Most renewal forecasts are assembled the week before they are needed, from a spreadsheet, by asking each CSM how they feel about their accounts. It produces a number. It does not produce confidence, and everyone in the room knows it.
Why the forecast breaks
It breaks because the two halves of the answer live apart. Contract dates and revenue sit in the CRM. The judgement about whether a customer is in good shape sits in somebody's head, or in a tool the finance team cannot see. Joining them by hand, once a quarter, under time pressure, is where the accuracy goes.
What a forecast needs to be defensible
- The same source as the team's daily work. If the forecast is rebuilt somewhere else, it will disagree with the team, and the team will win.
- A drill-down to accounts. Any figure should open into the customers behind it in one click.
- A stable definition of at-risk. Not one that gets renegotiated whenever the number looks bad.
- Enough warning. A forecast that turns red in the renewal month is a report, not a forecast.
Rank by revenue at stake, not by date
Working renewals in date order is how teams spend a fortnight on small accounts while a strategic customer drifts. Rank by the revenue at stake and combine it with health, and attention lands where it changes the outcome. This one change usually does more for retention than any new tooling.
Two quarters of renewals in view, sorted by ARR at risk, reviewed weekly for fifteen minutes. Not a quarterly forecasting exercise.
A second platform rarely fixes a forecasting problem
The instinct is to buy a system that produces the forecast. But the forecast was never missing — the agreement about what at-risk means was missing, and so was the habit of looking early. A new platform that holds a copy of your data adds a sync to monitor and a second version of the truth, and the underlying disagreement survives the migration.
If the contract data, the health judgement and the renewal view sit in one place — the place your team already works — the forecast stops being an exercise and starts being a by-product of how the team operates.
What to do with a bad forecast
A forecast showing 18% of next quarter at risk is not a failure of Customer Success. It is the first useful quarter you have had, because now the number arrived early enough to do something about it. The failure mode is the forecast that looked fine until the month it did not.