sales forecast
Also called: bottom-up sales forecast, deal forecast
An estimate of what will be signed or billed in a period, built bottom-up from existing accounts, named open deals and a realistic allowance for new ones. A stuck deal counted at full value is the most common way a forecast lies.
A sales forecast estimates what will be signed or billed in a period. A reliable one is built bottom-up: the base from existing customers (renewals you can see), named open deals, and, if needed, a modest allowance for deals not yet in the pipeline. Each named deal gets a probability from the team's own history for its stage, not from the owner's gut feeling.
It helps to give three numbers rather than one. Commit: what you would bet on, typically renewals and live deals at the contract stage. Best case: commit plus live deals at the proposal stage. Expected: the base plus every live deal weighted by the pooled conversion from its stage to won. Stale deals (no buyer activity for a month, close dates moved twice) are taken out or listed separately, and deals that cannot sign in the period are moved to the period when they can.
A forecast is not a target. When it falls short, the useful reactions are to add pipeline, unblock specific deals or change the plan; raising probabilities until the number reaches the target only hides the gap until the quarter ends. Review the forecast weekly, and compare each past forecast with what actually closed to learn how far off your method runs.
Example
At the review on 5 October 2026, Klasna for Teams' deal owners multiplied each deal by their own gut probability and got 10,614,240 UAH for the fourth quarter. Of that, 6,144,000 (57.88%) came from Veloria, Cobalt and Orchid, none of which could sign before 2027, and 2,344,800 came from seven stale deals.
Rebuilt bottom-up:
The gut forecast was 10,614,240 ÷ 1,587,439.56 ≈ 6.7 times the expected one. Method matters even inside the honest version: with averaged instead of pooled stage rates the expected forecast would be 1,798,634.29, about 211,000 UAH higher for no reason in the data.
Common mistakes
- Gut-feel probabilities. Owners' optimism compounds across deals; use stage conversion from your own history.
- Stale deals at full value. Take them out or list them separately until the buyer moves again.
- Dates the buyer never agreed to. A deal that cannot sign this quarter belongs in the quarter when it can, whatever its size.
- Averaged stage rates. Averaging per-quarter rates over-weights small quarters; use pooled conversion.
- Editing the forecast to meet the target. Raising a probability to close the gap hides the problem instead of solving it.
Learn it in the course
- Forecasting bottom-up: key accounts, renewals, new business · Build a sales forecast in three parts: key accounts judged one by one, the base of renewals, and new deals weighted by pooled stage conversion; then commit, best case and expected.
- Case: the quarter's forecast review · Thirty open deals and an owner-feel forecast of 10.6 million: move the key accounts, flag the stale deals, weight the rest by pooled conversion and give Marta commit, best case and expected.