Expansion Forecast Accuracy

Accuracy measured on its own line

Expansion forecast accuracy is the gap between committed expansion revenue and closed expansion revenue, measured as its own line rather than inside a blended number. A company can hit total plan while missing expansion badly, because a strong new-logo quarter covers the shortfall. That masking is the reason expansion accuracy deserves separate tracking. It also changes the response: a new-business miss points at demand and win rates, while an expansion miss points at adoption, product headroom, and whether anyone owned the account.

What good looks like

Pete Furseth of ORM puts typical accuracy on new plus expansion business at around 90%. The cost of that number is high. It takes considerable manual effort to produce, and it is static, so it stops reflecting reality as the quarter moves. ORM targets 95% accuracy without manual adjustments, and that accuracy holds from day 1 to day 90 of the quarter, updating as conditions change. ORM builds the model from a company's own historical sales performance, and a fully trained model takes 4 to 6 weeks.

The distinction that matters is not the percentage. It is whether accuracy on day 1 resembles accuracy on day 80. A forecast that only becomes accurate in the final week has no operational value, because the quarter has already happened by then.

Why expansion forecasts drift

Expansion models decay for a specific reason. They are built on assumptions about how the base behaved historically, and the base changes. ORM's view on why forecasts miss applies directly here: something in the business or the market shifts, the model still runs on old assumptions, and the miss follows. Price pressure compresses expansion deal sizes. Buying uncertainty stretches the time from qualified to closed. A territory change distracts the reps who own the accounts.

Seasonality compounds the drift. ORM observes that Q2 and Q4 typically run stronger than Q1 and Q3, and that the third month of a quarter outperforms the first and second. An expansion model that spreads the number evenly across a quarter will look wrong every month, then correct itself at the end for reasons that have nothing to do with the model being right.

Improving it

Separate the commit into carry-over expansion and in-quarter created expansion, then measure accuracy on each. Most of the error concentrates in the second bucket. Feed the model product usage rather than opportunity records alone, and re-fit it on rolling recent data instead of an annual assumption. Read forecast accuracy for the underlying measurement, and sales forecasting best practices for the process discipline that keeps the number honest.