Sales Pipeline Stages: Exit Rules That Show a Deal at Risk
Sales pipeline stages are the steps a deal moves through from first conversation to signed contract, each one opened by something the buyer has done. Every stage needs an exit rule that someone off the call can check in the CRM, plus a time limit. A deal that overstays its limit, or advances on seller activity alone, is at risk well before its close date slips.
Open a package's tracking page the day you order. It says "label created", which means the seller printed a sticker. The box hasn't moved. Every status after that changes only when a scanner at a depot reads the barcode, so each line on the page is a fact somebody else recorded. When a package sits at one depot for six days, you know it's lost before the delivery date passes.
Most startup pipelines read like that first line. A deal moves to "Proposal sent" because the rep sent a proposal, and to "Negotiation" because the call felt warm. Nothing in the record shows what the buyer did, so the deal looks healthy right up to the day its close date slips.
Two rules turn a pipeline into a tracking page. Each stage opens on something the buyer did, and each stage has a time limit.
Sales pipeline stages defined
A tracking page counts scans, and a pipeline should count buyer commitments. Name each stage after the thing the buyer is doing, so a seller's busy week can't move a deal. Five open stages suit a startup with one or two people selling:
| Stage | Buyer has | Exit proof | Stalls after |
|---|---|---|---|
| Discovery | Taken a call and named a problem | The problem and its cost, in the buyer's words | 11 days |
| Qualification | Confirmed a budget range and who decides | The decision-maker named, and a meeting with them booked | 8 days |
| Evaluation | Started comparing you with alternatives | Their evaluator's follow-up held, with the objections recorded | 23 days |
| Business case | Put the case in front of the decision-maker | The decision-maker's response, and a target signature date | 18 days |
| Negotiation | Committed verbally and started on terms | A signed contract, or a named reviewer with a date | 15 days |
Closed won and closed lost sit after these, and a lost deal always carries a reason. A common rule sets each stall limit at one and a half times a typical stage length, counted from the day a deal enters the stage. Evaluation usually takes about 15 days, so it stalls after 23. Treat those lengths as a starting guess, then replace them with your own once a few quarters of deals have closed.
These are deal stages. The stages before a deal exists, from first contact to a qualified lead, belong to MQL vs SQL and to HubSpot lifecycle stages.
Exit rules anyone can check
A depot scan works because nobody has to ask the driver how the trip felt. An exit rule should pass the same test: someone who wasn't on the call can confirm it from the CRM record alone.
"Demo went well" fails that test. "The buyer's technical lead booked a follow-up for Thursday" passes it. The first is a mood, and the second is an event with a date. Write at least two rules of the second kind for every stage.
Exit rules also make a pipeline teachable. A new rep can't learn a stage from your feeling about it, and a written rule is the first page of the playbook a founder handing off the sale has to write.
When the buyer goes quiet mid-stage: leave the deal where it is and start the clock. Moving it forward on a promise hides the stall, and moving it back early writes off a deal that might still be alive.
Signs a deal is at risk
With buyer-action stages and exit rules in place, risk shows up as data before anyone admits it in a meeting. These signs come first:
| Sign | Cause | Fix |
|---|---|---|
| Past its stall limit | The buyer stopped acting, or the stage was entered too early | Review it this week, and ask what the buyer would need to act |
| Close date pushed once | Something changed on the buyer's side | Tie the new date to a buyer event, such as a board meeting |
| Close date pushed twice | The deal was never what the stage said | Qualify it again from the start |
| Late stage, seller activity only | Emails and proposals sent, nothing back | Move it back to the last stage the buyer earned |
| No decision-maker by Business case | Qualification was skipped | Get the meeting, or mark the deal at risk |
| Next step reads "follow up" | Nobody owns a specific action | Write one action, one owner and one date |
Silence has its own limits. Give a deal a senior review after 30 days with no contact the buyer started. After 60 days, take it out of the active pipeline until the buyer comes back. CRM data hygiene covers the rest of the rules that keep dead deals out of the record.
Reading the stages together
A single deal at risk is a problem for this week. A stage where deals keep stalling is a problem with the process. Count how many deals move from each stage to the next, and the lowest rate shows where the pipeline breaks. That stage is rarely the one that feels most dramatic.
Deals often break a stage before they show it. A run of losses at Business case with no decision-maker named points back to Qualification, so fix the rule where the gap opened. Pipeline velocity combines win rate, deal size and cycle length into one number and points at the constraint worth fixing first.
When stages can't help: a pipeline shows where deals stop, and it can't make buyers want the offer. If deals stall evenly across every stage, look at the offer and the market before you redesign the pipeline.
Setting up your stages
A tracking page earns your trust because every line on it came from a scan. Rebuild your stages the same way, each one named for a buyer commitment and closed by rules someone else can check. Then go through every open deal and ask what the buyer did to earn its stage. Expect some to move back.
I've built a pipeline from nothing, as the first sales rep at Everee: $430K from 89 sales-accepted opportunities, 90% of it outbound. Building the stages, the exit rules and the CRM fields behind them is part of my fractional CMO work, alongside the sales playbook your first rep will read. The free diagnostic there gives you a written roadmap either way.
Let the buyer move the deal.
Common questions
How many stages should a startup's sales pipeline have?
A startup's sales pipeline works best with five or six open stages, plus closed won and closed lost. One or two salespeople can keep five stages honest. Stages are easy to add later and hard to merge once months of deal history sit in them.
Can a deal move backward in the pipeline?
A deal can move backward in the pipeline when the evidence for its stage disappears, such as a champion leaving or a budget being pulled. Moving it back keeps the forecast honest. Record the reason on the deal, so the history shows why it moved.
Should small deals and large deals use the same pipeline stages?
Small deals and large deals can share a pipeline with two tracks. A small deal with one decision-maker skips the business case stage. A deal with several stakeholders or a security review runs every stage, and a small deal that picks up either one moves to the full track.
What win probability should each pipeline stage carry?
The win probability for each pipeline stage starts as an estimate that climbs stage by stage, such as 10% at discovery and 90% at negotiation. Once a company has a few quarters of closed deals, it should replace each estimate with the share of deals that actually closed from that stage.