast that catches a non-representative slice.
Free-text verbatims. Every open-ended response, parsed and clustered against your driver dictionary - themes carry forward across cycles.
Customer cohort metadata. ARR band, persona, contract age, geography - drives the cross-tab without exposing individual identity.
Product-usage signal. What each respondent does in-product - claimed sentiment is checked against actual behavior so 'I love it' from a 0-DAU detractor reads correctly.

What lands per cycle

12-month rolling track. NPS over time with segment mix per month - leadership sees the slope, not the single-quarter point.
Driver synthesis. Top 3 promoter drivers and top 3 detractor drivers per cycle, sourced to verbatim and quantified by mention frequency.
Theme × segment heatmap. Where each driver concentrates - so PM, CS, and pricing each see the cohort their work moves.
Anonymized verbatim. Direct quotes per top driver, identifying detail stripped - so the exec readout starts with raw voice, not a synthesized paraphrase.

What you get, every week