Time to Value: How Long New Customers Wait Before the Product Pays Off

Time to value is the elapsed time between a defined starting point and the moment a customer first receives real value from the product. It is a duration, not a percentage, and it is reported as a median across a cohort rather than as a single account's experience.

The phrase real value is doing the work in that sentence. Logging in is not value. Finishing configuration is not value. Value is the first time the product does the thing the customer bought it to do, with the customer's own data.

For one account: TTV = timestamp(value event) − timestamp(start event). For a cohort: report the median of those durations, plus the share of accounts that never reached the value event.

The start event

Signup, contract signature, or kickoff call. All three are legitimate and all three give different answers. Self-serve products usually start at signup; sales-assisted implementations usually start at signature or kickoff. Pick one, declare it, and label every chart with it.

The value event

An observable, logged event that indicates value was delivered. It should be the same event you use for activation. If the two differ, you now have two definitions of value and neither team will believe the other's numbers.

The clock

Elapsed calendar time, including weekends and waiting periods. In-product time excludes exactly the delays you most want to see: the three days spent waiting for an administrator, or the week lost to an approval.

The summary statistic

Median, with percentiles if you want spread. The mean will describe none of your customers.

These numbers are invented to show the method and are not benchmarks.

A sales CRM defines the start as signup and the value event as the first follow-up sent to an imported contact. A cohort of 200 accounts signs up in March and is observed for ninety days. Of those, 150 reach the value event and 50 do not.

The 150 durations are sorted, and the middle of that list falls at six days, so the median is six days. The seventy-fifth percentile falls at fifteen days, showing the spread. The mean of the same 150 durations is eleven days, dragged upward by a dozen accounts that returned after several weeks.

The honest reported statement is therefore: median time to first value is six days among the seventy-five percent of the cohort that reached first value, with a seventy-fifth percentile of fifteen days and twenty-five percent of the cohort not reaching it inside ninety days. That third clause is what most reports omit, and omitting it is how a struggling onboarding process comes to be described as fast.

Time to value decides where onboarding investment goes and how much it is worth. It is the clearest early predictor of whether a new customer will still be a customer later, because a long, silent gap after purchase is where enthusiasm dies and refunds begin.

It also arbitrates a recurring internal argument. Product teams want to build features; onboarding teams want to hire; sales wants a faster pilot. Breaking the elapsed time into segments settles it with evidence, because the segment consuming most of the duration is almost never the one people assumed.

Silently dropping accounts that never activated

This is the definitional trap of the whole metric. Accounts with no value event do not have a missing duration; they have an unknown one that is longer than the observation period. Removing them makes the median fall exactly when onboarding is getting worse, because the accounts that struggle are the ones being excluded.

Reporting the mean

A skewed distribution summarised by its average produces a figure that no individual customer resembles and that moves for reasons nobody can trace.

Measuring in-product time

Counting only active session time removes waiting from the measurement, and waiting is usually the problem. Elapsed calendar time is less flattering and much more useful.

Blending self-serve with implementation projects

A two-day self-serve onboarding and a six-week rollout in one median describes neither. Segment by motion, plan size, or team size before drawing conclusions.

Redefining the value event to look faster

Moving the finish line earlier, from a real follow-up to a completed import, shortens the metric without helping any customer. If the event changes, the series breaks.

Good is recognisable without any benchmark. The median is short relative to your sales cycle, the gap between the median and the seventy-fifth percentile is modest rather than enormous, the never-activated share is small and falling, and the value event has not been redefined in a year. Customers describe the first week as straightforward rather than as a project.

Bad looks like a short median sitting on top of a large excluded population, a seventy-fifth percentile several times the median, a metric that improves in the same quarter the definition changed, and an onboarding process where the first visible result arrives only after configuration is complete.

Most teams do not need dedicated tooling for this. Two date fields on the account record, a start date and a first-value date, are enough to produce the median, the percentiles, and the never-activated share. Because the fields sit on the record, the accounts that are late become a working list rather than a statistic, which is the only way the number ever improves.

Continue with what a CRM is, CRM for small business, sales automation, lead management software, CRM versus spreadsheets, and the feature overview.