Sales velocity has been explained on more calculator pages than any one person could read. They agree on the formula, they offer a box to type four numbers into, and they stop there. Then a board member asks how you got the win rate, or why the cycle length jumped this quarter, and the calculator has nothing to say.

That question is where the metric lives or dies. Two of the four inputs are shaped by CRM behavior most teams rarely look at, and both CRMs document that behavior in plain sight.

What is the sales velocity formula?

HubSpot’s sales blog gives the equation as Sales Velocity = Number of Opportunities x Deal Value x Win Rate / Length of Sales Cycle, built from four metrics: number of opportunities, average deal value, win rate and length of sales cycle (HubSpot Sales Blog, Sales velocity, May 2025). The result is revenue per unit of time, and the unit is whatever you measured cycle length in. Measure in days and you get revenue per day.

Measure in months and you get revenue per month.

Written out:

Sales velocity = (Opportunities x Average deal value x Win rate) / Sales cycle length

Fill it in with letters before you fill it in with numbers. O opportunities, at an average value of V, winning at a rate of W, over a cycle of L days, produce O x V x W / L in revenue per day. Swap L for months and the same deals produce a much bigger-looking number that means the same thing. State the unit each time you report it, or someone will compare your monthly figure to last year’s daily one.

You build this number yourself from report data, which means four inputs and four places for it to drift. Move any one of them and velocity moves with it. The rest of this paper is about those inputs.

Win rate needs closed deals in the denominator

Two vendor sources define win rate two ways. HubSpot’s sales blog says to “divide the number of sales won by the total number of sales opportunities” (HubSpot Sales Blog, Sales velocity, May 2025). Salesforce’s Trailhead builds win rate on closed opportunities only, filtering an opportunities report to closed deals and adding the summary formula WON:SUM/CLOSED:SUM (Salesforce Trailhead, Improve Sales Performance with Report Formulas). For velocity, use won divided by closed, meaning won plus lost, for the same set of deals.

Read literally, the blog’s definition puts open deals in the denominator. Those deals haven’t won or lost yet. Count them and win rate drops whenever pipeline grows, which tends to be the exact quarter you want velocity to look good. (Explaining to a board that more pipeline made your win rate worse is not a meeting you want to have twice.)

The closed-only version answers a cleaner question: of the deals that reached a decision, how many went your way. That’s the rate you can multiply by a count of opportunities without counting the same open deals twice.

Both CRMs tie the won and closed flags to the deal stage, which makes stage hygiene the real input here.

A deal that died in March but still sits in Proposal isn’t lost as far as either report is concerned. It’s open, it stays out of your closed denominator, and your win rate looks better than your team’s quarter felt. If your stages and forecast categories have drifted apart, the forecast categories glossary covers how they’re supposed to line up.

Which HubSpot and Salesforce fields feed each input?

In HubSpot, opportunities are deals created in your period, deal value is the Amount property, win rate is deals with Is Closed Won true divided by deals where either closed flag is true, and cycle length is Days to close on closed-won deals. In Salesforce, build the same inputs from an opportunities report dated by Created Date or Close Date, with a WON:SUM/CLOSED:SUM summary formula for win rate.

The HubSpot pieces come straight from its property definitions. Amount is “the total value of the deal in the deal’s currency,” and accounts that use more than one currency also get Amount in company currency, which is the one to average if you sell in several currencies. Days to close is “the time between Create date and Close date,” and HubSpot notes you can use it as a filter when building a custom single object report (HubSpot Knowledge Base, HubSpot’s default deal properties, updated September 2026).

On the Salesforce side, Trailhead’s reports module shows opportunity reports carrying standard filters for Opportunity Status (Closed Won or Open, for example) and a date field you can set to Close Date or Created Date (Salesforce Trailhead, Improve Your Data Analysis with Report Formats). Summary and matrix reports take up to five summary formulas (Salesforce Trailhead, Improve Sales Performance with Report Formulas), so the win rate formula fits in the same report you use for the deal count.

InputHubSpotSalesforceWhat to check
Number of opportunitiesDeals with a Create date in the periodOpportunities report dated by Created DateSame cohort as the other three inputs
Average deal valueAverage of Amount (Amount in company currency if you sell in more than one currency)Average of the deal value your bookings reports already useWhether you average won deals or all deals, written down next to the number
Win rateDeals with Is Closed Won true divided by deals with Is Closed Won or Is closed lost trueReport filtered to closed opportunities, summary formula WON:SUM/CLOSED:SUMLost deals still parked in open stages
Sales cycle lengthAverage Days to close on closed-won dealsCreated Date to Close Date on closed-won opportunitiesEdited create dates and close dates stamped during cleanup (next section)

Build it once as a saved set of reports, not a spreadsheet someone rebuilds each quarter. The second version is where the definitions quietly change.

Why is my cycle length wrong?

HubSpot’s Days to close runs from Create date to Close date, and both dates can drift from what happened. Create date “is set automatically by HubSpot, but can be manually edited by users.” When a deal moves into a closed-won or closed-lost stage, “the close date is set to today’s date,” an automation you can turn off in your deal pipeline settings (HubSpot Knowledge Base, HubSpot’s default deal properties, updated September 2026).

Put those two sentences next to each other and the drag is easy to picture. A deal signs, the rep is on to the next one, and the stage doesn’t move until a quarter-end cleanup. HubSpot stamps Close date with the cleanup day, and that deal’s cycle now includes however long it sat waiting for someone to click. Edit a Create date and you stretch or shrink the cycle on that deal by hand, with nothing on the report to show it.

Neither of these is HubSpot misbehaving. It’s doing what its documentation says. The question is whether your team knows the automation is on and whether anyone edits Create date as a habit. Check your pipeline settings before you check the math.

Open deals carry their own version of this. A deal created in an open stage with no close date gets the last day of the month, or your custom default if you’ve set one. Those placeholder dates don’t feed Days to close on won deals, but they do feed your forecast, and they’re one reason a HubSpot forecast comes out wrong.

If you run Salesforce, check what sets Close Date in your org when an opportunity closes, a person or an automation, before you average it.

Close dates that keep moving are worth watching for a second reason. Ebsta and Pavilion’s B2B Sales Benchmark Report 2024, which analyzed 4.2 million opportunities from 530 companies, found that opportunities whose close dates were moved more than 3 times saw win rates reduce by 77% (Ebsta x Pavilion, B2B Sales Benchmark Report 2024). That belongs next to your velocity number, not inside it. A deal whose close date has slipped repeatedly is one to look at before you trust the averages around it.

Opportunities whose close dates moved more than 3 times saw win rates reduce by 77%.

77%lower win rates after more than3 close-date moves4.2 millionopportunities analyzed530companies in the sample
77%lower win rates aftermore than 3 close-datemoves4.2 millionopportunitiesanalyzed530companies in thesample
Opportunities whose close dates moved more than 3 times saw win rates reduce by 77%. (source: Ebsta x Pavilion, B2B Sales Benchmark Report 2024, 2024)
FigureLabel
77%lower win rates after more than 3 close-date moves
4.2 millionopportunities analyzed
530companies in the sample

Cohorts and periods: keep the four inputs from the same deals

Pick one set of deals and run all four inputs from it: deals created in the period, or deals closed in the period. Mixing them, with an opportunity count from deals created this quarter and a win rate from deals closed this quarter, produces a number that describes no real group of deals. Whichever cohort you pick, say so next to the number.

Both CRMs make either choice easy. Salesforce opportunity reports can be dated by Close Date or Created Date. HubSpot gives you Create date and Close date, and on a Professional or Enterprise subscription it also records Date entered and Date exited for each deal stage (HubSpot Knowledge Base, HubSpot’s default deal properties, updated September 2026). That second option matters if your team creates deals long before anyone works them. Dating the cohort by when deals entered your first qualified stage measures the cycle you’re trying to manage, not the time a deal spent as a placeholder.

The two cohorts trade off in opposite directions.

Created in the period. Honest about the deals you generated, but incomplete. Recent deals are still open, so win rate and cycle length fill in over time and early readings look worse than they’ll end up.

Closed in the period. Complete, because each deal in it has an outcome, but it looks backward. It describes deals that started months ago under whatever conditions held then.

Neither is wrong. Switching between them from one board meeting to the next is.

One more cut worth making: if your deal sizes differ widely, run velocity per segment. An average deal value that blends a mid-market motion with an enterprise motion describes neither of them.

What should you check before you show the board the number?

Run four checks before the number goes in a deck: win rate divides by closed deals, cycle length has a stated unit and you know whether close-date automation is on, Create date edits have been spot-checked, and all four inputs come from one cohort, split by segment if deal sizes differ widely.

  1. Win rate denominator is closed deals. Won divided by won plus lost, with stale deals moved out of open stages first.
  2. Cycle length unit is stated, and close-date automation is known. Days or months, written next to the number, plus a yes or no on whether HubSpot is stamping Close date on the day deals move to closed.
  3. Create date edits are spot-checked. Pull a handful of won deals with unusually short or long Days to close and look at their property history.
  4. All inputs come from one cohort. Created in period or closed in period, applied to all four, segmented if deal sizes vary widely.

If the number survives those four, it’ll survive the follow-up questions. If cycle length is the input that moved, why your sales cycle is getting longer walks through the causes, and sales forecast accuracy covers the close-date and stage issues from the forecasting side.

One blended number or one per segment, with the inputs shown

You can report sales velocity as a single number. It fits on a slide, it trends nicely, and the first time it drops, the room will spend the meeting guessing which of four inputs moved. Or you can report it per segment with the four inputs printed underneath, which takes more space and turns the conversation into which input to work on.

Start this week: pull last quarter’s four inputs for your largest segment (opportunities, win rate, average deal size and cycle length), calculate velocity once by hand, and write down which input you trust least. That input is the first one to clean.

The first version is easier to present. The second is the one you can defend when a CFO asks how you got it, and the one that doesn’t fall apart when your board numbers don’t match the CRM. If you want to know whether your pipeline data can support either version, the free CRM diagnostic checks the stage and date hygiene these inputs depend on.