Sales Velocity: The Formula, and the Averages Hiding Inside It
Sales velocity multiplies the number of opportunities by average deal value and win rate, then divides by average sales cycle length, producing revenue per day. Three of the four inputs are averages, so a figure blended across segments with different deal sizes overstates the motion and describes neither segment.
Key takeaways
- The formula never observes revenue; it models revenue per day from four averages, so it earns its place only through the decomposition.
- Blending the calculation across motions with different deal sizes attaches large deal values to small deal counts and overstates the result.
- Cycle length sits in the denominator, so closing losses out faster raises velocity while leaving actual revenue completely unchanged.
- The four levers are coupled in practice, and moving upmarket raises deal value while lengthening the cycle and lowering the win rate together.
Sales velocity is a single figure describing how much revenue a sales motion produces per unit of time, built by multiplying the number of open opportunities by the average deal value and by the win rate, then dividing the result by the average sales cycle length in days. The output is expressed as dollars per day, and it is used to compare periods, teams or segments on one axis instead of four.
The formula is popular because it decomposes. Any change in the output can be traced to one of exactly four inputs, which makes it look like a diagnostic. Most of what is worth knowing about the metric concerns what those four inputs are averages of.
The four inputs and where each one comes from
Number of opportunities. How many qualified opportunities are open, or were created, in the period being described. The choice between those two matters and is usually left unstated.
Average deal value. Total value of those opportunities divided by their count, taken either from won deals historically or from the values sitting on open records.
Win rate. The share of opportunities that ends in a closed deal, which carries its own denominator problem into the formula. The win rate entry works through why four defensible denominators produce four different numbers here.
Sales cycle length. The average number of days from opportunity creation to close, sometimes across won deals only and sometimes across everything that closed.
- Step 1Opportunities
How many are open or were created, over a period somebody has to define
- Step 2Average deal value
Total value over count, drawn from won history or from open records
- Step 3Win rate
The share that closes, carrying whichever denominator was chosen upstream
- Step 4Cycle length
Average days to close, divided into the first three to give revenue per day
The first input hides a fork that is almost never declared. Counting opportunities currently open describes a stock, a thing that exists at a moment. Counting opportunities created during the period describes a flow, a thing that happened over an interval. Dividing a stock by an average duration produces a rate that assumes the stock replenishes at exactly the speed it drains, which is a strong assumption about a pipeline nobody has checked. Two teams using the two conventions will report velocities that cannot be compared, and neither will be able to say why the other's number looks wrong.
Note what the formula does not do: it never observes revenue. Actual revenue per day is closed revenue divided by elapsed days, and any finance system can produce it directly. Sales velocity is a model of that quantity assembled from four averages, and it earns its place only when the decomposition tells you something the direct measurement cannot.
The blind spot: three of the four inputs are averages over a population that is not uniform
Deal value, win rate and cycle length are all means. A mean is only informative when the underlying distribution has a single centre, and a sales pipeline carrying two genuinely different motions does not.
Take an illustrative company running a mid-market motion and an enterprise motion side by side. Suppose the mid-market side carries eighty opportunities worth ten thousand dollars each, wins one in four, and closes in sixty days. Suppose the enterprise side carries twenty opportunities worth a hundred thousand dollars each, wins one in five, and closes in a hundred and eighty days.
Computed separately, the mid-market motion produces about three thousand three hundred dollars a day and the enterprise motion about two thousand two hundred, for a combined figure near five thousand six hundred.
Computed as one blended pipeline, the same company reports a hundred opportunities at an average value of twenty eight thousand dollars, a blended win rate of just under one in four, and an average cycle of eighty four days, which multiplies out to eight thousand dollars a day. That is close to half as much again as the sum of the two real motions, and no input was fabricated to get there.
The distortion has a mechanical cause worth understanding rather than memorising. Averaging deal value across the whole pipeline attaches enterprise-sized values to mid-market opportunity counts, and the multiplication then applies that inflated average to all hundred deals. Any blended velocity across motions with different deal sizes will overstate in the same direction, and the overstatement grows as the gap between the segments widens.
- Mid-market: eighty opportunities at ten thousand, one in four, sixty days
- Enterprise: twenty opportunities at a hundred thousand, one in five, a hundred and eighty days
- Combined output near five thousand six hundred dollars a day
- Each number describes a real population
- Tells you which motion to invest in
- A hundred opportunities at an average of twenty eight thousand
- One blended win rate just under one in four
- One blended cycle of eighty four days
- Output of eight thousand dollars a day
- Describes neither motion, and overstates both together
Where else the formula misleads
It rewards shortening the cycle even when the shortening comes from losing faster. Cycle length sits in the denominator, so anything that reduces it raises velocity. Killing weak deals earlier reduces it. So does abandoning long, valuable pursuits that would eventually have closed. If cycle length is measured across everything that closed rather than across wins only, a quarter spent aggressively closing out losses will show rising velocity alongside flat revenue, and the formula will read that as progress.
The four levers are not independent, and the multiplication implies they are. Moving upmarket raises average deal value and almost always lengthens the cycle and lowers the win rate at the same time, because larger deals carry more stakeholders and a procurement process. Loosening targeting raises opportunity count and lowers win rate. Tightening qualification raises win rate and lowers opportunity count. A plan built by improving each term separately is a plan built on an assumption the business does not satisfy.
Open-record values are forecasts wearing the costume of data. Where average deal value is drawn from open opportunities rather than from closed history, it inherits every optimistic number a seller entered and never revised.
At individual seller level it is mostly noise. All four inputs are averages, and a seller carrying a handful of open opportunities supplies a sample too small for any of them to stabilise. One large deal landing moves that seller's average deal value, win rate and cycle length at once, in the same direction, and the multiplication compounds all three movements into a velocity swing that looks like a transformation in performance. Reading that swing as information about the person produces the behaviour you would predict, which is opportunities created and held open to manage the shape of a ratio.
It is a rate, so it says nothing about capacity. A motion can have excellent velocity and still miss the number, simply because there are not enough opportunities in it. Velocity divided into a target gives days required, which is the only form in which it answers a planning question, and pairing it with pipeline coverage is what turns it from a comparison into a forecast.
- Yes: Split it wherever deal sizes differ by an order of magnitude
- Yes: Say whether cycle length covers won deals only or everything closed
- Yes: Say whether deal value came from closed history or open records
- Yes: Check whether a fall in cycle length came from faster wins or faster losses
- Depends: Treat a lever moving alone as suspicious rather than as a result
- Depends: Read it beside absolute closed revenue for the same period
Reading it well
Sales velocity is at its most useful as a comparison between two things that are genuinely alike: the same segment across two quarters, or two territories running the same motion into the same market. It is at its least useful as a company-level headline, where the averaging problem is at its worst and the number is least likely to describe anything real.
The most valuable question the decomposition supports is which lever is cheapest to move. For a company selling into a market where deal sizes are set by what the buyer is, deal value is close to fixed and the honest choices are opportunity count and win rate. For a company with abundant opportunities and a poor conversion, more volume makes the number worse rather than better. Working out which case you are in is worth more than any improvement in the headline figure, and the answer usually comes from segmenting the pipeline rather than from studying the total. Where the answer is deal size, how channel mix changes with deal size covers what has to change alongside it, and where opportunities should and should not sit governs whether the opportunity count is even honest.
On the opportunity-count lever specifically, our own position is narrow and worth stating in our own voice. We send one message per campaign, built on one premise, and a later approach is a separate campaign with its own reason to exist. That constrains how quickly opportunity count can rise, and it pushes the work upstream into who gets contacted, because there is no second attempt to compensate for a badly chosen audience. In velocity terms the trade is deliberate: fewer opportunities entering, at a higher win rate, which moves the numerator through the term that does not drag the cycle out behind it.
There is one form in which the metric answers a question directly, and it is worth using more than the headline. Divide the period's target by the velocity figure and the result is the number of selling days the current motion needs to deliver it. Compare that against the days actually remaining and the gap is immediate, arithmetic and hard to argue with, which is more than a dollars-per-day figure achieves sitting on its own in a report. It also fails honestly: when the required days exceed the days available by a wide margin, no adjustment to the inputs is going to close it inside the period, and the useful conversation moves to the following one.
The last thing to hold onto is that a velocity figure is only as trustworthy as the least examined of its four inputs, and in most companies that is cycle length. It is the input nobody owns, the one most sensitive to how stale opportunities are cleaned up, and the only one sitting in the denominator, where a small error moves the answer most. Defining the target list with the arithmetic attached is where the first two inputs are actually set, and what a good meeting costs is the arithmetic that tells you whether buying opportunity volume makes sense at your deal size. Where it does, our pay per qualified meeting offer is built around that unit deliberately.
Frequently asked questions.
Frequently asked questions- How do you calculate sales velocity?
- Multiply the number of opportunities by the average deal value, multiply that by the win rate, then divide by the average sales cycle length in days. The output is revenue per day. Every one of those four inputs needs its population named, particularly whether opportunity count means currently open or created during the period.
- Why does our sales velocity look high when revenue is flat?
- Two common causes. The pipeline may be blended across segments with very different deal sizes, which inflates the average value applied to every opportunity. Or cycle length may have fallen because losses were closed out faster rather than because wins arrived sooner, and cycle length sits in the denominator where any reduction raises the result.
- Should sales velocity be measured per rep?
- Rarely usefully. All four inputs are averages, and a seller carrying a handful of open opportunities gives none of them enough data to settle. A single large deal moves deal value, win rate and cycle length at once, and the multiplication compounds all three into a swing that looks like a change in performance.
- What is a good sales velocity figure?
- The question does not travel between companies, because the number is denominated in your own deal sizes and cycle lengths. The useful form is comparative: the same segment across consecutive quarters, or two territories running an identical motion. Dividing a target by the figure gives selling days required, which is the version that supports a decision.