Your CRM says the pipeline is worth six figures. Your bank account disagrees.
Use six sales pipeline stages: Prospecting, Qualification, Discovery, Proposal, Negotiation and Closed Won, plus Closed Lost. Move a deal only when the buyer does something you can check, and set each stage’s win probability from your own closed deals.
We built that pipeline in a real CRM and ran 300 made-up deals through it. The stages held up. The forecast still overshot by more than half, for a reason most guides miss.
What are the 6 sales pipeline stages?
A sales pipeline is the path a deal takes from first contact to signed contract. Each stage answers one question about the buyer.

| Stage | The question it answers | Exit when the buyer… |
|---|---|---|
| 1. Prospecting | Is this someone worth talking to? | Replies and agrees to a first call |
| 2. Qualification | Is there a real problem, budget and a decision-maker? | Confirms a budget range and who signs |
| 3. Discovery | What exactly do they need, and by when? | Attends a demo or working session and agrees on scope |
| 4. Proposal | Does our offer fit? | Accepts the proposal in principle or asks for the contract |
| 5. Negotiation | What terms will they sign? | Signs the contract or sends a purchase order |
| 6. Closed Won | Did we win? | Nothing left: it’s revenue now |
Closed Lost sits off the line because a deal can die at any stage. When one does, record two things: why, and where it died.
Why six? It’s the smallest set where every stage changes what you do next.
A one-call sale can merge Qualification and Discovery. A deal that needs a security review might add a stage. Past a point, extra stages hurt, and we’ll show you where.
Still picking software? Our walkthrough on how to choose a CRM puts Close, FluentCRM and HubSpot side by side.
What moves a deal to the next stage?
A deal moves forward when the buyer does something, not when you do. That one rule decides whether your pipeline means anything.

Think about the last deal you moved to Proposal. Did the buyer ask for pricing, or did you just send it?
Sending the email, logging the call and sharing the deck feel like progress. None of it proves the buyer moved. Confirming a budget, showing up to the demo or accepting the proposal does.
HubSpot’s default deal pipeline shows how easily this slips. Its open stages are Appointment scheduled (20%), Qualified to buy (40%), Presentation scheduled (60%), Decision maker bought-in (80%) and Contract sent (90%).
Three of those five describe the rep’s calendar or outbox. Contract sent at 90% says nothing about whether anyone will sign. Setting HubSpot up now? Our HubSpot free trial guide covers the free plan.
Our take: keep HubSpot’s structure if you like it, but rename the activity stages to buyer outcomes. Presentation scheduled becomes Demo attended. Contract sent becomes Terms agreed.
We measured the damage. In our test pipeline, we bumped about half the open Qualification, Discovery and Proposal deals one stage, as if a rep had logged a call.
Same deals, same buyers. The weighted forecast rose from $132,475 to $158,975: 20% more pipeline from labels alone.
How we tested these stages (and what was fake)
Real CRM, fake deals. We’d rather tell you that than dress up a model as customer data.
- The CRM: EspoCRM 10.1.1 on our test server, the same install behind our CRM implementation plan. We added a Discovery stage and set 10%, 20%, 35%, 50% and 75% for the five open stages.
- The deals: 300 synthetic deals from a seeded script, worth $2,000 to $15,000 and created over 180 days. Each stage had a set chance of moving forward, a typical time in stage and an 8% chance the buyer went quiet for about two months.
- The snapshot: we froze the pipeline on day 150 and loaded the 57 open deals into EspoCRM. Its weighted total matched our script to the dollar: $132,475.
We wrote the rules behind these deals, so the percentages say nothing about your market. They show the mechanics: how conversion, forecasts and stuck-deal reports behave on a pipeline like yours.
Stage-to-stage conversion: where do deals leak?
Stage-to-stage conversion is the share of deals reaching one stage that make it to the next. Count only won and lost deals, because an open deal hasn’t told you anything yet.
By day 150, 203 test deals had closed and 19 were won. Here’s how they flowed:
| Stage | Deals that reached it | Moved to next stage | Won from here |
|---|---|---|---|
| Prospecting | 203 | 56.7% | 9.4% |
| Qualification | 115 | 61.7% | 16.5% |
| Discovery | 71 | 56.3% | 26.8% |
| Proposal | 40 | 60.0% | 47.5% |
| Negotiation | 24 | 79.2% | 79.2% |
Read the last column first. That’s the stage probability your history supports, and it belongs in your CRM settings.
Ours didn’t match the defaults we’d typed in. Discovery earned 27%, not 35%. Negotiation earned 79%, not 75%.
Then scan the middle column for leaks. Nearly half our Prospecting deals never qualified, which points at the lead source before the sales script.
Why did the weighted forecast still come in high?
Because the open pipeline isn’t a fair sample of your deals. Fast deals close and leave. Slow, quiet deals stay behind, and they win less.
Here’s the 57 open deals on day 150, next to what they brought in by the end:
| Forecast method | Amount |
|---|---|
| Unweighted (every deal at full value) | $424,500 |
| Weighted, our starting probabilities | $132,475 |
| Weighted, probabilities from closed deals | $121,338 |
| Actually won from those 57 deals | $78,500 (8 deals) |
Calibrating helped a little. Even the calibrated number ran 55% high.
Why? In the closed history, 13% of deals had gone quiet at some point. In the open snapshot, 51% had. Probabilities learned from mostly healthy deals got applied to a pile of mostly stalled ones.
We replayed the calibrated forecast 20,000 times, letting each deal win or lose by its odds. Eight runs in ten landed between $89,500 and $154,000. The real $78,500 beat only 4% of them.
Real-world tip: report the forecast as a range, and give stalled deals their own line. A single weighted number hides both problems.
How to spot stuck deals with days in stage
A deal is stuck once it has sat in a stage for more than twice the usual time. Your own medians set the bar.
Ours were 6 days in Prospecting, 8.6 in Qualification, 11 in Discovery, 8.3 in Proposal and 9.2 in Negotiation.
Doubling those flagged 22 of the 57 open deals, $169,500 of pipeline. Three still closed later for $29,000. So a flag means call the buyer, not delete the deal.
Here’s the catch in EspoCRM: there’s no days-in-stage field. It does log each stage change as a timestamped note, so you can work it out. This query listed all 57 open deals in our test:
-- EspoCRM: days each open deal has sat in its current stage
SELECT o.name, o.stage, o.amount,
DATEDIFF(NOW(), MAX(n.created_at)) AS days_in_stage
FROM opportunity o
JOIN note n ON n.parent_id = o.id AND n.parent_type = 'Opportunity'
AND (n.type = 'Create'
OR (n.type = 'Update' AND JSON_EXTRACT(n.data, '$.value') IS NOT NULL))
WHERE o.deleted = 0
AND o.stage NOT IN ('Closed Won', 'Closed Lost')
GROUP BY o.id
ORDER BY days_in_stage DESC;
HubSpot tracks Date entered and Date exited for every stage, but only on Professional or Enterprise plans. On a free plan, a weekly n8n workflow that compares each deal’s last stage change with today does the same job.
Three CRM settings that quietly wreck your forecast
Perfect stages won’t save you if the CRM stores probability badly. We hit three traps in one afternoon.
- A new stage starts at 0%. We added Discovery through EspoCRM’s field manager API. The stage saved, its probability didn’t, and our first Discovery deal went in at 0%. A $5,000 deal counted for nothing.
- New probabilities skip existing deals. After we set Discovery to 35%, that deal stayed at 0% until someone edited it. Change the numbers and the forecast lags until every deal gets touched.
- Hand edits don’t last. We set one deal to 60% by hand. The next API edit, which only changed the description, reset it to 35%. HubSpot does the opposite: edit a probability by hand and it stops following the stage for good.
One more trap hits your loss reports. We imported an $8,000 test deal straight as Closed Lost, and EspoCRM’s funnel report filed it under Prospecting. The same deal moved to Proposal first, then lost, landed under Proposal.
The fix is dull: after any stage or probability change, bulk-update every open deal, then spot-check five.
How many pipeline stages is too many?
A stage is one too many when it holds too few closed deals to give you a believable win rate. Our rule of thumb: fewer than about 30 closed deals in your calibration window is too thin.
Every stage you add slices the same history thinner. With 40 closed deals reaching Proposal, the honest range for its win rate was 33% to 63%. Split Proposal in two and each half sees about 20 deals, so the range stretches to roughly 30% to 70%.
That’s not a probability anymore. It’s a shrug.
Add a stage only when it changes who acts, like a legal review or a paid pilot. Merge stages that happen in the same meeting. Rebuilding as part of a bigger system change? Our CRM vs ERP guide shows where the pipeline ends and order handling begins.
Set up your pipeline stages this week
No new software needed. Just an hour, a few months of closed deals and a spreadsheet.
A pipeline you can forecast from
Five steps, in this order
-
Step 1 of 5
Write one exit rule per stage
Phrase each as something the buyer did: confirmed budget, attended the demo, accepted the proposal. Paste the rules into the stage descriptions so every rep sees them.
-
Step 2 of 5
Set probabilities from closed deals
For each stage, divide deals won by deals that reached it, counting won and lost deals only. Under about 30 closed deals, keep the default and mark it as a guess.
-
Step 3 of 5
Make Closed Lost record why
Require a loss reason, and walk deals through their real stages before closing them so the last stage is right.
-
Step 4 of 5
Flag stuck deals weekly
List every open deal past twice its stage median. Decide each one: call, move the close date, or close it as lost.
-
Step 5 of 5
Re-check every quarter
Recalculate the probabilities, then bulk-update open deals so the new numbers reach the forecast.
Which CRM? For a free start, HubSpot’s free CRM ships with a deal pipeline and editable stage probabilities.
On WordPress, FluentCRM keeps contacts on your own server, and our FluentCRM review says where it’s thin for sales teams. n8n can run the weekly stuck-deal check.
Disclosure: HubSpot, FluentCRM and n8n links are affiliate links. We may earn a commission if you sign up, and it doesn’t change what we recommend.
Frequently Asked Questions
What’s the difference between a sales pipeline and a sales funnel?
A pipeline tracks individual deals and what each one needs next. A funnel shows totals: how many dropped out between steps. Same stages, different view. Sales teams work the pipeline every day, while marketers usually read the funnel, which is what your CRM’s conversion report shows.
Should leads be a stage in my sales pipeline?
No. Unqualified leads swamp your conversion numbers if they sit in the deal pipeline. HubSpot, for example, ships a separate Leads pipeline with New, Attempting, Connected, Qualified and Disqualified. Create a deal only once someone agrees to a first real conversation.
Can a deal move backwards through the stages?
Yes, whenever an exit rule stops being true. If your champion leaves or the budget gets pulled, move the deal back to Qualification instead of leaving it in Negotiation. A backwards move is information. A deal parked in the wrong stage is a forecast error.
Do I need a separate pipeline for renewals or upsells?
Usually, yes. A renewal skips prospecting and discovery, and its odds look nothing like a new deal’s, so mixing the two skews both sets of probabilities. Give renewals a shorter pipeline of their own, such as Renewal due, Terms sent, Renewed and Churned.
What’s a good stage-to-stage conversion rate?
No universal benchmark is worth trusting, because it depends on deal size, lead source and how strict your exit rules are. Compare each stage with its own last quarter instead. A conversion rate that drops five points in a quarter tells you more than any industry average.
How often should I update sales pipeline stage probabilities?
Once a quarter suits most small teams, and only after each stage has around 30 closed deals behind it. Monthly updates on a handful of deals mostly chase noise. Bulk-update open deals afterwards, or the forecast keeps using the old numbers.
How much pipeline do I need to hit my target?
Divide your target by your win rate by value, then add a buffer for stalled deals. In our test, $424,500 of open pipeline produced $78,500 of wins, about 5.4 to 1. Work out your own ratio from closed deals rather than borrowing a rule of thumb.
The bottom line
Six sales pipeline stages cover most small teams, as long as each one ends on a buyer action you can check.
Take probabilities from your closed deals, treat the weighted total as a range, and keep stuck deals on their own line. That’s where the forecast goes wrong.
Start with the exit rules. Write five sentences today, one per open stage, and paste them into your CRM before your next pipeline review.



