# Sales win rate in Pipedrive: the version you can trust

https://salesbond.app/blog/sales-win-rate-pipedrive/
Published: 2026-10-01 · Updated: 2026-10-25

Sales win rate is the number of deals you won divided by the number you won plus lost in the same period. Open deals are not in the bottom of the fraction, because they have not decided anything yet. A quarter with 30 deals won and 70 lost has a 30% win rate, however many deals are still sitting in the pipeline. The same goes for deals still “nurturing”, and for anything counted by a date other than the day the deal was actually marked won or lost.

Your win rate is probably wrong before it is low.

The formula is simple. The trouble is that a small team’s CRM holds deals closed lost just to tidy the board, deals reopened after a loss, and placeholders created “just in case”, and each of those moves the number. Open deals stay out on purpose. They have no outcome yet, and putting them in the denominator drags the rate down for no reason other than timing. The rest of this piece covers which definition to pick, how to slice it so it tells you something, and what has to be true in Pipedrive before you show the number to anyone.

## Three ways to calculate win rate, three different numbers

Take one hypothetical quarter for a five-person team. During the quarter they created 150 deals. In the same quarter, 100 deals reached a decision: 30 won, 70 lost. Of those 70 losses, 25 were closed from the very first stage. They were leads nobody ever had a real conversation with, a couple of duplicates, and placeholders someone created after a trade show.

Where the line sits between a lead and a deal is a separate question, covered in [leads vs deals in Pipedrive](https://salesbond.app/blog/pipedrive-leads-vs-deals/).

Three people on that team can report three win rates and all of them will be “correct”.

One: won ÷ all deals created in the quarter. 30 ÷ 150 = 20%.

Two: won ÷ deals won or lost in the quarter. 30 ÷ (30 + 70) = 30%.

Three: won ÷ decided deals, leaving out the never-qualified. 30 ÷ (30 + 45) = 40%.

The first one mixes two different groups of deals. Some of the 30 wins were created last quarter, and plenty of the 150 new deals are still open and will close next quarter. It also punishes the team for adding pipeline: create 50 more deals in the last week and the rate drops, even though nothing was lost. Some teams call this first number a close rate and keep “win rate” for the second. Use whichever names you like, as long as each number has its own.

The second is our default and the one in the first paragraph. It counts deals by the date they were marked won or lost, not by the expected close date a rep typed in months ago.

The third is a fair measure of how well the team sells once a deal is real, and we like it for that. It only works if the exclusion is mechanical. “Lost from the first stage” or “lost with the reason Not qualified” is a rule. “The rep decided it didn’t count” is not, and a win rate built on that drifts up every quarter without anyone selling better. Tenders are worth their own line: a no-bid is a deal you chose not to play, not a loss, so leave it out of the denominator and read the [RFP win rate](https://salesbond.app/blog/rfp-bid-no-bid-pipedrive/) separately.

Pick one, write it down, and keep it for at least a year. If a report shows two of them, label both. And if you later turn win rates into stage probabilities for a [weighted pipeline forecast](https://salesbond.app/blog/weighted-pipeline-forecast/), use the same definition there, or the forecast and the win-rate slide will disagree in the same meeting.

The quarterly review is where this number gets read; here is [how a small team runs that meeting from five Pipedrive numbers](https://salesbond.app/blog/sales-qbr-small-team/).

## Cut it four ways, and know when a slice is too thin

A team-level 30% is an average, and averages hide the useful part. We cut it four ways, in this order.

Source comes first because it changes planning. In our hypothetical quarter, inbound deals won 16 of 40 decided (40%), outbound won 9 of 45 (20%), and referrals won 5 of 15 (33%). Same team, same product, and the blended 30% describes none of them. If next quarter’s plan leans on outbound, the planning number is 20%, not 30%. Planning on the blend means booking pipeline that was never going to close. The referral slice grows only if someone asks, so we wrote down [how to ask customers for referrals and track them](https://salesbond.app/blog/ask-customers-referrals-pipedrive/).

Deal size is the second cut. Split your decided deals into three bands based on your own won-deal values, small, middle and large, and calculate each band separately. A 30% blended rate means something quite different if the large band wins at a third of the rate of the small one. That is usually a qualification or pricing conversation, and its worth having before anyone asks for a bigger quota.

For recurring contracts, base the bands on [annual contract value](https://salesbond.app/blog/acv-vs-deal-value-pipedrive/).

The third cut is the stage the deal was in when it was lost, and it is the one we’d look at first if we had to pick. In the example the 70 losses split like this: 25 at the first stage, 18 at Discovery, 19 at Proposal, 8 at Negotiation. Assume nobody skipped stages. Then 57 deals reached Proposal (30 won plus 19 and 8 lost later), so the win rate from Proposal is 30 of 57, about 53%. From Negotiation it is 30 of 38, about 79%. The stage-to-stage conversion from Proposal to Negotiation is 38 of 57, about 67%.

![Hypothetical quarter: deals lost at each of four stages (25, 18, 19, 8), win rate from Proposal 53%, from Negotiation 79%](https://salesbond.app/wp-content/uploads/2026/10/sales-win-rate-pipedrive-2.png)

Those last two numbers answer different questions, and it helps to keep them apart. Stage-to-stage conversion tells you where deals leak; win rate from a stage tells you how likely a deal sitting there today is to close. [ORM Tech](https://orm-tech.com/blog/how-to-calculate-win-rate-by-stage) draws the same line between them. For this team the reading is plain: almost as many qualified deals die at Proposal as at Discovery, so the next thing to open is the lost reasons on those 19 deals. How to make that list readable is in our piece on [lost reasons in Pipedrive](https://salesbond.app/blog/pipedrive-lost-reasons/), and if your stages don’t have clear entry criteria the stage cut will be noise, which is where [defining your sales pipeline stages](https://salesbond.app/blog/sales-pipeline-stages/) comes in.

Rep is last, and it’s the cut people want first. A rep with 8 decided deals and 3 wins is at 38%. One more loss and they’re at 33%; one more win and they’re at 44%. A single deal moves the number by four to seven points, which is roughly the whole gap people argue about in a rep ranking. It is also [why we do not rank reps by win rate](https://salesbond.app/blog/sales-leaderboard-small-team-pipedrive/) on a leaderboard. Our working rule is arithmetic, not statistics: we don’t read a slice until one more win moves it by less than three points, which at a 30% rate takes about 25 decided deals. On a small team that often means reading reps over two quarters or a full year. The same rule applies to every other cut, and by it the referral slice above, with 15 deals, is still a hint rather than a finding. The same arithmetic is behind a common slump verdict for reps with [ten to twenty deals a quarter: too early to say](https://salesbond.app/blog/sales-slump-small-team-pipedrive/).

Each cut also points at a different fix, which is the honest answer to “how do I raise my win rate?” A pile of losses at the first stage is a qualification problem, and the fix happens before the deal is ever created. Losses at Proposal send you back to the lost reasons. Losses right after the first call can send you to a [win rate by pitch version](https://salesbond.app/blog/sales-pitch-small-team-pipedrive/), if the team tracks one. A wide gap between sources is a planning question: put next quarter’s pipeline targets on the sources that close, instead of asking the team to push harder where deals don’t.

## What your Pipedrive data needs before the number is honest

Pipedrive already calculates a win rate for you, in Insights. Create a report, choose Deal and then Conversion, and you get two types. The funnel conversion report shows your overall win rate at the top and, for each pipeline stage, the share of deals that reached it. The win/loss conversion report puts deal owners on the columns, so the rep cut is there out of the box.

Both start from the same default filters: deals created this year, with status Won or Lost. That is a cohort by creation date, not the second definition above. Change the date filter to close time and you have our default; set View by to source channel and you have the source cut. Measure by can be the number of deals or the deal value, and those are two different win rates, so say on the slide which one it is. One quirk of the funnel report matters for the stage cut: Pipedrive treats a won deal as having completed every stage, wherever it was closed, while a lost deal stops at the stage where it was lost. ( [Pipedrive’s guide to deal conversion reports](https://support.pipedrive.com/en/article/insights-reports-deal-conversion) has the full list of filters.)

None of those reports can fix teh data underneath. Most of the work is here, and this has to be true in your Pipedrive data first:

- Dead deals are closed as lost, not left open. An abandoned deal stays out of the fraction, so the win rate comes out higher than it should. See [when to close a deal as lost](https://salesbond.app/blog/when-to-close-a-deal-lost/) and the difference between [stalled and dead deals](https://salesbond.app/blog/stalled-deals-vs-dead-deals/).
- Every lost deal has a lost reason, including “Not qualified” and “Duplicate”. Those two reasons are what make the third definition mechanical instead of a judgement call.
- Board clean-ups get their own lost reason. A rep who closes a pile of old deals on the last Friday of the quarter drops them all into one quarter’s denominator. Tag them so you can report with and without them.
- The created date means something. A deal entered weeks after first contact, or a batch of placeholders created on the day of an import, breaks the first definition and any cycle-length number you calculate later.
- Reopened deals count once. Decide whether a deal lost in March and won in May belongs to the period of its final outcome, and write the decision next to the win-rate definition.
- Deals move through stages instead of jumping from the first stage to Won. Otherwise the stage-of-loss cut describes the data entry habits, not the sales process.

If you already track [CRM data quality metrics](https://salesbond.app/blog/crm-data-quality-metrics/), empty lost reasons and deals open long past their expected close date are the two to look at right before a win-rate report goes out.

## What is a good win rate? Why we don’t quote benchmarks

A good win rate is one calculated the same way as last quarter and moving up, with one of the four cuts explaining why; a benchmark range that doesn’t state its definition can’t tell you that. Many pages that rank for this question open with a range of what a “good” win rate is, and we deliberately don’t print one. The example above produced 20%, 30% and 40% from the same quarter and the same team, and a benchmark without its definition, its period and its deal mix can’t be compared with any of the three. A team that sells mostly through referrals and a team that lives on outbound will land far apart for reasons that have nothing to do with with how well either of them sells. If you want an outside reference, the useful one is which of your won deals you’d like more of, and [building an ICP from your won deals](https://salesbond.app/blog/ideal-customer-profile-won-deals/) starts exactly there. The same goes for people: don’t rank reps on 8 deals, whatever the table says.

## Put the definition where everyone can see it

Before the next pipeline meeting, write one sentence at the top of the win-rate report, something like: “Win rate = deals won ÷ deals won or lost, counted by the date they were marked won or lost, excluding deals lost as Not qualified or Duplicate.” Then leave it alone for a year. The first quarter-on-quarter comparison you make with it will be the first one that means anything.