Stages and probabilities you can defend

Why probabilities matter

Pipedrive multiplies every open deal's value by its stage probability to produce the weighted pipeline. That number flows into the revenue forecast, into Insights reports and into whatever your management reads on Monday morning. If the probabilities are guesses, the forecast is a guess wearing a suit.

The failure mode is predictable. Someone sets 20, 40, 60, 80 because the numbers climb nicely. Real pipelines do not climb nicely. In most B2B accounts we audit, the true conversion from the proposal stage is nowhere near the 60 percent the account has it set to, and the forecast has been flattering everyone for years.

Probabilities also shape behaviour. A rep who knows stage entry moves the forecast treats stage changes seriously. That only works when each stage has a clear entry test, which is why this guide leans on the exit criteria from pipeline design.

The design decisions that matter

Decision one: derive, do not vote. For each stage, count deals that entered it over the last twelve months, then count how many of those eventually won. That ratio is the probability. The deal conversion report in Insights gives you the funnel per stage, so the math takes an hour, not a week.

Decision two: allow deal-level overrides, but treat them as exceptions. Pipedrive lets you enable deal-specific probability per pipeline. Use it for the odd deal with a signed letter of intent, not as a daily mood dial.

Decision three: set rotting limits per stage at the same time. Rotting is configured in the same pipeline editor, in days per stage. A sensible default is twice the median time a won deal spends in that stage. The deal duration report in Insights gives you the median.

Decision four: decide what a stage skip means. Deals that jump from Qualified straight to Verbal agreement are either brilliant or fictional. Agree that skipped stages still count as entered for conversion purposes, and let the numbers absorb reality.

A worked example

A 14-person software services team, one new business pipeline, twelve months of history, 214 deals entered stage one. Here is what the derivation produced.

Two things stand out. The old guessed value for Proposal sent was 60 percent, against a real 42. And Verbal agreement was set to 100, which the three deals that died on paperwork politely disproved. After correction, the weighted pipeline dropped by about a quarter. Painful for a week, then useful forever.

Common mistakes

The classic is the 50 percent parking lot. One stage carries half the pipeline because nobody knows what else to do with those deals. The fix is process, not probability: enforce the exit criterion, and push stalled deals to lost with a reason. Your hygiene routine should catch the rest.

Second, silent recalibration. Someone changes probabilities without telling the team, and Monday's forecast moves with no deal having changed. Announce every recalibration with the date and the evidence, or trust in the number evaporates.

Third, mistaking weighted pipeline for a forecast. Weighted value is an average across many deals. On small deal counts it produces nonsense, because you cannot win 42 percent of one proposal. The forecasting guide covers when to trust which number.

Maintenance

Put a quarterly recalibration in the calendar. Pull the deal conversion report, compare actual conversion per stage against the set probability, and adjust anything more than ten points off. Log each change in a note the whole team can see.

Watch rotting compliance weekly, not quarterly. Rotting deals show up red in the pipeline view and can be filtered. A pipeline with a fifth of its deals rotting is not a probability problem. It is a coaching problem, and the reporting setup should make it visible before it becomes a forecast problem.

Questions