How to Calculate Sales Cycle Length for Open Deals

Why does cycle length from closed deals come out too short?

Because the calculation only includes deals that finished, and the ones still open are the slow ones. Averaging days-to-close across closed-won opportunities looks reasonable and produces a number that is wrong in a consistent direction. Every deal currently sitting at 200 days is invisible to that average until it resolves. When it finally closes, it joins the pool and drags the number up, which reads as the cycle lengthening when it was always that long.

The effect is large in businesses with long tails. ORM groups opportunities with a machine learning model and predicts a close curve for each group, and those curves run from 1 to 80 weeks. Most of the expectation lands before week 12 and very few groups extend past 52 weeks, which means the tail is thin but real. A closed-deal average deletes that tail entirely.

How do you calculate cycle length including open deals?

Use the cohort method: fix a creation period, then measure what share of that cohort had closed won by each checkpoint, keeping unresolved deals in the denominator.

``` Share closed by day N = Deals from cohort closed won by day N / Total deals created in cohort ```

Pick a creation cohort old enough that most of it has resolved, usually four to six quarters back. Then walk forward and measure the share closed won at day 30, day 90, day 180, and day 365. The curve flattens where deals stop converting, and the point at which it flattens is your real cycle boundary.

This produces a curve rather than a single number, which is the honest output. A single average implies deals close at a consistent pace. They do not. They close in a distribution with a heavy front and a long, thin back.

What does the age profile of open pipeline show?

Whether the pipeline you are looking at still fits the timeline the forecast assumes. The snapshot below is a hypothetical pipeline, with the age-band conversion rates standing in for the ones you would pull from your own closed history. The median open deal in this set sits around day 95. If winning deals typically close near day 76, more than half the open pipeline is already past the window in which comparable deals converted. That is a specific, checkable statement about the quarter, and no closed-deal average will surface it.

The right two columns are where the money is. The $4,300,000 sitting past 180 days converts at 6% and 2% in this example, contributing roughly $194,000 of realistic expectation while occupying 27% of the reported pipeline value. Any pipeline coverage ratio built on the full $15,700,000 is counting that value at face.

How do you handle deals that will never close?

Apply an aging rule and close them out rather than carrying them. ORM uses a 12-month rule for most customers, where meaningful activity means a change in stage, close date, or amount. Across ORM customers, 10% or more of open pipeline has had none of those changes in 12 months.

Those opportunities do more damage than their value suggests. They inflate the pipeline, they stretch every duration calculation that eventually includes them, and they give reps a reason to describe a quarter as covered when the coverage is fictional. Closing them out is a data hygiene action with a direct forecasting payoff.

The absence of change is the signal, not the age by itself. A 14-month-old opportunity with a stage move last week and a revised amount is a real enterprise deal. A four-month-old opportunity with no change since creation is already dead and has not been marked.

What does this change about the forecast?

It changes how much of the visible pipeline you count toward the current period. Across ORM customers, roughly 20% of the value carrying in-quarter close dates on day one of the quarter actually closes inside that quarter. The other 80% moves, shrinks, or dies.

That ratio is only explicable through the age and curve view. A forecast that assumes every in-quarter close date is real assumes an implied cycle far shorter than the cohort data supports. Reconciling the two is the fastest way to improve forecast accuracy, because the correction is arithmetic rather than judgment.

Which threshold should trigger action?

Twice the median cycle for the segment, measured from creation. A deal past that point has left the distribution its comparables occupied. It is not necessarily lost, but it should not be forecast at the same weight as a deal inside the window.

The strongest confirming signal is close-date movement. A rep changing a close date is the clearest indication that a deal is drifting, and a deal that slips from one quarter to the next is less likely to close even when it sits in commit. The earlier signal is the absence of any signal at all: no activity, no field changes, no notes. Track deal slippage against the age bands and the combination identifies the deals to scrub first.

What breaks the calculation?

Reopened opportunities and records created after the deal was already in flight.

Reopened deals restart the clock in most CRM configurations, which makes a nine-month cycle look like a three-week one. Use original creation date rather than any field that resets.

Late record creation is harder to detect. When reps open the opportunity at proposal stage rather than at first qualified conversation, the measured cycle excludes everything that happened before, and the business looks faster than it is. Compare creation-stage distribution across reps, since one team member logging deals at a different point will skew the segment they belong to.