How to test a sales forecast in one week

Most forecasts survive the management presentation and fail the first quarter after close. They fail because nobody tested them against the history that already exists in the CRM. That history is usually enough to tell you, in about a week, how much of the plan is built on pipeline and how much on hope.

This is the sequence we use. It needs one export and a few careful hours per step.

What do you need from the target?

One read-only export of all opportunities for the last eight quarters. Each record needs created date, close date, stage history, amount, owner, source, segment, outcome and loss reason. Add the forecast snapshots management used, weekly or monthly, and the stage definitions with their exit criteria.

If the target cannot produce stage history, that is already a finding. A forecast without history cannot be audited, and a management team that cannot audit its own forecast is guessing with confidence.

Does the pipeline cover the plan?

Pipeline coverage is open pipeline divided by the target still to close. Many teams plan for three times coverage. The test is not whether management says they have it. The test is whether the records support it after you remove deals with no activity in 60 days and deals past their close date.

An example: management claims 3.0 times coverage on a CHF 6 million quarter. After removing stale and overdue deals, the real figure is 1.8. That gap has to be explained by a very high win rate or it is a forecast problem.

How old is the pipeline?

Stage ageing shows how long deals sit in each stage compared with the stages' historic median. Deals that sit well beyond the median close less often. A pipeline full of old deals looks large and converts poorly.

Plot the age distribution of the current pipeline against the age at which won deals closed. If most of the current pipeline is already older than the typical winner, it is inflated.

How often do deals slip?

Count deals whose close date moved out more than twice. Slippage is the most honest measure of forecast quality, because it records every time someone said "this quarter" and was wrong.

Look at it by rep and by segment. One team with chronic slippage is a coaching problem. Slippage across the board points to qualification or to a definition of stage that is too generous.

Did past forecasts come true?

Compare each quarter's forecast, taken at the start and at the midpoint, with what closed. Accuracy within about 10% for a quarter is good. Bias matters more than noise: a team that is always 15% optimistic is easier to correct for than one that swings both ways.

Then rebuild the next four quarters bottom-up from current pipeline, historic conversion by stage and age, and the new-business run rate. Put that line next to the management line. The distance between them is the number the committee needs.

Key takeaways

• One export with stage history is enough to audit a forecast.
• Test coverage after removing stale and overdue deals.
• Compare the age of current pipeline with the age of past winners.
• Count slipped deals by rep and segment.
• Rebuild the forecast bottom-up and size the gap against management.

This is one of ten areas a GTM due diligence scores. To see how each is rated, try the scorecard.

Frequently asked questions

What if the CRM data is messy?

Messy data is a finding, not a blocker. We report how much of the pipeline could not be tested and treat unverifiable pipeline as unproven.

How many quarters of history do you need?

Eight is ideal, four is the minimum. Fewer than four cannot separate seasonality from trend.

Is this useful after close?

Yes. The same test gives a new owner a baseline and shows where forecasting discipline needs to improve in the first 100 days.