How to Use Feedback to Reduce SaaS Churn (A Data-Driven Playbook)

TL;DR

Customer feedback is the best early warning system for SaaS churn -- but only if you know which signals to watch, act on them before users cancel, and close the loop so customers know you listened. Feedback alone won't eliminate churn, but ignoring it guarantees you'll lose customers you could have saved.

Here's the hard truth: 96% of unhappy customers never complain -- they just leave. And since increasing retention by just 5% can boost profits by 25-95% (Bain & Company), every churned customer you could have saved represents real money walking out the door.

This guide covers:

The economics of churn (why prevention beats acquisition every time)
Hidden churn signals buried in your feedback data
How to build a feedback-based early warning system
Exit interview best practices that actually reduce cancellations
Proactive outreach strategies based on sentiment shifts
How AI can detect churn risk 2-4 weeks before cancellation
A complete step-by-step churn prevention workflow

Why Does SaaS Churn Cost More Than You Think?

SaaS churn destroys growth because lost revenue compounds over time -- a 5% monthly churn rate means you're replacing half your customer base every year just to stay flat. For startups, churn is the silent killer that makes fundraising harder, growth slower, and unit economics unsustainable.

The numbers tell a brutal story.

Acquiring a new customer costs 5 to 25 times more than retaining an existing one, according to Harvard Business Review. Yet most SaaS companies pour resources into acquisition while neglecting the customers already paying them. That neglect carries a product-side price tag too -- the true cost of not analyzing customer feedback puts a number on what the unread signals cost a 50-person team before a single customer cancels.

Here's what that looks like by segment:

Segment	Monthly Churn Rate	Annual Churn Rate	What This Means
SMB SaaS	3-5%	31-46%	Replacing nearly half your base annually
Mid-Market SaaS	1.5-3%	17-31%	Significant revenue leakage
Enterprise SaaS	0.5-1.5%	6-17%	Still meaningful at scale
Average B2B SaaS	3.5%	35%	Unsustainable without strong acquisition

Source: Recurly 2025 Churn Report via Vitally

The financial upside of retention is equally stark. Companies with high retention rates grow revenue 2.5x faster than industry peers (Bain & Company). And existing customers spend 67% more on average than new customers (BIA Advisory Services), making retention not just a defense against loss -- but a growth engine.

As Lincoln Murphy, customer success consultant and co-author of Customer Success: How Innovative Companies Are Reducing Churn and Growing Recurring Revenue, puts it: "The seeds of churn are planted early." By the time a customer hits the cancel button, you've already lost them weeks or months ago. The question is whether you had a system to detect it.

That system starts with feedback. To see the full dollar breakdown of what flying blind costs your startup -- from wasted engineering to reputation damage -- see our guide on the hidden cost of ignoring customer feedback.

One check before you build any of this: if churn is spread evenly across your base rather than concentrated in one segment, you may have a fit problem rather than a retention problem -- and no early-warning system fixes that. Our guide on how to know if you have product-market fit walks through how to tell the difference.

What Are the Hidden Churn Signals in Customer Feedback?

The most dangerous churn signals aren't angry emails or one-star reviews -- they're silence, subtle frustration, and declining engagement patterns that most teams miss entirely. For every customer who bothers to complain, 26 others remain silent (Lee Resources Inc.).

This means the feedback you aren't getting is often more telling than the feedback you are. For a deeper look at that specific failure mode -- the accounts that leave without ever filing a complaint -- see our guide on why your users churn without telling you why.

Here are the four categories of feedback-based churn signals, ranked by urgency:

1. Frustration Signals (High Risk -- Act Immediately)

These show up as negative sentiment in support tickets, feedback submissions, and survey responses. Look for:

Repeated complaints about the same issue
Escalating language ("still broken," "this is unacceptable," "considering alternatives")
Feedback that mentions competitors by name
CSAT or CES scores declining over consecutive interactions

73% of consumers cite poor customer service as the number one reason for churn (Microsoft Global State of Customer Service Report). When customers express frustration through feedback, they're giving you a chance to fix things -- but that window is small.

2. Silence Signals (High Risk -- Often Missed)

Customers who used to give feedback but stopped are sending you a message by not sending one. Watch for:

Accounts that previously responded to NPS surveys but now ignore them
Users who submitted feature requests months ago but have gone quiet
Declining engagement with in-app feedback prompts
No response to satisfaction check-ins

68% of customers leave because they feel the company is indifferent to them (Rockefeller Corporation). Silence often means they've already mentally checked out.

3. Declining Engagement Patterns (Medium Risk -- Monitor Closely)

Usage data combined with feedback tells a powerful story:

Fewer logins paired with unresolved feedback items
Reduced feature adoption after expressing confusion
Support ticket volume dropping (not because things are better -- because they've given up)
Shorter session times following negative survey responses

4. Negative Sentiment Shifts (Medium Risk -- Early Warning)

This is where AI sentiment analysis becomes valuable. Look for:

NPS scores dropping from promoter (9-10) to passive (7-8) territory
Open-ended responses shifting from positive to neutral or negative tone
Increasing mentions of price, value, or ROI concerns in feedback
Emotion patterns shifting from satisfaction to frustration over time

NPS promoters have a 95% retention rate (Bain & Company). When a promoter becomes a passive or detractor, that's a concrete signal that something changed -- and your feedback data can tell you what. NPS is a far sharper signal at the individual-account level than as a company-wide churn predictor; our guide to average NPS scores by SaaS category breaks down where that correlation actually holds.

Churn Signal Detection Quick Reference

Signal Type	What to Watch	Risk Level	Response Window	Recommended Action
Frustration in feedback	Negative language, repeated complaints, competitor mentions	Critical	24-48 hours	Personal outreach from CS or founder
Feedback silence	Previously active users going quiet	High	1-2 weeks	Proactive check-in, re-engagement survey
Declining engagement	Fewer logins + unresolved feedback	Medium-High	2-4 weeks	Usage review + targeted outreach
Sentiment shift	NPS drops, tone changes in open-ended responses	Medium	2-4 weeks	Follow-up survey, 1:1 conversation
Feature request frustration	"When will X ship?" asked multiple times	Medium	Ongoing	Transparent roadmap update

How Do You Build a Feedback-Based Churn Early Warning System?

A feedback-based early warning system combines structured feedback data (NPS, CSAT, CES scores) with unstructured signals (open-ended responses, support conversations, sentiment patterns) to flag at-risk accounts before they cancel. The goal is detection, not prediction -- giving your team time to intervene.

Here's how to build one in four layers:

Layer 1: Structured Metric Monitoring

Start with the metrics you're already collecting. If you're not sure which to use, our NPS vs CSAT vs CES comparison breaks down when each metric fits best.

Set threshold alerts for:

NPS score drops of 3+ points between survey cycles
Any NPS detractor response (0-6) from a previously happy customer
CSAT scores below 3/5 on two consecutive interactions
CES scores indicating high effort after support contacts

The probability of selling to an existing customer is 60-70%, compared to just 5-20% for a new prospect (Marketing Metrics). Protecting these relationships is worth the effort of monitoring.

Layer 2: Unstructured Sentiment Analysis

This is where most teams have a blind spot. Open-ended feedback contains churn signals that numerical scores miss entirely.

What to analyze:

Open-ended NPS follow-ups ("What's the main reason for your score?")
Feature request language and urgency
Support ticket tone and escalation patterns
In-app feedback submissions

Teresa Torres, author of Continuous Discovery Habits and product discovery coach, advocates for building continuous discovery habits where teams maintain regular touchpoints with customers. In her work on continuous discovery as a tool for churn reduction, Torres emphasizes that "the best product teams don't just wait for feedback -- they build habits around continuous customer contact so churn signals surface before they become cancellations."

Layer 3: Behavioral + Feedback Correlation

Neither feedback nor usage data tells the full story alone. The real power comes from combining them:

Low usage + no feedback = silent churn risk (they're disengaging without telling you)
Low usage + negative feedback = critical risk (they told you, and they're leaving)
High usage + negative feedback = fixable risk (they need you, but something's wrong)
High usage + positive feedback = healthy (but don't ignore them)

For the critical-risk and fixable-risk quadrants, the response itself is the retention lever -- see how to respond to negative customer feedback for the ALAR framework, timing targets by channel, and a 3-tier escalation model.

Layer 4: Automated Alerting and Routing

Turn detection into action with automated workflows:

Critical alerts (immediate): Detractor NPS from high-value accounts, competitor mentions, cancellation-intent language
Warning alerts (same-day): Sentiment decline, feedback silence from previously active users
Watch alerts (weekly review): Gradual engagement decline, mixed sentiment patterns

Predictive analytics helps businesses cut customer churn by 15-25% by identifying at-risk customers early, allowing teams to act before the cancellation happens.

What Should You Ask in Exit Interviews (And When)?

Exit interviews -- the feedback you collect at the moment of cancellation -- are your last chance to learn why customers leave and, in some cases, to save them. Exit surveys combined with smart cancellation flows cut voluntary churn by 12-15% when done right.

Timing: When to Trigger the Exit Survey

In-app, at the moment they click "Cancel." Not after. Not via email the next day. Right when they initiate the cancellation flow.

Why? Context and emotion are fresh. They're most willing to tell you the real reason while they're in the act. In-app surveys triggered contextually have significantly higher response rates than follow-up emails -- and for exit surveys specifically, you're catching someone who's already decided to engage with your cancellation flow.

The 5 Questions That Actually Matter

Keep it to 2-3 questions maximum. Brevity gets honest answers.

Question 1 (Required, Multiple Choice): "What's the main reason you're canceling?"

Too expensive for the value I get
Missing features I need
Too complicated / hard to use
Switched to a different tool
No longer need this type of product
Not getting enough value
Other (please specify)

Question 2 (Required, Open-Ended): "Is there anything we could have done differently to keep you?"

Question 3 (Optional, Scale): "How likely are you to come back in the future?" (1-5 scale)

The multiple-choice question gives you quantifiable patterns. The open-ended question gives you the real story. And the third question identifies potential win-back candidates.

What to Do with Exit Survey Data

Collecting exit feedback without acting on it is just bureaucratic noise. Here's the workflow:

Aggregate monthly: Look for patterns, not individual responses. If "too expensive" is your #1 cancellation reason, that's a pricing or value communication problem.
Route fixable issues: If someone cancels because of a missing feature that's on your roadmap, flag them for win-back outreach when you ship it.
Feed it into product decisions: Exit data should directly inform your feature request prioritization.
Track trends over time: Are cancellation reasons shifting? That tells you whether your retention efforts are working.

For a complete framework on turning feedback insights into product decisions, see our feedback strategy playbook.

How Does Proactive Outreach Based on Sentiment Prevent Churn?

Proactive outreach -- reaching out to at-risk customers before they initiate cancellation -- is the highest-ROI churn prevention tactic available. Gartner research shows proactive customer outreach reduces churn by 15-25%. The difference between reactive and proactive retention is the difference between saving some customers and preventing most losses.

Reactive vs. Proactive Churn Prevention

Approach	Reactive	Proactive
When you act	After customer cancels or requests cancellation	Weeks before, based on early signals
Data used	Cancellation reason, exit survey	Sentiment trends, engagement data, feedback patterns
Typical save rate	5-15% of cancellations	15-25% churn reduction overall
Customer experience	Feels desperate ("Please don't leave!")	Feels caring ("We noticed X -- how can we help?")
Cost per save	High (often requires discounts)	Low (often just requires attention)
Long-term impact	Delays churn, doesn't fix root cause	Fixes root cause, builds loyalty

Sources: Gartner via GrowSurf, Phoenix Strategy Group

The Proactive Outreach Playbook

Trigger 1: Sentiment Drop Detected When a customer's feedback sentiment shifts negative (NPS drops, negative language in open-ended responses):

"Hey [Name], I noticed your recent feedback mentioned some frustration with [specific issue]. I wanted to reach out personally -- can we jump on a 15-minute call this week? I'd love to understand what's going on and see if there's something we can fix."

Trigger 2: Feedback Silence from Active User When a previously engaged customer stops responding to feedback requests:

"Hi [Name], we haven't heard from you in a while and I wanted to check in. Is everything working well, or is there something we could be doing better? No survey, no form -- just reply to this email."

Trigger 3: Usage Decline + Previous Negative Feedback When someone expressed dissatisfaction and then reduced their usage:

"Hey [Name], I saw your feedback about [issue] from [date]. I want you to know we [took specific action]. I also noticed you haven't been using [feature] as much recently -- is that related, or is something else going on?"

Personalized customer experiences increase retention rates by 20% (McKinsey & Company). The key word is personalized. Generic "we miss you" emails don't work. Specific, feedback-informed outreach does.

Real-World Patterns: What Proactive Retention Looks Like

Pattern 1: The Feature Frustration Spiral A mid-market B2B SaaS company noticed a cluster of accounts submitting repeated feature requests for the same capability. None had canceled yet, but sentiment scores were declining. They sent a personalized roadmap update to each account showing the feature was planned for Q3. Result: zero churn from that cohort, and three accounts upgraded when the feature shipped.

Pattern 2: The Silent Disengagement An SMB-focused SaaS saw that accounts who stopped responding to quarterly NPS surveys churned at 3x the rate of respondents. They implemented a "silence trigger" -- any account that missed two consecutive surveys got a personal email from the founder. Result: 40% of silent accounts re-engaged, and the team identified three product issues they'd never heard about.

Pattern 3: The Price Sensitivity Signal After implementing exit surveys, a startup noticed "too expensive" was their #2 cancellation reason. But digging into the feedback data, they found these customers weren't actually price-sensitive -- they weren't seeing enough value because they hadn't adopted key features. Proactive onboarding outreach to accounts showing similar patterns reduced price-related churn by 35%.

Why Does Closing the Feedback Loop Reduce Churn?

Closing the feedback loop -- telling customers what you did with their input -- directly reduces churn because it proves you're listening. Customer feedback programs that close the loop reduce churn by 10-15% (Bain & Company). That's not a small number when compounded across your entire customer base.

The reason is straightforward: 68% of customers leave because they feel the company is indifferent to them (Rockefeller Corporation). Not because the product is bad. Not because the price is wrong. Because they feel ignored.

Nick Mehta, CEO of Gainsight and co-author of Customer Success, has repeatedly emphasized that "the number one priority for any business right now is keeping customers. It's a shift in mindset from closing as many new deals as possible to ensuring existing customers get value, grow, and remain with the company over time."

The Feedback Loop → Retention Connection

Here's how it works mechanically:

Customer submits feedback about a pain point
You acknowledge it within 48 hours ("Got it, we're looking into this")
You act on it (or decide not to, and document why)
You tell them the outcome ("We fixed X" or "Here's why we went a different direction")
Customer feels heard → Trust increases → Churn decreases

That fourth step is where most companies fail. Only 5% of companies actually tell customers they used their feedback to improve the product. The other 95% collect feedback into a void.

For a complete guide on implementing this process, see our deep dive on how to close the feedback loop without annoying your customers. And for building a broader systematic approach, our guide to building a Voice of Customer program covers the full lifecycle from collection to action.

What Closing the Loop Looks Like for Retention

For at-risk accounts specifically:

Acknowledge their frustration personally (not automated)
Share a concrete timeline for addressing their issue
Follow up when you've made progress (even if it's partial)
Ask if the resolution met their expectations

This turns a churn risk into a loyalty opportunity. Customers who experience a service failure but receive excellent resolution are 4x more likely to remain loyal than customers who switch to competitors (Bain & Company).

How Can AI Predict Churn Risk from Feedback Data?

AI-powered feedback analysis detects churn signals that humans miss by processing volume, identifying patterns across thousands of feedback submissions, and flagging sentiment shifts in real time. Businesses using AI for churn prevention reduce customer loss by 25-40% compared to manual methods.

Here's what AI can do that manual review can't:

Sentiment Trend Detection

AI tracks sentiment across every feedback touchpoint -- surveys, support tickets, in-app submissions, open-ended responses -- and identifies when an account's sentiment is trending negative, even if no single interaction looks alarming on its own.

A customer who goes from "Love the product!" (Q1) to "It's fine" (Q2) to "Still waiting on that feature" (Q3) is showing a clear trajectory. AI catches this pattern across hundreds of accounts simultaneously.

Emotion Analysis Beyond Positive/Negative

Advanced NLP goes beyond simple positive/negative classification. It detects specific emotions: frustration, disappointment, confusion, resignation. Each maps to different churn risk levels and different intervention strategies.

Frustration → Fix the specific issue
Disappointment → Reset expectations, show roadmap
Confusion → Onboarding/education gap, offer help
Resignation → Highest risk -- immediate personal outreach

FeedSense uses AI-powered sentiment and emotion analysis to automatically categorize incoming feedback and flag accounts showing negative sentiment trends -- giving you the early warning system that makes proactive outreach possible.

Keyword and Topic Clustering

AI groups feedback by topic and tracks volume changes. A sudden spike in "slow," "buggy," or "expensive" across multiple accounts indicates a systemic issue that's likely to drive churn if unaddressed.

46% of SaaS companies now use churn prediction models, and advanced implementations achieve up to 88.6% precision in predicting which customers will churn. When paired with human judgment, AI has been shown to achieve a 71% churn prevention rate -- far surpassing manual methods alone.

What AI Won't Do

Let's be honest about the limitations:

AI won't write your outreach emails (that should stay human)
AI won't replace the judgment call on whether to build a requested feature
AI can flag risk, but a person needs to decide the intervention
AI is only as good as the feedback data you're collecting -- garbage in, garbage out

The best results come from AI detection paired with human action. The machine tells you who's at risk and why. You decide what to do about it.

What Does a Complete Feedback-Driven Churn Prevention Workflow Look Like?

A complete churn prevention workflow connects feedback collection, analysis, detection, intervention, and loop closure into a repeatable system. Here's the step-by-step process, designed for teams of any size.

Step 1: Collect Feedback at Every Key Touchpoint

You can't detect churn signals in feedback you never collected. Build a comprehensive feedback collection system that covers:

Onboarding (Day 1, Day 7, Day 30): CES + CSAT surveys
Ongoing usage: In-app feedback widget, feature-specific micro-surveys
Support interactions: Post-resolution CSAT
Quarterly relationship: NPS survey to full customer base
Cancellation: In-app exit survey

Step 2: Analyze and Score

Every piece of feedback gets scored for churn risk:

Numerical scores (NPS, CSAT, CES): Track trends over time, flag declines
Open-ended responses: Run through sentiment analysis, flag negative patterns
Combined signals: Correlate feedback sentiment with usage data

Step 3: Detect and Alert

Based on the churn signal categories we covered earlier, set up automated alerts:

Alert Level	Trigger	Who Gets Notified	SLA
Critical	Detractor NPS from top-20% revenue account	CS lead + founder	4 hours
High	Two consecutive negative feedback submissions	Assigned CSM	24 hours
Medium	Feedback silence (missed 2+ survey cycles)	CS team queue	48 hours
Low	Minor sentiment decline, single negative interaction	Weekly review	1 week

Step 4: Intervene with Context

Use the feedback data to personalize your intervention. Don't just say "we noticed you might be unhappy." Reference specific feedback they gave, acknowledge specific issues they raised, and offer specific solutions.

Personalized experiences increase retention by 20% (McKinsey). Generic retention outreach doesn't work. Feedback-informed outreach does.

Step 5: Close the Loop (Always)

Every intervention needs a follow-up. Did your fix work? Did the customer re-engage? Did their sentiment improve? Close the feedback loop by telling them what you did and asking if it made a difference.

Step 6: Measure and Iterate

Track these metrics monthly:

Churn rate by feedback engagement: Do customers who give feedback churn less than those who don't?
Intervention success rate: What percentage of at-risk accounts were saved?
Time-to-detection: How early are you catching churn signals?
Loop closure rate: What percentage of at-risk interventions resulted in follow-up communication?

Make Feedback Your Best Retention Tool

Churn isn't a mystery. In most cases, customers tell you -- directly or indirectly -- before they leave. The question is whether you have a system to hear them.

The playbook is straightforward:

Collect feedback systematically at every touchpoint
Watch for churn signals in both what customers say and what they don't
Build early warning alerts that give your team time to intervene
Reach out proactively before the cancellation, not after
Close the loop so customers know you listened
Use AI to detect patterns at scale that humans miss
Measure everything and iterate monthly

Feedback alone won't eliminate churn. Pricing problems, missing features, market shifts -- some churn is structural and no amount of listening will fix it. But feedback is the best early warning system you have. It's the difference between watching customers leave through the rearview mirror and catching them before they reach the door.

Companies with high retention grow revenue 2.5x faster (Bain & Company). Start with your feedback data. The signals are already there.

Use FeedSense to Catch Churn Before It Happens

FeedSense helps you turn customer feedback into a churn early warning system.

What you get:

AI-powered sentiment analysis that flags negative trends across all feedback channels
Emotion detection that identifies frustration, confusion, and resignation patterns
Automatic categorization so churn signals don't get buried in noise
Feedback tracking with customer context preserved for personalized outreach

What it won't do:

Make your product decisions for you (that's still your job)
Replace human outreach to at-risk customers (AI flags, you act)
Eliminate all churn (some churn is structural -- we're honest about that)

If you're tracking feedback in spreadsheets and missing the patterns that predict cancellation, FeedSense gives you the system to detect churn signals and act on them before it's too late.

Start free with FeedSense -- free tier includes 100 feedback items/month, no credit card required.

FAQ

How does customer feedback help reduce SaaS churn?

Customer feedback reduces SaaS churn by revealing dissatisfaction before users cancel. Feedback data contains churn signals -- frustration patterns, sentiment declines, feature gaps, and silence from previously engaged users. Companies that systematically collect and act on feedback reduce churn by 10-15%, according to Bain & Company research. The key is acting on signals proactively, not just collecting data.

What is the average churn rate for B2B SaaS companies?

The average B2B SaaS monthly churn rate is 3.5%, according to the 2025 Recurly Churn Report. This breaks down by segment: SMB SaaS sees 3-5% monthly churn, mid-market runs 1.5-3%, and enterprise stays below 1.5%. These rates mean SMB-focused companies replace nearly half their customer base annually just to stay flat.

What are the early warning signs of customer churn?

Early churn signals include declining NPS or CSAT scores, negative sentiment in open-ended feedback, silence from previously active feedback providers, repeated complaints about the same issue, competitor mentions in support conversations, and declining product usage paired with unresolved feedback. The most dangerous signal is silence -- for every customer who complains, 26 others leave without saying a word.

How much does it cost to acquire a new customer vs. retaining one?

Acquiring a new customer costs 5 to 25 times more than retaining an existing one, according to Harvard Business Review. The exact ratio depends on industry and business model, with B2B SaaS typically landing in the 5-10x range. Existing customers also spend 67% more than new ones, making retention both cheaper and more profitable than acquisition.

Can AI predict customer churn from feedback data?

Yes. AI sentiment analysis processes feedback volume at scale and identifies churn patterns humans miss. Businesses using AI for churn prevention reduce customer loss by 25-40% compared to manual methods. AI detects sentiment trends, emotion shifts, and topic clusters across thousands of feedback submissions, flagging at-risk accounts weeks before cancellation. However, AI detection works best paired with human intervention.

What should I ask in a SaaS exit survey?

Keep exit surveys to 2-3 questions maximum, triggered in-app at the moment of cancellation. Ask a multiple-choice question about the main cancellation reason (too expensive, missing features, switched tools, etc.), an open-ended question about what you could have done differently, and optionally a likelihood-to-return scale. Exit surveys with smart cancellation flows reduce voluntary churn by 12-15%.

Why do most customers churn without complaining?

Research from 1st Financial Training Services shows 96% of unhappy customers never complain directly. Rockefeller Corporation found that 68% of customers leave because they feel the company is indifferent to them. Most people choose the path of least resistance -- canceling is easier than explaining. This is why proactive outreach based on behavioral and sentiment signals matters more than waiting for complaints.

How does closing the feedback loop reduce churn?

Closing the feedback loop -- telling customers what happened with their input -- reduces churn by 10-15% (Bain & Company) because it proves you're listening. Only 5% of companies actually communicate feedback outcomes to customers. When you acknowledge feedback, act on it, and tell the customer what changed, you build trust that prevents the "they don't care about me" feeling that drives 68% of churn.

What is the ROI of investing in churn prevention?

The ROI is significant: increasing customer retention by just 5% boosts profits by 25-95% (Bain & Company). Proactive outreach based on feedback signals reduces churn by 15-25% (Gartner). Each percentage point of monthly churn reduction increases SaaS company valuation by 15-25%. For a company with $1M ARR and 5% monthly churn, reducing churn to 4% saves $120K annually in retained revenue.

How often should I survey customers to detect churn risk?

Survey at key touchpoints rather than on fixed schedules: after onboarding (Day 1, 7, 30), after support interactions, quarterly for NPS, and at cancellation. Limit surveys to 1-2 per customer per month to avoid fatigue. The most important signal is change over time, so consistent periodic measurement matters more than frequency. Watch for customers who stop responding -- survey silence is itself a churn signal.