A health score you can explain
A health score is only useful if you can say why an account got it. If the answer is "the software said so", you will either ignore it or argue with it.
Retenova takes the opposite approach: every score comes from visible rules, and every rule can be edited.
How the score is built
Each account starts from a baseline of 75. Specific signals then move the score up or down, and the panel for each account lists every adjustment next to its reason.
The starting rules
These are the default thresholds in the preview. They are starting points, not proven answers, and you can change all of them in the Playbook.
| Signal | Default rule | Effect |
|---|---|---|
| Usage drop | Down 25% or more over 14 days | Watch, or At risk at twice that |
| Renewal | Within 30 days and no review in the last 90 days | At risk |
| Satisfaction | NPS of 6 or below | At risk |
| Silence | No contact for 21 days or more | At risk |
| Growth | Seat count up 15% or more | Expansion signal |
Why not a black box
- Trust. A founder will act on a flag they understand, and ignore one they do not.
- Fit. A ten-customer startup and a sixty-customer startup need different thresholds. Editing a number is faster than waiting for a vendor.
- Honesty. With little data per customer, a model that claims precision would be pretending. Simple rules are easier to check.
What we do not know yet
The thresholds above are our starting guesses. We expect design partners to tell us which are too strict or too loose, and we will adjust them. Learning from many customers' outcomes is a later step, and it will need real data and clear consent.
You can watch the rules at work in the demo on the home page or in the live preview.