Before the board starts relying on your renewal forecast, war-game the model. Separate renewals, contraction, logo loss, and expansion so one cannot hide another. Calibrate each probability weight against the last four quarters of outcomes. Cap expansion that has no live conversation behind it, and show a range with the assumptions that move it. Paste the model, the renewing accounts, historical outcomes by stage, and your cohort rules into one request. Instruct Pingpong to flag weights that do not match history, expansion counted without evidence, large accounts hidden inside averages, and cohort rules built on an old customer mix. Review your written answers in a second round.
This page is for chief executives, finance leaders, and customer leaders preparing a retention forecast the board will use to judge the plan. What to do after a quarter misses is covered in stress-test a forecast miss review before the board hears it. Executive starting points live under Pingpong for executives.
Four movements, kept apart
- Gross renewal: how much of the renewing base comes back at all, before any upsell. Directors ask about this first.
- Contraction: customers who stay but buy less.
- Logo loss: customers who leave entirely, counted separately from contraction so a single large departure is visible.
- Expansion: new seats, products, or usage from existing customers. The easiest to inflate.
A single net retention figure can look healthy while gross renewal quietly erodes. Show the four, then the net.
Rules that cap the optimism
- Weights come from history. If renewals marked 80 percent likely closed at 65 percent over the last year, the model uses 65.
- Expansion counts only with an active conversation, a named contact, and a date. Everything else goes in an upside line.
- Risk signals lower the weight automatically: a departed champion, a sharp drop in usage, an open escalation.
- Any account above a set share of renewing revenue is listed by name, with its own assumption.
- Cohort rules match the customers you have now, and say when they were last checked.
Customer teams resist the first rule more than any other. A historical rate can feel like an accusation to someone who knows the customer. Showing it by segment and stage, so each team sees its own history, helps.
A worked example
This example is illustrative and does not describe a customer. A software company with $40 million in annual recurring revenue has $18 million up for renewal over the next two quarters. The model forecasts 118 percent net retention on that base, and the board plan depends on it.
The finance and customer leaders paste the model, the account list, four quarters of outcomes, and the cohort assumptions.
A useful pass finds that the model uses new-business stage weights for renewals, overstating late-stage renewals by about 12 points. Expansion makes up most of the gain, and 60 percent of it sits on accounts with no expansion conversation open. One customer worth 8 percent of the renewing base is marked 90 percent likely, though its executive sponsor left last month and usage is down by a third. The churn rate comes from a year dominated by large annual contracts; today's renewing base leans toward smaller monthly ones. Gross renewal is never shown on its own.
The revised model shows gross renewal at 89 percent, against an implied 94 in the original. Weights are recalibrated by segment. Expansion without an open conversation moves to an upside line. The large account drops to 50 percent and appears by name, with the customer leader owning a recovery plan. Net retention becomes a range: 99 percent low, 106 percent base, 112 percent high, with the three assumptions that move it written next to each figure. The chief executive asks the board to plan on the base case.
What the board needs before trusting the number
Directors want gross renewal by itself, the largest accounts by name, last year's forecast against actuals, and the two or three assumptions that would change the answer most. Offer them before anyone asks. If customers are already leaving and the question is what to do about it, see war-game a churn response before the board asks why.
A request you can copy
Below are our renewal forecast model, the renewing accounts with detail, four quarters of renewal outcomes by stage, and our cohort rules. Show gross renewal, contraction, logo loss, and expansion separately. Compare each probability weight with historical outcomes. Flag expansion counted without an active conversation, and large accounts with risk signals. Check whether cohort rules fit the current customer mix. Suggest a low, base, and high case with the assumptions that drive each. Stop there so we can answer in writing, then review our answers.
How Pingpong runs the review
The web review app sends your request through several models in order. Each later model receives the original request and every earlier answer, with instructions to assess the work so far. For the steps inside the app, see running your first review.
What a model can't review
A model cannot call your customers or sense how a renewal conversation went. Pingpong checks whether the forecast's assumptions match the evidence you give it.
When the decision goes to the board, see war-game a decision before the board meeting. More guides live under work decisions before you commit.