Designing pipeline stages people agree on
If two people would put the same deal in different stages, the stages are wrong.
Pipeline stages should be defined by observable buyer actions rather than by seller sentiment — 'demo completed' rather than 'interested'. Five to seven stages is usually right, with written entry criteria so any two people would classify the same deal identically.
Key takeaways
- Define stages by what the buyer did, not by how the seller feels.
- Five to seven stages; more than that and nobody maintains them accurately.
- Write entry criteria for each stage and test them on ten real deals.
- Capture loss reasons — the pipeline's most useful output is why deals do not close.
Take ten open deals and ask two salespeople to assign stages independently. Where they disagree, the stage definitions are broken — and every forecast built on them is fiction.
Buyer actions, not seller feelings
| Weak stage | Better stage | Why |
|---|---|---|
| Interested | Discovery call completed | Something happened that can be verified |
| Hot lead | Site visit booked | A commitment was made |
| Negotiating | Proposal sent | A dated artefact exists |
| Almost closed | Verbal agreement, contract sent | Specific and checkable |
How many stages
Five to seven for most businesses. Fewer and the pipeline tells you nothing about progress; more and salespeople stop updating accurately, which means the extra granularity is imaginary anyway.
Ageing and stalls
Set an expected duration per stage and alert on deals that exceed it. Most deals are not lost at a decision point; they are lost by sitting untouched for six weeks while everyone assumes someone else is on it.
Capture loss reasons
A required, short list on closing a deal as lost — price, timing, competitor, no response, not qualified. Six months of that data tells you more about your business than any dashboard: whether you have a pricing problem, a follow-up problem or a lead quality problem.