# Average deal size in Pipedrive: why the median tells you more

https://salesbond.app/blog/average-deal-size-median-pipedrive/
Published: 2026-10-07 · Updated: 2026-10-07

Average deal size is the value of the deals you won in a period divided by how many deals that was. It is the right number for quota and pipeline arithmetic, and the wrong number for describing a typical deal, because one large deal pulls the average up while the deals your team actually closes stay the same. So calculate two numbers from the same list of won deals: the average, for the math, and the median (the middle deal when you sort them by value), for what a normal deal looks like. In Pipedrive the average is two Insights measures divided by each other; the median is a sorted list or a spreadsheet column.

## The quarter when average deal size rose 46% and nothing changed

Beatrix runs a team of six. For a year they closed 40 deals between $3,000 and $12,000, which comes to $280,000 and an average of $7,000. Then one of her reps won a $140,000 contract with a company that had found them through a partner. Good news and it was celebrated as good news.

Three weeks later the founder opened the dashboard, saw average deal size at $10,244, and started planning as if that were the new normal: fewer deals needed per quarter, a higher price for the standard package, a bigger target for the new rep. The team had not changed what it sold or whom it sold to. Forty-one deals, $420,000, and one third of that revenue from a single contract. The median of those 41 deals was $7,000. The deal in the middle of the list looked exactly like the deal in the middle of the list the year before.

Nobody did the math wrong. An average does what averages do. The mistake was reading one number as if it answered two different questions. All figures in this article are a worked example, not customer data.

## Why the average misleads: five deals

Average deal size = total value of won deals ÷ number of won deals. That is the whole formula, and it is also the whole weakness.

Take five won deals: $4,000, $6,000, $7,000, $9,000 and $90,000. The average is $23,200. The median is $7,000. Not one of the five deals is anywhere near $23,200, and four of the five are within $3,000 of the median.

Now swap the $90,000 deal for an $11,000 one. The average drops to $7,400 and the median stays at $7,000. The median barely notices the outlier; the average is governed by it. Prospeo makes the same point for enterprise deals in a mid-market quarter and recommends the median for that reason ( [Prospeo: Average Deal Size](https://prospeo.io/s/average-deal-size)). We would go one step further. Don’t pick one. Keep both, and treat the gap between them as information.

When your deals are roughly the same size, the two numbers agree. They disagree when you have outliers or when two different kinds of business are sitting in one list. A manager should know about either, and the gap is the cheapest way to find out.

## What goes into the list

Won deals only, and a short window. Our rule is the last 90 days for a number you look at monthly and the last 365 days for anything you plan on, like quota or pricing. “All time” is the default in a lot of dashboards and it is almost never what you want: a team that sold $4,000 deals three years ago and sells $9,000 deals today should not be planning on $6,500.

Four things to exclude before you divide anything:

- Open and lost deals. Average deal size describes what you closed. Open deals carry the value a rep typed in on day one, and lost deals carry whatever was left when they were marked lost.
- Deals with a value of zero. A free pilot or a deal someone forgot to price will drag the average down without saying anything about your business. Either give the pilot its real value or leave it out.
- Test and duplicate deals. Every CRM has them, and won is the status where they hurt most.
- A second currency, unless you have converted everything to one. A list with $ and € mixed in has no average at all; see [how totals work with multiple currencies in Pipedrive](https://salesbond.app/blog/pipedrive-multiple-currencies-totals/) before you trust the sum.

Once you have more than about twenty won deals in the window, split the list by segment first and compute both numbers per segment. Rep, product line, company size, source: whichever one you actually make decisions on.

## Where to find average and median deal size in Pipedrive

The average comes from Insights as two measures you divide yourself. We did not find a median in Pipedrive’s help pages for Insights, so the median comes from a sorted list or a spreadsheet. Both take under ten minutes.

1. Open Insights and create a deal report filtered to won deals and the period you chose. Pipedrive’s help lists the **Measure by** options as “Number of deals, Deal value, Weighted value, Custom numerical field, Custom monetary field” ( [Pipedrive: Insights reports](https://support.pipedrive.com/en/article/insights-reports-deal-conversion), updated 2026-09-03).
2. Run it once by **Deal value** and once by **Number of deals**. Divide the first by the second. That is your average deal size for the window. If you want the average of a custom field rather than deal value, that is a different job; we covered it in [custom field averages in Pipedrive Insights](https://salesbond.app/blog/pipedrive-insights-custom-field-averages/).
3. For the median, open the deals list view and apply the predefined won filter from the Filter dropdown, or create one with “Add new filter” ( [Pipedrive: Filtering for won, lost or deleted deals](https://support.pipedrive.com/en/article/filtering-for-my-won-lost-or-deleted-deals), updated 2026-09-03). Narrow it to your window, then sort the list by value. With 41 deals, the median is the 21st row. With an even count it is the midpoint between the two middle rows.
4. If the list is long, choose the columns with the gear icon, then click “…” > Export filter results and pick XLSX or CSV. Only global admins and users whose permission set allows it can export ( [Pipedrive: Exporting data](https://support.pipedrive.com/en/article/exporting-data-from-pipedrive), updated 2026-09-03). In the spreadsheet, put `MEDIAN()` on the value column with `AVERAGE()` next to it, so the two numbers sit side by side. Keep the file: you will want the same list for the segment split.
5. Write both numbers down with the window and the segment. “Average $10,244, median $7,000, won deals, last 365 days, all pipelines” is a number you can reuse next quarter. “$10K” is not.

We checked the Insights overview page as well, and it does not describe how measures are aggregated ( [Pipedrive: Insights](https://support.pipedrive.com/en/article/insights-feature), updated 2026-09-03). So we are not saying Pipedrive cannot show a median. We are saying we did not find it, and the sorted list is quick enough that we stopped looking.

![41 deal cards sorted by value: the median sits mid-list, the average is dragged right by one $140,000 deal.](https://salesbond.app/wp-content/uploads/2026/10/average-deal-size-median-pipedrive-2.png)

## Which number for which job

Beatrix’s founder was not wrong to use the average. He used it for teh wrong jobs. The average is the number that connects revenue to deal count, because revenue is count times average, by definition. Anything that starts from a revenue target and ends in “how many deals” belongs to the average. Anything that starts from “what does one of our deals look like” belongs to the median.

| Decision | Number to use | Why |
| --- | --- | --- |
| Quota and pipeline coverage | Average (trailing 365 days, per segment) | Revenue = deals × average; the math only closes with the mean. [How to set sales quotas](https://salesbond.app/blog/how-to-set-sales-quotas/) walks through it. |
| Forecast and velocity | Average, with the outlier handled | The formula in [sales velocity in Pipedrive](https://salesbond.app/blog/sales-velocity-pipedrive/) needs a mean, but a mean that one deal set is a forecast of that one deal. |
| List price, standard package, minimum deal size | Median | A package priced at the average is priced above what most customers paid. [Minimum deal size for a small team](https://salesbond.app/blog/minimum-deal-size-small-sales-team-pipedrive/) is a median question. |
| “Is this open deal normal?” | Median, per segment | A $40,000 open deal on a team with a $7,000 median is either a different segment or a deal to look at today. |
| Onboarding a new rep | Median | Tell Gideon the typical deal is $7,000, not $10,244, or he will chase the wrong accounts for a quarter. |

One caveat on the first row. The average is right for quota only if the deals that produced it will happen again. Run the quota math on the 41-deal list and a $600,000 target needs 59 deals at $10,244. Run it on the 40 ordinary deals at $7,000 and it needs 86. That is 27 deals, more than half a year of closing for this team, hidden inside a choice of denominator.

If the $140,000 contract came through a channel you can repeat, keep it in and build a segment for it. If it was a one-off, take it out of the list you plan on, and say so in the plan. [Win rate](https://salesbond.app/blog/sales-win-rate-pipedrive/) has the same outlier problem one step earlier in the funnel, for what it’s worth.

## When the gap is the finding

A gap between average and median is not an error to fix. It measures how much of your revenue depends on deals that do not look like the rest. On Beatrix’s list the average sits 46% above the median, and one deal is a third of the year’s revenue. That is a concentration fact, and it belongs in the forecast conversation, which we covered separately in [one big deal in the forecast](https://salesbond.app/blog/one-big-deal-forecast-pipedrive/). Here the point is narrower: the gap is the alarm, and the sorted list is where you look next.

Sometimes the sorted list shows you a second business instead of an outlier. Say twelve deals at $3,500 and twelve at $11,000. The median of all 24 is $7,250 and so is the average. The two numbers agree perfectly, and both are wrong: not one deal in the list is anywhere near $7,250. That is two segments with two prices, two sales cycles and probably two kinds of rep, and the combined number hides all of it. The test is simple. Sort the list and look for a a hole in the middle. If there is one, split the list there and compute everything twice.

The gap also shows up early, before the deal is won. An open deal several times the segment median deserves its own review: a bigger buying committee, a longer cycle than your [typical sales cycle](https://salesbond.app/blog/sales-cycle-length-pipedrive/), more people who can say no, and a value a rep may have typed in from hope. We build SalesBond, a sales management layer on top of the CRM a team already uses; we started with Pipedrive. Deal Health checks every deal in the connected CRM against more than 20 system rules, gives each deal a health score, and shows the specific violations with what to do about them. The oversized deal is the one you want checked on every rule, because it is the one that moves the number. [Check your CRM with SalesBond](https://salesbond.app/) if that review currently happens only when someone remembers to ask.

## Once a quarter, ten minutes

Beatrix now does this on the first Monday of each quarter. Apply the won filter, narrow it to the last 365 days, read the total and the count from Insights, divide. Sort the list, find the middle row. Write both numbers into the same note as last quarter’s, with the window and the segment next to them. Then one question: did the gap between the two get wider or narrower and which deal did that?

Most quarters the answer is boring, and boring is what you want from this number.

So here is the first step, before any process. Pull last year’s won deals, sort them by value, and put the middle row next to the average your dashboard shows. If the two are close, you’re done until next quarter. If they aren’t, you now know which deal to talk about first.