# Sales cycle: how long a deal should sit in each stage, and what to do when it sits longer

https://salesbond.app/blog/sales-cycle-how-long-a-deal-should-sit-in-each-stage/
Published: 2026-11-01 · Updated: 2026-11-13

The deal cycle, more often called the sales cycle, is the time a deal takes from first contact to won or lost. To manage yours, take the deals you won in the CRM and count how many days each one spent in every stage; the median per stage is what “normal” looks like for your team. On review day, look first at deals that have sat in a stage far longer than that median with no next step scheduled.

## Sales cycle vs sales process: length, not steps

Most teams run some version of the same stages: prospecting, qualification, discovery of the buyer’s needs, proposal, negotiation and close, and plenty of people call exactly that list “the sales cycle”. Strictly, the list is the sales process, the steps a deal goes through. The sales cycle is how long those steps take. Two teams can run an identical process and one of them closes in weeks while the other takes months.

We keep the two apart because they are set by different people. The process is a design decision: a manager decides which stages exist and what has to be true to leave each one (if yours are blurry, start with [how to define sales pipeline stages](https://salesbond.app/blog/sales-pipeline-stages/)). The cycle is a measurement. Nobody decides it, you read it off the deals.

So when someone says “our sales cycle is too long”, ask: in which stage? A cycle where most of the days pile up in Proposal points to pricing or approvals. The same total spread evenly is a different conversation, and you can’t have either one from the total alone.

## How to measure your sales cycle by stage in a CRM

You need four dates per deal plus the stage history, and most CRMs keep all of it without being asked.

The created date and the won or lost date give you the total cycle. The stage history, the log of when a deal moved from one stage to the next, gives you days in each stage. The expected close date and the next activity date come in later, for the stalled-deal checks, so pull them in the same export.

Stage history is the part that trips people up. CRMs do keep it, but each one exposes it differently. HubSpot, for example, stores a [Date entered property for each deal stage](https://knowledge.hubspot.com/properties/stage-calculated-properties); elsewhere it may be a history tab on the deal or a separate report. Before you plan the analysis, check that your CRM gives you the date each deal entered each past stage as well as the current one. With only today’s stage and its entry date you can measure open deals but you can’t reconstruct where won deals spent their time. The CRM-specific walkthroughs are in our articles on measuring deal time in [Pipedrive](https://salesbond.app/blog/sales-cycle-length-pipedrive/), [HubSpot](https://salesbond.app/blog/sales-cycle-length-hubspot/) and [Salesforce](https://salesbond.app/blog/sales-cycle-length-salesforce/).

Once the export is in a spreadsheet the calculation is dull. Filter to won deals from a period long enough that one odd deal can’t swing the result. For each stage, take the median days spent there. Median, not average: one deal that sat in Negotiation while the buyer changed CFOs drags an average up and says nothing about the rest.

Then do it again for lost deals, separately. Lost deals often sit longest in the the stage where they died, and comparing the two columns shows you which stage is the quiet graveyard.

![Timeline of one deal from Created to Won with a team-median marker per stage; Proposal runs past its median and is flagged for review](https://salesbond.app/wp-content/uploads/2026/11/sales-cycle-how-long-a-deal-should-sit-in-each-stage-2.png)

## What “too long” means for your team

It means longer than your won deals needed. Nothing else.

We say this bluntly because “what is the average sales cycle?” gets asked constantly, and the answer is no use for running a pipeline. An industry figure averages a freelancer selling a small project with a vendor selling a multi-year contract. Your won deals, in your stages, with your buyers, are the only benchmark that describes teh deal in front of you.

So the threshold is per stage, and it has two lines. The median is the centre: a deal past it is slower than typical and nothing more. The slowest won deal in that stage is the edge. An open deal that has been there longer than any deal you ever won from that stage is in territory where your team has no record of winning. It isn’t automatically dead. But it needs a reason to still be in the forecast, and “the rep says it is fine” is not one.

Write both lines down per stage where the team can see them. A manager who knows the median and the edge for Proposal asks a very different question than one who only knows the deal is “old”.

The same numbers tell you where to start if you want a shorter cycle. Don’t try to speed up everything. Find the one stage where your days pile up furthest past the median and work on why deals wait there: a missing decision-maker, an approval step nobody planned for, a proposal that takes too long to produce. Fixing one stage is a project you can give an owner and a deadline.

## Six signs a deal has stopped moving

Age alone is a weak signal. A deal can sit for weeks in Negotiation because procurement at the buyer is slow and the rep has a meeting booked for Thursday; that deal is fine. What separates slow from stuck is whether anything is still happening and whether the deal’s own data agrees with its stage. These are the signs we look for. They’re also default rules in Deal Health, the product we build, so the day counts below are defaults you can change:

- **No next activity scheduled.** The most reliable sign of the six. A live deal has a next step in the calendar, even if it’s “call back in two weeks”.
- No activity for 14 days (default, configurable). Nothing logged, nothing sent, nothing booked; the deal has fallen off the rep’s desk.
- 21 days in one stage (default, configurable). Past this point the question is why it is late.
- A skipped stage. If a deal jumped from Qualified to Contract, there was never a proposal discussion in the CRM. Either the stages are wrong or someone is pushing the deal to look further along than it is.
- Many steps back. One move back is honesty, but a deal that keeps bouncing between stages is one nobody understands.
- Expected close date in the past. It was supposed to close, it didn’t, and nobody updated the date, so the forecast still counts it.

Any one of these is a reason to ask a question. Two or more on the same deal and it is stalled, whatever its stage says. For the difference between a stalled deal you can restart and one you should close as lost, see [stalled deals vs dead deals](https://salesbond.app/blog/stalled-deals-vs-dead-deals/).

## Slow but alive, or stalled?

Here is the test we use on review day: is there a dated next step that the buyer knows about?

If the rep can say “they’re reviewing the contract, legal call on Thursday, the buyer sent the invite”, the deal is slow and alive. It may be past every median you have. Leave it in the forecast with the date it actually has today. A long cycle isn’t bad in itself. Some buyers take their time, and a cycle with a dated next step at every point is just one of those buyers.

If the answer is “I’m waiting to hear back”, the deal is stalled. Waiting is not an activity. It doesn’t matter that the buyer sounded keen on the last call: a keen buyer who hasn’t scheduled anything has other priorities, and the longer the silence runs the more the deal belongs in a nurture list rather then a pipeline stage.

The data version of the same test is the first sign on the list above. A dated next activity in the CRM is the deal’s proof of life. Its absence is the loudest signal a CRM can give you and the one most teams never look at, because the pipeline view shows amounts and stages and hides the calendar.

## Review day: one question per signal

A pipeline review goes badly when the manager asks “so what’s happening with this one?” about every deal. The reps narrate and nothing changes. It goes well when each flagged deal gets the question its signal implies, and the rep leaves with a dated task.

No next activity: “What’s the next step, and when?” If the rep can’t answer, the task is to find out by Friday.

Silent for two weeks: “When did the buyer last reply to you, not you to them?” Reps count their own emails as activity. A buyer who has gone quiet for two weeks has told you something.

Three weeks in a stage: “What has to be true for this to move to the next stage, and who at the buyer makes it true?” Then check whether that person has ever appeared in the CRM.

Skipped stage: “Walk me through the proposal conversation.” If there was one, log it and move on. If there wasn’t, move the deal back. It makes the pipeline report honest.

Moved backwards more than once: “What did we learn each time it came back?” If the answer is the same thing twice, the deal isn’t qualified.

Close date passed: “New date, and what changed to justify it?” A pushed date without a new reason is a guess. We’d rather see the deal move back a stage than carry a third close date, because the forecast is what suffers, and [revenue at risk in the pipeline](https://salesbond.app/blog/revenue-at-risk-sales-pipeline/) is mostly made of deals like this.

Six questions, each tied to a fact anyone can see in the CRM. After a couple of reviews reps start fixing the signals before the meeting because they know which question is coming.

## Checking every deal against these signs, every day

A disclosure, since this is where a tool comes in: we build SalesBond, a sales management layer on top of the CRM a team already uses; we started with Pipedrive. Deal Health is the part of it that does the review-day sorting on every deal, around the clock, so the Monday meeting starts from a list instead of from a pipeline view.

Deal Health runs more than 20 rules against each deal in the connected CRM, and the six signs above are among them. Others look at the deal’s data (no amount, no contact or organisation, no close date, a close date more than 180 days out), at ownership (frequent owner changes) and at how the deal ended (lost without a reason). The rules are grouped into weighted categories: Activity at 30, Progression at 25, Data at 20, Forecast at 15 and Ownership at 10.

Each deal gets a score from 0 to 100, a weighted average of those categories. At 80 and above the deal is fine, from 50 to 79 it needs attention, below 50 it’s at risk. If fewer than three rules apply, as with very new deals, you see “not enough data” instead of a misleading number. Next to the score are the rules the deal broke, a plain explanation of what lowered it, and a recommended action for each violation.

What it doesn’t do is act for you. The rep or manager resolves, snoozes or ignores each violation, and that decision stays with a person. SalesBond doesn’t create tasks on its own or write to buyers. The review-day questions are still yours to ask. For more on how the score is built, see [what a deal health score measures](https://salesbond.app/blog/deal-health-score/) and [how Deal Health surfaces at-risk deals](https://salesbond.app/blog/deal-health-salesbond-at-risk-deals/).

## Setting your own thresholds, and who sees what

Deal Health’s default thresholds, 14 days without activity and 21 days in a stage, are starting points. If your won deals show that Negotiation normally runs well past three weeks, a 21-day stage rule will flag a big part of your healthy pipeline, and a rule that cries wolf gets ignored fast. The thresholds can be changed, with a preview before you save, and if the change turns out wrong you can roll it back for seven days.

Our advice: set the stage-duration threshold near the edge from your own measurement, the longest a won deal ever spent in that stage, well above the median. The median flags slow deals. The edge flags deals with no winning precedent, and that’s the list you want on Monday morning; its shorter and every item on it deserves a question.

Visibility follows the org chart. A rep sees the scores and violations on their own deals. A manager sees the team’s, with an average score per rep and where each rep’s problems are. A rep who sees “no next activity” on their own screen on Tuesday fixes it on Tuesday, and the Monday review is left with the deals where the fix wasn’t obvious.

SalesBond works with Pipedrive, HubSpot, Salesforce, Zoho CRM, Microsoft Dynamics 365, Freshsales, Close, monday CRM, Copper, Zendesk Sell, Nutshell, Insightly, Attio, Bitrix24 and Kommo (amoCRM). The Sales Dashboard also has a Deal Cycle widget, but the spreadsheet method from earlier works in any CRM, with or without us. Start there: pull your won deals, find the median and the edge for each stage, and compare your open deals with them. If you’d like to see which of your deals fail which checks today, [check your CRM with SalesBond](https://salesbond.app/).
