for customer success · churn root-cause
Forty churn-cause interviews without making each customer write a two-page exit survey.
Lacudelph is the interview surface for CS leads who need to know why customers actually churned — beyond “price” on the cancellation form. Each at-risk or churned account gets the same structured conductor; you get a cohort view of which factors recurred, which were idiosyncratic, and which segment to triage before the next renewal cycle.
The current shape of the work
The reasons people give a cancellation flow are not the reasons they actually left. NPS scores told you a number; they didn’t tell you that the QBR slipped, the champion changed jobs, and the procurement team quietly insourced the workflow. Getting that signal usually means a CS manager chasing 30-minute Zooms with people who have already moved on.
- Exit-survey response rates sit at 5–10%, almost entirely from the angriest decile. The quiet majority of churned accounts ghost.
- Manual churn calls take a CS manager 30 minutes plus 30 minutes of write-up — and you only get to whoever picks up.
- NPS plus a free-text field surfaces complaints, not the cross-account pattern (“the four customers we lost in Q2 all had a champion change in the prior 90 days”).
- Renewal-risk signals (“quiet” accounts, low-touch trials) need the same treatment, and there’s never enough CS bandwidth to interview both groups.
How Lacudelph changes it
1
CS lead writes one brief
Five fields about what you want to learn from each at-risk or churned account — what shifted in their world, which factors they’d re-rank now, what would have changed the decision. Lacudelph generates the conductor and the meta-noticing rules.
2
Send the link to the cohort
Drop the URL into the existing cancellation flow, the QBR follow-up email, or a one-off CS outreach. The AI conductor adapts per respondent — drills where they say something concrete, moves on where they don't.
3
Each respondent gets a private reflection
At session close they receive their own structured reflection — sections picked from what they actually said: what they named, what shifted as they talked, and one factor they almost-but-didn't surface. Trust mechanism, not extraction. Higher response rates because the artifact is theirs.
4
You get the cohort aggregate
Convergent churn factors, divergent framings, recurring hedges across accounts (“everyone said price but none of the cohort lost on price”), and routing recommendations: which at-risk accounts to escalate, which segments to pull QBR cadence forward in. Pro tier.
What a churn-cause brief looks like
A worked example for a B2B SaaS reviewing last-quarter churn. Substitute your own accounts and hypotheses.
- Goal
- Surface the actual cluster of factors driving churn in last quarter's lost accounts — beyond what the cancellation form captured. Test which of our hypotheses about pricing pressure, champion changes, and feature gaps each respondent thinks were load-bearing.
- Audience
- Primary contacts at the 12 accounts that churned in the last 90 days, plus the 8 accounts currently flagged as at-risk in the next renewal cycle.
- Hypotheses to check
- (a) Renewals where the champion left in the prior 90 days are systematically harder to save; (b) the workflow we replaced is being insourced by procurement, not switched to a competitor; (c) feature parity matters less than time-to-first-value at re-onboarding.
- What respondents get back
- A reflection tailored to their account — sections picked from what they actually said: what shifted in their context, where their account converged or diverged from the cohort, and one factor they hadn't put into words before this conversation. Theirs to keep, not extracted.
Common questions
How is this different from sending an exit survey through Gainsight or HubSpot?
Exit surveys collect a static set of free-text fields — response rates sit at 5–10%, dominated by the loudest decile. Lacudelph runs an adaptive multi-turn conversation that probes Socratically when the customer says something concrete, and the participant gets their own structured reflection back at the end (not extraction, mutual exchange). Higher response rates because the artifact is theirs; cross-account patterns surface that no individual exit-survey field captures.
Can I send the link from my existing CSM-managed comms (HubSpot, Outreach, raw email)?
Yes — the participant link is just a URL. Drop it into your existing cancellation flow, QBR follow-up email, or one-off CS outreach. Each customer opens it when their calendar permits — no scheduling, no Calendly round-trip.
What tier do I need for cross-account aggregation?
Pro tier ($99/mo). Free and Solo tiers can run individual interviews and produce per-account takeaways; Pro is the tier when you want the cross-account cohort report — convergent churn factors, divergent framings, segment-level routing recommendations.
Does the customer know it's an AI conversation?
Yes — explicitly. The participant entry page tells them they're talking to an AI research moderator (not a human), and the takeaway carries an AI-authorship disclosure (EU AI Act Art. 50 compliant). Counter to intuition, this often increases candour: customers say things to a structured AI conductor they wouldn't say to a CSM whose comp depends on retention.
Can the conductor adapt to enterprise vs SMB segments?
Yes — the brief generation step takes your audience definition and produces an interviewer persona calibrated to it. You can also run two parallel rounds (enterprise + SMB) with two briefs and compare them; or one round with both segments and read the cohort report's convergent vs divergent splits.
Run your next churn review on Lacudelph
Free tier covers a single brief and 5 sessions — enough to pilot one churn cohort. Cross-account aggregation lands on Pro at $99/mo.