How to Automate NPS Surveys with n8n and Smart Segmentation

A Net Promoter Score that just sits in a dashboard is a vanity metric. The number moved, nobody acted, the detractor who scored you a 3 churned three weeks later, and the promoter who'd have left a glowing review never got asked. n8n NPS survey automation earns its keep only when each score triggers a different next move.

Search results for this job collect scores and stop there. n8n's library has a GoHighLevel-Gmail-Notion NPS flow and a cross-platform NPS tracker built on Bright Data, and automation blogs like Restflow cover monitoring and tagging. They capture and store. What they skip is the segmentation flywheel: the explicit 0-6 / 7-8 / 9-10 logic and the three different downstream actions that turn a score into a retained customer or a public review.

The score is the input, not the output

Here's the take that reframes the whole build: NPS is a router, not a report. The standard buckets exist precisely because each one demands a different response. A detractor needs a human in the next hour. A passive needs education about the feature they're not using. A promoter needs to be asked, today, while they still feel it. Most NPS tooling treats all three the same, dumps them in a spreadsheet, and calls it analytics. That's the gap.

If you build the routing and nothing else, you've already beaten the ranking pages.

What you can automate in an NPS program

• Sending the survey on a trigger (post-purchase, day-30, renewal, support close)
• Capturing the 0-10 score plus the free-text reason
• Bucketing into detractor, passive, and promoter automatically
• Alerting a human the moment a detractor scores you low
• Dropping passives into a feature-education nurture sequence
• Routing promoters to a review or referral ask while sentiment is high
• Logging every response to Sheets so the trend line is real, not anecdotal

The bucketing and the three routes are the part worth getting right. Everything else is standard n8n.

The NPS routing pipeline

Trigger (Schedule / event)
  → Send survey (email or SMS with the 0-10 link)
  → Webhook (capture score + comment)
  → Code (bucket: detractor / passive / promoter)
  → Switch (route by bucket)
       detractor → Slack alert + recovery email
       passive   → nurture drip
       promoter  → review / referral request
  → Google Sheets (log)

The Code node in the middle is four lines. It's also the difference between an NPS dashboard and an NPS engine.

1. Send the survey

A Schedule trigger fires the survey on your cadence, or an event webhook fires it after a specific moment (order delivered, ticket closed). Send a one-click 0-10 link by email or SMS. Keep the survey itself dumb: one number, one optional comment box. The intelligence lives downstream.

A note on the Schedule trigger that the tutorials gloss over: if it fires every minute while a previous run is still in flight, n8n can silently drop the overlapping execution. For a daily survey send that's irrelevant, but if you batch-send to a large list, move to a Cron expression with a sane interval and let each batch finish.

2. Capture the response

The survey's submit button hits an n8n Webhook node. Pull the score (0-10) and the free-text comment into clean fields. If a comment exists on a low score, this is where an optional OpenAI call earns its place: summarize the complaint into a one-line reason so the detractor alert is actually actionable, not just "customer scored 2."

3. Bucket the score

The whole flywheel turns on this Code node:

const score = Number($json.score);
let segment;
if (score <= 6)      segment = 'detractor';
else if (score <= 8) segment = 'passive';
else                 segment = 'promoter';
return [{ json: { ...$json, segment } }];

That's it. Three buckets, NPS-standard. Now the Switch has something concrete to route on.

4. Route each segment differently

The Switch node reads segment and forks three ways.

Detractor (0-6): fire a Slack alert to the account owner now, with the score and the AI-summarized reason. Queue a recovery email or, better, a human reach-out task. Speed matters here. A detractor reached within a day is a save; reached next week, they're already gone.

Passive (7-8): these customers are fine, not thrilled. Drop them into a short nurture drip that highlights the feature they're probably not using. A Wait node spaces the sends. Passives are the cheapest segment to convert into promoters, and almost nobody works them.

Promoter (9-10): ask. Send a Google review request or a referral link while the goodwill is fresh. This is the segment that funds the whole program, and the one most NPS setups forget to act on.

5. Log everything

A Google Sheets append on every path: timestamp, customer, score, segment, comment, and the action taken. That sheet is your real NPS trend plus a record of what you did about it. When someone asks "did the detractor outreach work," you have rows, not vibes.

Implementation patterns

Pattern A — the detractor SLA. Treat a low score like an incident. Stamp the alert with a target response time and, if no human marks it handled within the window, escalate to a second Slack channel via a Wait node and a status check. Detractors decay fast.

Pattern B — promoter gating before the public ask. Don't send every promoter straight to a public review link blind. Confirm the comment isn't secretly negative (a 9 with a "but the onboarding was rough" note). A quick OpenAI sentiment check on the comment catches the mismatch, so only genuinely happy promoters hit the public review step.

n8n nodes you'll use most

Getting started

1. Pick a trigger: a Schedule for cadence surveys or an event webhook for post-purchase.
2. Send the 0-10 survey link by email or SMS, comment box optional.
3. Capture the response in a Webhook node and flatten score and comment.
4. Add the four-line Code node to bucket the score into a segment field.
5. Wire a Switch on segment with three distinct follow-up branches.
6. Build the detractor alert, the passive drip, and the promoter review ask.
7. Append every response to Google Sheets and watch the segment mix over time.

The promoter branch is where NPS pays you back. The ReviewFlow review request engine ships the promoter-to-public-review flow with a built-in pre-screen that sends unhappy customers to a private form instead, and the User Feedback Loop handles the multi-source ingest and sentiment scoring that feeds the whole pipeline.

An NPS score that triggers nothing is a number you'll regret collecting. Bucket it, route it, and let each segment pull its own weight. For the upstream collection side read How to Automate Customer Feedback Collection with n8n, and for the retention angle on detractors see n8n Churn Prevention. The detractor gets a human. The promoter gets asked. The passive finally learns about the feature they paid for.