Renewal forecasting needs its own model: a practical revenue guide

A new-business forecast asks whether a buyer will start a commercial relationship. A renewal forecast asks whether an existing customer will continue, expand, contract or leave. The evidence, timing and interventions are different enough that one probability model should not be copied across both motions.

A practical renewal forecast combines authoritative contract data with adoption, value, support, stakeholder and commercial signals. It keeps observed evidence separate from the team’s risk judgment and from the final commit category.

Start with a complete renewal population

The denominator is every in-scope contract approaching a decision inside the forecast horizon. Missing records are not zero risk; they are coverage gaps. Preserve original amount and currency, renewal date, term, product scope and accountable owners. Keep unknown values unknown until verified.

• One stable customer and contract identity.
• Verified renewal, notice and commercial decision dates.
• Current recurring amount in the original currency.
• Products, seats or scope included in the renewal.
• Customer-success, commercial and executive owners.
• Prior commitments, exceptions and open actions.

Use six evidence dimensions

No single health score should erase the underlying facts. Show each dimension and its freshness.

Separate observed risk from forecast category

A risk signal is an observation, such as declining use or an unresolved escalation. A forecast category is a judgment about the likely commercial outcome by a date. The same risk may have different consequences depending on product criticality, sponsor commitment and recovery progress.

Record both views. A customer can be high risk but still committed after an approved recovery plan. Another can look healthy operationally while procurement has not confirmed budget. The forecast should explain the gap instead of compressing it into one color.

Define commit, best case and risk with evidence

Use language the customer-success, sales and finance teams can apply to the same account.

Turn the forecast into a managed action queue

Every exception should have one observation, hypothesis, action, owner role, check date and result. Do not create duplicate tasks for the same underlying blocker. Revisit the existing action until the result is verified.

Review the model for calibration and action

Compare forecast categories with mature outcomes by horizon and segment. Report absolute counts and value by currency; never relabel or combine raw currencies without a verified conversion policy. Review whether interventions completed and whether the underlying evidence changed.

• Population coverage is explicit and missing contracts are visible.
• Evidence dates meet the required freshness window.
• Forecast categories use stable definitions across teams.
• Expansion, contraction and flat renewals remain distinct.
• Actions close only after a verified result.
• Calibration uses mature outcomes and comparable horizons.