How to build a sales pipeline that forecasts accurately
A wrong forecast is not a reporting problem, it is a definition problem. When a stage means something different to every rep, the sum of probabilities cannot mean anything at company level.
This guide shows how to rebuild a pipeline around verifiable criteria and how to measure forecast quality month by month.
Rebuilding it in 5 steps
1Cut stages down to six at most
Beyond six stages, reps pick approximately. A working B2B set: qualified → need confirmed → quote sent → negotiation → won/lost. Each stage describes what the customer did, not what the rep feels.
2Define observable exit criteria
"Need confirmed" means the approximate budget, the decision maker and the timeline are in the system. If you cannot show the evidence in the CRM, the opportunity does not advance. This single rule fixes most pipeline inflation.
3Weight on history, not intuition
Compute the actual conversion rate from each stage to won over the last 12 months. Those are your probabilities, not the software defaults. Recompute quarterly.
4Use time in stage as a signal
An opportunity stuck three times longer than the stage average has a far lower real probability than the theoretical one. Flag it automatically and force a decision: advance or close.
5Compare forecast with actuals every month
Record the forecast at the start of the month and actuals at the end. Track the percentage error. Under 15%, the forecast can support stock and hiring decisions; above 30%, the pipeline is still declarative.
Three common mistakes
- Opportunities left open forever because nobody wants to report a loss
- Estimated values entered once and never updated after negotiation
- A forecast built only on the weighted sum, with no committed minimum scenario
What to report to the board
Three numbers: committed value (opportunities in negotiation with a confirmed date), weighted pipeline value, and coverage - pipeline over target. Below a 3:1 ratio the quarterly target is at risk regardless of effort.
Aktualizacja:
Najczęstsze pytania
How many stages are too many?+
More than six. Each extra stage adds ambiguity and admin time without improving forecast accuracy.
What forecast error is acceptable?+
Under 15% over 30 days is a good B2B result. Over 90 days, under 25% is realistic for 2-3 month sales cycles.
Should reps set probability manually?+
No. Probability is derived from the stage and historical data. Left to the rep, it becomes an internal negotiation tool rather than a measurement.
Powiązane pojęcia ze słownika
Gdzie sprawdza się CRMconnect
Chcesz zobaczyć CRMconnect w działaniu?
Spersonalizowane demo w 30 minut - bez karty, bez zobowiązań.
Umów demoPowiązane poradniki
How to measure CRM ROI: the formula and the numbers to track
A ROI formula for a B2B CRM, with the four measurable sources of gain and a method for setting the baseline before implementation.
CRMHow to choose a CRM for a B2B sales team
The criteria that decide a B2B CRM: field adoption, forecast quality, stock and contract pricing at quoting time. Includes the questions to ask in a demo.