Customer success · health to renewal
A health score nobody can open is a colour, not a signal.
Every product in this category ships a proprietary model and a weighting slider. The first time it calls a good account red, nobody can see why, and from then on the column is decoration.
So the formula is one your team writes, over signals you choose, and you can run it against your own accounts before it goes anywhere near a dashboard.
- A formula you write, validated and simulated first
- Signals from usage, billing, support and surveys
- A missing signal is neutral, never zero
- Product usage28-day active users against seats74
- Time since anybody spokelast activity 24 days ago31
- Invoices past duenone outstanding100
- Survey scorenever surveyedneutral
- Open ticketstwo open, one high severity62
Three answers a competitor does not have
The questions to ask a customer-success vendor.
Each of these is answerable here, and each is a page you can read before you ever speak to us.
Where the signals come from
Four other modules, which is the point of it not being a separate product.
Product usage. How recently anybody actually spoke to them. Invoices past due. The three survey types. How engaged their executives were in the last review. Open support tickets, weighted by severity.
A bolt-on customer-success tool gets most of that through an integration that breaks quietly, or does not get it at all. Here they are the same records the rest of the business is already writing.
- A band change does four things, and records whether anybody was told —
A ledger entry, a notification to the owner that reaches email or chat by their own preference, the identifier of that notification written back onto the entry, and an event.
- A drop into red creates one task, not one per hop —
Only on a genuine drop — and it checks for an existing open one first, so a flapping account does not produce a task a day.
- A human's risk flag beats the automatic one, permanently —
The band mirrors onto open renewals as a risk level unless somebody has set it by hand.
Renewals
Five reminders, and two things that actually happen.
Seeded from both subscriptions and contracts, four months out, with value normalised to a twelve-month figure so a monthly and an annual plan are comparable.
At ninety days your playbook applies itself. At sixty the renewal opens as a deal — routed through the same path a person would use, specifically so the automated one still produces stage history and an audit row.
- 120It appears, seeded from the subscription or the contract
- 90Your renewal playbook applies itself
- 60A real deal opens — in the pipeline, forecast, with history
- 30Reminder
- 7Reminder
A discount becomes a real approval
Above a threshold you set, with a clock, escalating to the owner's manager by adding them rather than replacing them.
One bad record cannot stop the night
Every renewal, every reminder and every playbook run gets its own savepoint.
Churn does not mark the deal won
Three of the four outcomes mirror onto the linked deal. The fourth deliberately does not.
The whole module
Every capability, with a page that names the screen.
Written after reading how each one actually behaves — which is why each page tells you what it refuses to do as plainly as what it does.
Knowing where you stand
Getting them live
Six parts, bought separately, on one account record.
Health, usage, renewals, reviews and plans, surveys and onboarding are separate purchases — and every one of them survives the day somebody merges two duplicate accounts, which a bolt-on integrated by interface does not.
Sales CRM
The renewal is a deal in your pipeline, and the account with its whole history is the account the score is about.
ExploreFinance
Subscription periods and contract terms seed the renewals, and past-due invoices are a health signal.
ExploreSupport
Open tickets and their severity feed the score, so a quiet account with three critical tickets does not read as healthy.
ExploreCall intelligence
A renewal conversation can hear what the last quarter actually sounded like, rather than reading a number.
ExploreQuestions that decide it
Including the ones about what it will not claim.
›Is there a churn prediction model?
No, and deliberately not. The score is arithmetic your team wrote and can read. Churn is an outcome somebody records on a closed renewal, not a probability the product asserts.
›Does it report net revenue retention?
No. Recurring revenue and churned recurring revenue are computed in the finance module; a retention rate is not calculated anywhere. If that is a board metric for you, plan to derive it.
›Does the renewal forecast handle several currencies?
It groups by currency and shows a line each, rather than adding unlike units into one number. The product can convert elsewhere; this does not.
›Is there a customer-facing portal?
No. The two things a customer sees are a survey they respond to and a success plan you share by link — and that link stores only a hash of itself, can be revoked, and records the first time it was opened.
›Can we export health scores or survey responses?
Not today. The screens are where that data is read. Worth knowing if your reporting lives elsewhere.
›Does closing a renewal extend the contract?
No. Marking one renewed records the outcome; the term itself is advanced by billing on its own cycle. Treat them as two acts.
Book a demo
Bring ten accounts and your own idea of what healthy means.
The third one is the demo that decides it.
- Write the formula, and watch the editor refuse a bad one
- Simulate it against those ten before activating anything
- Find an account that has never been surveyed and read its score
- Set a renewal sixty days out and watch the deal appear