Sentiment is a CVM signal, not a support metric
Most teams file customer sentiment under 'CX'. That's a mistake. Tone is one of the earliest churn predictors you have.
Sentiment usually lives in the support org: a CSAT score, a quarterly NPS deck, a wall of dashboards nobody opens.
But sentiment is a leading indicator of value at risk. A frustrated customer rarely files a "cancellation intent" form. They leave clues — a sharper tone in a ticket, a lukewarm review, a survey comment that trails off.
Avg. sentiment (90 days pre-churn)
From vibes to segments
The reason sentiment stayed a support metric is that reading it at scale was impossible. You sampled, you summarized, you moved on.
Modern models change the economics. You can now:
- Score every conversation, not a sample.
- Attach that score to a customer, not just a ticket.
- Feed it straight into a segment — "declining sentiment + high value + renewal in 60 days."
Ticket
Score
0.82 → 0.41
Customer
Segment
High value + risk
Play
That last line is a CVM play, not a CX report.
The point
If your sentiment data can't be turned into a segment, it isn't finished — it's decoration.
This is exactly the itch behind Echo, a small experiment I've been building to turn messy customer signals into something a lifecycle team can actually act on.