Find the Gap in Your Sales Funnel Conversion

Divide the opportunities reaching the next stage by the eligible opportunities in the previous stage, using the same cohort and observation window. Report pending and unknown records separately. Overall funnel conversion uses the starting cohort as its denominator.

Your team booked 100 appointments this month and attended 60. Is attendance 60%? Only if those 60 belong to those same 100 bookings, the attendance window has closed, and your counting rules handle reschedules and cancellations consistently. If some attended calls were booked last month, or some new appointments are scheduled next week, dividing the two monthly totals answers a different question.

Sales funnel conversion is useful when the count and its context stay together. For anyone reviewing their own sales team’s process, the first investigation should establish what the number represents. This guide uses a completely synthetic cohort to show how. It provides no industry benchmark, diagnosis of a real team, or estimate of recoverable revenue.

Define the group before calculating the rate

Choose the unit you are following. A person, an opportunity and an appointment are not interchangeable. One person may have two opportunities; one opportunity may have several appointment attempts. In this example, the unit is an accountable opportunity, counted once at each stage it reaches.

Our fictional cohort contains 200 opportunities assigned September 1–7. We observe each for 14 days after assignment and review the cohort only after all those windows close. Every later count is a subset of the earlier stage, from the same 200 opportunities. The example assumes complete records and one agreed stage definition throughout.

For attendance here, we ask whether a booked opportunity had at least one qualifying attended appointment within its window. A reschedule stays with the same opportunity. A cancelled opportunity with no subsequent attendance stays in the booked count and does not count as attended. A business may also need an appointment-level show rate, but that is a separate metric with different denominators.

Amplitude’s primary documentation likewise makes event sequence and conversion window explicit, and distinguishes overall conversion from conversion between steps. Its product counts users or events; our opportunity-based example adapts the measurement principle to a sales process. Read the funnel documentation.

How do you calculate sales funnel conversion?

The qualified-to-won rate is 14 divided by 42. Overall assigned-to-won conversion is 14 divided by 200, or 7%. Both are valid here; they answer different questions. Calling both “the close rate” invites an argument about performance that is actually an argument about definitions.

The largest percentage drop is from qualified to won: about 66.7% did not progress within the window. The largest count drop is from contacted to booked: 60 opportunities did not progress. Neither observation establishes the cause or automatically decides where to invest. A later-stage gap could reflect fit, timing, offer constraints or incomplete evidence. The table tells you where to inspect records. A drop-off can signal potential revenue at risk, but it does not establish a recoverable amount or verified lost cash.

Keep pending and unknown visible

The clean example hides the difficulty of a live operation. Imagine that 10 of the 100 booked opportunities still have unresolved attendance records. You know 60 attended and 30 did not; you do not know the remaining 10. Report those counts and the uncertainty. A 60% confirmed-attendance share is not the same as a final show rate with complete records.

Also separate missing evidence from time that has not elapsed. An appointment next week is pending, not a no-show. A completed observation window with a missing meeting record is an evidence problem. Silently labeling both “did not attend” makes a reporting failure look like a sales failure.

When the denominator is zero, show “not applicable” or “no eligible opportunities,” rather than 0%. If one rep won 1 of 2 qualified opportunities and another won 9 of 30, the combined rate is 10 of 32, or 31.25%. Averaging 50% and 30% would yield 40%, which gives unequal-sized groups equal weight.

Why do bookings and attendance reports disagree?

Use this synthetic reconciliation before judging a rep. Four opportunities have five appointment records: A attended its first appointment; B rescheduled once and attended the replacement; C was confirmed as a no-show; D is booked for next week. There are four unique booked opportunities, five appointment records and two attended opportunities. The original B appointment is a reschedule, not another prospect or a no-show.

For a question restricted to the three opportunities whose scheduled outcomes are resolved, two attended: 2 / 3, about 66.7%. For the full four-opportunity group, two have confirmed attendance so far: 2 / 4, or 50%, with one pending. These describe different populations. Neither justifies reporting a completed four-opportunity show rate. Record the cutoff and eligibility rule beside the count.

Calendly documents that rescheduling creates cancellation and creation events. Its no-show feature records a separate, correctable status. Those mechanics illustrate why counting scheduling events is different from counting opportunities. They do not establish an active Obavia integration.

Also compare report settings. GA4 distinguishes open and closed funnels and specified event sequences; a skipped step can exclude a user from later steps. GA4 counts users, not sales opportunities, but the diagnostic question transfers: are both reports following the same unit, sequence and window?

Finally, separate source credit from attendance. An ad platform, a CRM owner field and a payment record answer different questions. Google defines attribution models as assigning credit to touchpoints. That does not determine your reps’ commission credit or prove collected cash. Reconcile the records first; investigate the conversation next.

Turn the gap into a bounded investigation

Suppose you choose the booked-to-attended transition. Compare cohorts with the same definitions and enough time to mature. Then inspect the underlying appointment history: lead source, time from booking to appointment, reschedules, cancellations, reminder delivery records and the agreed purpose of the call. Keep the comparison narrow enough to understand the differences.

Read a small set of both attended and unattended opportunities. If you inspect only failures, you cannot tell whether the suspected pattern also appears in successes. Preserve unknowns and avoid assigning blame from the chart alone. A change in lead mix or calendar capacity can coincide with a different rate without proving that a rep’s behavior caused it.

Write the investigation as a question: “Did this cohort receive a clear agenda before the appointment?” Link it to evidence you can actually inspect. If the handoff lacks that context, use the handoff brief to make future records clearer. A subsequent process change still needs its own evaluation; an improved percentage alone does not isolate its effect.

Keep wins, cash and scenarios distinct

A won opportunity in this example means the defined contract milestone was reached. It says nothing about the amount collected. Collection needs its own authorized payment evidence, currency, timing and treatment of adjustments. Do not multiply a stage gap by an average deal value and label the result “lost cash.” At most, an explicitly stated what-if calculation is a scenario whose assumptions need review.

Start the next operating review with one metric written as a sentence: “Of these eligible opportunities, this many reached this stage within this window; these records remain pending or unknown.” Compare the transition across reps with comparable opportunities, inspect the relevant records and conversations, and select one supported next step. Track subsequent change using the same definitions. The chart directs the investigation; it does not supply the cause.