Why CRM Probability Percentages Are Usually Wrong
Jun 30, 2026
Most CRM systems assign a probability to every sales stage.
Qualified: 20%
Proposal: 60%
Negotiation: 80%
It looks scientific.
In reality, those percentages are often little more than educated guesses.
If your sales stages weren't built using your own historical conversion data, your forecast is resting on assumptions instead of evidence.
Where CRM Probabilities Come From
Explain that they're usually:
- CRM defaults
- copied from a previous company
- decided in a meeting
- adjusted because they "felt about right"
None of those are evidence.
Why This Creates Poor Forecasts
If Proposal is set to 75% but historically only 38% of proposals close...
Every forecast is inflated.
The CRM isn't lying.
It's doing exactly what you told it.
The Better Approach
Use historical data.
For example:
| Stage | Historical conversion to Won |
|---|---|
| Qualified | 18% |
| Discovery Complete | 31% |
| Commercial Validation | 47% |
| Proposal | 64% |
| Commit | 91% |
Now your forecast reflects reality.
Why Most Companies Never Do This
Because their stages don't have consistent exit criteria.
This links perfectly to the previous article.
If people enter "Proposal" under different circumstances...
Historical conversion becomes meaningless.
Historical Data Is Better Than Opinion
Don't ask:
What probability should Proposal be?
Ask:
What percentage of Proposal opportunities became customers during the last two years?
The answer already exists.
AI Doesn't Fix This
If probabilities are wrong...
AI produces better-looking wrong forecasts.
Garbage in.
More convincing garbage out.
Key Takeaways
- CRM probabilities should be earned, not guessed.
- Historical conversion data is the best source of forecast probabilities.
- Exit criteria make conversion data trustworthy.
- Better data leads to better forecasts.