A conversion window is the allowed interval between a defined entry event and a qualifying outcome. Choose it using the decision and observed journey timing, then compare cohorts at equal maturity. Report the window, identity rules and delayed outcomes so a short cut-off does not silently redefine success.
Define the journey before choosing a duration
Specify what begins observation: an enquiry, trial start, quote or another event. Define the qualifying outcome and whether the unit is a person, organisation or opportunity. A customer can submit several enquiries before one purchase, so conversion at enquiry level differs from conversion at customer level. Record deduplication and re-entry rules before examining which duration gives the most attractive percentage.
The window should reflect the question. A seven-day measure may support a rapid follow-up process, while a ninety-day measure may better describe a long purchasing cycle. Neither is universally more truthful. Inspect elapsed times among observed outcomes and discuss operational constraints, but recognise that missing or immature outcomes can distort that distribution. A window is a measurement choice that should be justified rather than optimised for a headline.
Specify boundaries, identity and credit rules
Write the interval precisely, including whether an outcome at exactly the final timestamp qualifies. State the reporting time zone and whether elapsed time uses calendar days or business days. Decide how outcomes are linked to entries: one outcome may be credited to the earliest eligible enquiry, latest eligible enquiry or a defined opportunity. Avoid counting the same purchase several times unless the metric deliberately measures overlapping opportunities.
Keep conversion measurement distinct from marketing attribution. A purchase within a window demonstrates temporal eligibility under the rule; it does not prove the entry event caused it. Where several channels touch the same customer, attribution requires its own convention. Preserve unmatched outcomes and unresolved identities rather than allocating them through guesswork that changes both numerator and denominator without visible evidence.
Compare completed observation windows
A recent cohort has had less time to convert than an older cohort. Show immature cohorts as provisional or compare both after the same elapsed duration. Provide counts as well as rates, including the number still eligible to convert later. When extending a window, retain the original shorter-window result so readers can distinguish a definition change from additional conversions under the existing definition.
The analytics lifetime reference provides background on entry and subsequent events. The worked case is fictional and does not recommend a particular platform setting. Check whether data retention and identity coverage support the entire chosen duration. A ninety-day metric is not fully observed if the source retains only thirty days of linking information, even when a query returns a neatly formatted percentage.
Include a worked boundary case in the definition: an enquiry at noon with a purchase exactly thirty elapsed days later. That case forces a clear inclusion rule and prevents different query implementations from disagreeing about an otherwise identical customer journey.
Fictional enquiries: seven days versus thirty days
A fictional scenario to explain the method, not a customer result or a market benchmark.
A fictional consultancy receives 100 distinct eligible enquiries during a cohort week. Twelve produce a signed engagement within seven elapsed days, and another eighteen produce one between day eight and day thirty. Under a seven-day definition, conversion is 12 divided by 100, or 12%. Under a thirty-day definition, it is 30 divided by 100, or 30%.
Neither result shows that extending follow-up caused eighteen additional engagements. The window changes which observed outcomes qualify. The report retains both measures and labels the thirty-day figure unavailable for a newly formed cohort that has only ten days of observation.
One customer sends two enquiries before signing once. A reviewed opportunity identifier assigns that engagement once under the stated rule. These invented records illustrate duration, maturity and identity choices. They do not establish the right sales cycle for another business or a causal effect from any acquisition channel.
TAKE THIS INTO YOUR NEXT CONVERSATION
A practical next-step checklist.
- Specify entry, outcome and identity unit.
- Write duration and boundary rules exactly.
- Deduplicate outcome credit deliberately.
- Compare equally mature cohorts and show counts.
Common questions.
Is a longer conversion window always better?
No. It captures more delayed outcomes but may answer a different operational question and complicate attribution.
Can a purchase inside the window prove campaign impact?
No. Temporal eligibility and causal attribution are separate questions requiring different evidence.
Further reading & source context.
These primary references provide background. They are not evidence for the fictional example above and do not endorse Accuralytix.
Found an error or a missing qualification? Send an editorial correction.
Check my data readiness.
Use a practical tool to turn this idea into a clearer starting point.
Check my data readiness