How to calculate pipeline velocity

It is 3:20 p.m. on the last Thursday of the quarter. A sales manager opens the board and sees 48 open deals. The total value is roughly twice the target, so on paper nothing looks broken. Yet the team has closed only four deals in the past three weeks. The manager says the pipeline is full; finance asks when the cash actually arrives. Both statements are true, and neither answers the same question. What is missing is not the amount in the pipeline but how fast that amount is moving.

This article puts a number on that movement. We start with the pipeline velocity formula and its four variables, then cover where each variable comes from in a CRM, how the math looks across two example quarters, how pipeline velocity differs from sales velocity, why cycle length is the most contested input, which stage is actually slowing you down, which of the four levers to pull first, the mistakes that quietly corrupt the calculation, why a single average misleads, and how to fold the measurement into a weekly rhythm.

What does the pipeline velocity formula actually measure?

Pipeline velocity tells you how much revenue your open deals generate per day. The standard calculation multiplies three things — the number of qualified open opportunities, the average deal value, and the win rate — and divides the result by the average sales cycle in days. The output is not a total; it is a tempo. It answers the sentence "our pipeline produces roughly this much per day."

That per-day unit is what separates velocity from every other pipeline number. Total pipeline value takes a photograph; velocity plays the film. The same $2M in open deals belongs to a completely different company at a 45-day cycle than at a 120-day cycle. The first company closes the quarter calmly. The second one spends the last two weeks negotiating discounts it did not plan to give.

The practical value comes from the decomposition. A single score tells you nothing actionable, but because the score is built from four parts, you can isolate which part broke. If velocity fell, either fewer deals entered, deals got smaller, the win rate slipped, or the process got longer. They rarely deteriorate together — usually one of them drags the rest.

One caveat worth stating early: this is not a benchmarking metric. A team selling enterprise software and a team selling consumables cannot compare their numbers in any meaningful way. The only honest comparison is against your own history, your own segment, and your own team's previous quarter.

Where does each of the four variables come from?

The credibility of the whole calculation rests on how these four fields are defined. "Number of deals" means opportunities that sit in the pipeline during a defined window, have passed your qualification bar, and are genuinely active. Unqualified interest, records nobody has touched in six months, and test opportunities created during onboarding do not belong in that count. Include them and you inflate the numerator, which makes velocity look better precisely when it is getting worse.

Average deal value should come from closed amounts, not from the optimistic figure a rep typed when the record was created. Win rate is calculated on decided deals only — won divided by won plus lost. Put open deals in the denominator and the rate will always look depressed, and the team ends up staring at a number that punishes them for having a healthy pipeline. Running opportunities with disciplined stages and probabilities is the habit that keeps all three of these fields consistent.

Cycle length is the gap in days between the date a deal entered the first stage and the date it closed. The detail that matters most is where the clock starts: at record creation, or at the moment sales accepts the deal as qualified? Weeks can separate those two definitions. Which one you pick matters far less than picking the same one every quarter.

None of this needs to be assembled by hand. The Sales Pipeline Analytics screen already keeps win rate, average time to close, and average deal size in one place, so three of the four inputs read off a single view; your only real decision is which pipeline and which date range you are counting.

How does the math look across two quarters?

Numbers make the formula concrete. The table below compares two invented quarters — read it for the mechanism, not the values. The point is what happens in Q2: pipeline value stays flat, the win rate improves by a point, and velocity still falls, because the cycle stretched.

Variable	Q1 (example)	Q2 (example)	What changed
Qualified open deals	60	58	Essentially flat
Average deal value	$18,000	$19,000	Slightly up
Win rate	22%	23%	Slightly up
Average cycle	45 days	62 days	17 days longer
Velocity per day	≈ $5,280	≈ $4,090	Clear decline

The lesson is that improving three variables did not offset the fourth. Because cycle length sits in the denominator, its proportional weight is larger than the others: seventeen extra days erase the pricing gain and the win rate gain combined. The same discipline applies when reading average deal size — a team that understands how average revenue per account is calculated will also spot the quarters where two outsized deals carried the mean.

Is sales velocity the same thing as pipeline velocity?

In everyday use the two terms are swapped freely, and for most teams that causes no harm, because the arithmetic is identical. The emphasis differs, though. Sales velocity frames the number as the pace of revenue — the commercial outcome. Pipeline velocity puts the emphasis on the pipeline itself: how fluidly deals travel between stages.

The practical translation is simple. Sales velocity is a leadership metric; pipeline velocity is a sales operations metric. One answers "will we make the quarter," the other answers "where is it stuck." We cover the first framing separately in the four levers that accelerate revenue, and reading that alongside the stage-level breakdown here gives you the full picture.

Some teams calculate both deliberately: one company-wide sales velocity, plus a separate pipeline velocity for each pipeline. That split earns its keep as soon as you run separate processes for different product lines, because a 60-day enterprise rhythm and a 10-day SMB rhythm should never be melted into a single average.

Why is average cycle length the most contested input?

Of the four variables, cycle length is the easiest to miscalculate and the easiest to game. The first problem is arithmetic: a mean is dragged by a few long-tail deals. If nine of ten deals close in 30 days and one closes in 300, the mean lands near 57 — a number nobody on the team has ever experienced. That is why the median belongs next to the mean in every report.

The second problem is subtler. Because you only count closed deals, the ones still sitting open and untouched for months never enter the calculation at all. That is survivorship bias, and it makes the process look nimbler than it is. The fix is not to redefine the metric but to track the age of open deals as a separate indicator alongside it.

The third problem is definitional drift. When stage names or qualification criteria change mid-quarter, historical cycle data stops being comparable. Defining stages on paper before building the pipeline exists precisely to prevent that drift. If nobody has written down when a stage counts as complete, you are not measuring the process — you are measuring how reps happen to log their work.

When velocity drops, which stage do you inspect?

Total cycle time looks like one number but it is a sum of stage durations. So "the process got slower" is never an action. The action begins with knowing how many days are spent where. Once you break waiting time down by stage, the problem almost always concentrates in one place — and in most teams that place is not the first meeting, it is the dead zone after a quote goes out.

The Pipeline Velocity view answers exactly this question: it shows which stage deals get stuck in and how many days they sit there, which names the bottleneck instead of describing it. The Deal Health Score works the same information at the record level — every open deal carries a 0-100 score, and 14 days of inactivity, 30 days of stagnation, or a close date now in the past all pull that score down. The vague observation "the pipeline feels slow" turns into a list of twelve deals that have not moved in two weeks.

Post-quote waiting is especially treacherous because nothing is ever declared lost; the record simply stops. The Approval Queue Report collects every sent but unresolved quote in one queue and shows how many days each has been waiting, which drags that silent zone into daylight. The same logic extends downstream — teams that have built out the full path from quote to cash can tell whether the delay lives in selling or in approval and contracting without arguing about it.

The step after finding the bottleneck is reminder discipline. Automatic Reminders raise a notification or open a follow-up task on their own when a deal has sat in the same stage longer than the number of days you set. It is the cheapest way to shorten a cycle: no new headcount, no new spend — forgotten records simply stop being forgotten.

Which of the four levers should you pull first?

The four variables do not cost the same to move. Adding deals usually costs budget. Raising average value requires pricing and packaging decisions. Lifting the win rate takes coaching and stricter qualification. Shortening the cycle, by contrast, is mostly process work. That is why the first move is almost always on the cycle side: it pays back fastest and requires no additional spend.

Lever	Typical move	Time to effect	Risk
Cycle length	Reminders and ownership rules on the stuck stage	Weeks	Rushing past qualification
Win rate	Raise the qualification bar, work the loss reasons	A quarter	Deal count drops
Deal value	Bundle, add cross-sell	A quarter	Cycle gets longer
Deal count	Channel and campaign spend	Months	Acquisition cost rises

The risk column is the reminder that these levers push against each other. Average deal value should grow through bundling rather than discounting, because revenue bought with a discount comes straight out of margin; a team that knows how gross profit margin is calculated sees that trade before making it, and does not buy speed with profitability.

Adding more deals is the most visible lever and the most expensive one. More records mean a longer queue for the same team; a longer queue means fewer touches per deal, and the cycle lengthens on its own. So before increasing channel spend, measure where your customer acquisition cost currently sits. If acceleration is being purchased with expensive volume, the gain never leaves the dashboard.

What quietly corrupts the calculation?

The first corruption is mixed currency. Averaging deals opened in different currencies into one total means the number shifts every time rates move. On the Multi-Currency side, 16 exchange rates are pulled daily from the central bank and locked at the moment the record is saved, so a past quarter's average does not silently rewrite itself later and period-over-period comparison stays meaningful.

The second is keeping dead records in the pipeline. Deals with close dates long past, untouched for months, inflate both the deal count and the pipeline value when nobody clears them out. Velocity reads that bloat as good news, because it enlarges the numerator. Cleaning the board at quarter close is the most basic step toward an honest metric.

The third is the win rate denominator, covered above. The fourth is reading a single score instead of a trend. One quarter of decline can be seasonal; three consecutive quarters of decline is structural. Velocity should always be read next to the coverage ratio your target requires, because however fast the pipeline moves, a pipeline that is too thin will still miss the number.

Why does a single average mislead?

A company-wide velocity figure is usually the average of business lines that have nothing in common. Put 90-day enterprise deals and 15-day SMB deals in the same bucket and the resulting number describes no team's reality. Without segmentation, the metric gives leadership false comfort and gives the field an unfair scoreboard.

Three breakdowns carry almost all the value: segment, source, and rep. Segment shows product-market fit, source shows marketing quality, rep shows where coaching is needed. The Lost Source Matrix completes the third dimension by crossing loss reason with the source a deal came from, producing an actionable sentence — "paid leads fall out on price, referrals stall on timing" — instead of a vague channel verdict.

What makes segmentation sustainable is not rebuilding a report every month. Saved Board Views let you name and store pipeline, owner, stage, and attention filters, so the whole team looks at the same number under the same definition. Stage design is part of that consistency too, and starting from a proven pipeline template removes most of the definition repair work you would otherwise do six months later.

How do you fold this into a weekly rhythm?

Pipeline velocity is useless as a quarter-end report, because by the time the quarter ends the window for intervention has closed too. A workable rhythm looks like this: the weekly meeting looks at exactly two things — the list of deals aging in a stage, and the last four weeks of the velocity curve. The monthly meeting discusses win rate and average value. The quarterly review revisits stage definitions and the qualification bar.

Team composition belongs in that rhythm as well. A new rep's first deals reliably take longer to close than a veteran's, so when hiring or attrition is high, the average cycle can stretch without the process breaking at all. Tracking the effect of turnover in a sales team separately keeps you from blaming the process for something staffing caused.

Finally, set the target correctly. "Increase velocity by 20%" is not a target. "Cut average waiting time in the quote stage from eighteen days to ten" is a target. The first produces debate; the second produces tasks. That is where stage-triggered automation and reminders earn their place: the goal attaches to the process rather than to a person. If you want to try the measurement on your own pipeline, open a free account and start by calculating the last ninety days of a single pipeline.

FAQ

How often should I calculate pipeline velocity?

Monthly for decisions, weekly for monitoring. A weekly figure captures short-term swings but is noisy on its own, so real decisions should follow a four-week moving average. Stage-level waiting time, on the other hand, can be reviewed every week because the response to it is immediate.

I have very few deals. Is the formula still useful?

With low deal counts, win rate and average value swing hard, so a weekly number is mostly noise. Calculate quarterly instead, and use the median cycle rather than the mean. It is also worth flagging any quarter where one large deal carried the average, so nobody reads it as a trend.

Should lost deals be part of the cycle calculation?

They belong in the win rate, since the denominator is every decided deal. For cycle length, keep two numbers: average days to win and average days to lose. Losing early is good news; losing late is the most expensive form of process cost you can carry.

Does discounting make sense as a speed tactic?

It shortens the cycle now and charges you twice later: average deal value drops, and the buyer expects the same discount on the next purchase. A discount is a closing tool, not a velocity tool. Once it becomes systematic, it erodes the numerator of the formula — which is the part you actually get paid on.