Conversion rate

Editorial approval scope: The team checked current source fidelity, scope boundaries, dates, arithmetic, reader usefulness, practical workflow, and general-information limitations.

A conversion is a named event

The word “conversion” can mean a form submission, answered call, scheduled consultation, conditional onboarding offer, assessment completed, first service, or paid claim. These events answer different questions. A lead-to-consultation rate says nothing by itself about access to care.

Write the measure before calculating it:

conversion rate = unique eligible records completing the defined event by the cutoff / unique eligible records entering the cohort

The numerator must come from the same cohort as the denominator. State whether the unit is a person, household, referral, organization, opportunity, or campaign response.

Lock the cohort and maturity window

Select a cohort by its starting event and date, then allow the agreed exposure window to pass. A seven-day booking measure for inquiries received through June 30 should be calculated only after every included inquiry has had seven full days.

Records still open at the cutoff remain in the denominator when they had a full opportunity to convert. Report them by age and next action. Excluding slow cases inflates the rate and hides operational work.

Deduplicate repeat forms, forwarded emails, and multiple calls according to a written identity rule. Preserve separate referrals when they truly represent different requested services or people.

Keep funnel stages separate

An ABA practice might track:

• inquiry to first human response
• response to completed intake review
• review to conditional onboarding offer
• offer to accepted onboarding
• accepted onboarding to assessment
• assessment to first authorized service

Each transition has a distinct owner and denominator. A payer decision, clinical recommendation, capacity decision, and family choice should remain separately attributable.

The SBA market-research guide recommends examining demand, market size, location, saturation, and pricing. Conversion results become more useful when segmented by service area, requested service, access need, payer route, source, and capacity state. Use groups large enough to protect privacy and support responsible interpretation.

A fictional calculation

A fictional practice locks a cohort of 40 unique inquiries received in June. Every inquiry has a full ten-business-day intake-review window by the reporting date. Thirty-two receive a human response within two business days. Twenty-four complete intake review. Fifteen receive a conditional onboarding offer, and twelve accept.

The practice reports:

• response rate: 32 of 40, or 80%
• review completion rate: 24 of 40, or 60%
• offer rate among completed reviews: 15 of 24, or 62.5%
• accepted-offer rate: 12 of 15, or 80%
• inquiry-to-accepted-onboarding rate: 12 of 40, or 30%

Sixteen inquiries have no completed review by the cutoff. They stay in the original-cohort denominator with status, age, owner, and next action. Calling 12 of 15 the overall conversion rate would erase those records.

Higher is not always better

A high rate can result from narrow intake criteria, misleading promises, easy form completion, or premature offers. A lower rate can follow better routing, honest capacity information, or appropriate referral to another provider.

Pair the measure with response time, wait time, access actions, withdrawal reasons, clinical appropriateness, family choice, complaints, and downstream service outcomes. Review both successful and unsuccessful paths.

Segment without creating false precision

Useful segments may include source channel, geography, payer product, requested hours, setting, language, and access needs. Predefine segments and combine very small groups. A change from one conversion out of two to two out of three is large in percentage points and weak as operating evidence.

Show counts beside percentages. Record changes to forms, staffing, hours, eligibility, advertising, and tracking logic so trend breaks have context.

Marketing claims need a separate review

An internal conversion metric rarely supports a public outcome claim. The FTC advertising FAQ explains that objective claims need a reasonable basis before release and that express and implied messages both matter.

“Most families start quickly” needs a defined population, event, time window, current evidence, and appropriate qualification. A conversion to an administrative step does not prove clinical benefit, coverage, or payment.

Build a reproducible metric record

Keep the metric name, purpose, owner, event definitions, unit, eligibility rule, deduplication logic, source fields, cohort dates, maturity window, exclusions, numerator, denominator, segment, and version. Reconcile the result to the underlying records.

When definitions change, retain the prior version and start a new series. Recalculating history under a new rule may be useful, but label it clearly.

Assign a named data steward to reconcile exceptions and approve each reporting version before leaders use it.

Tie each threshold to a named decision. For example, a response-rate decline may trigger a staffing review, while a fall in review completion may trigger a form or access audit. The metric should direct investigation, not automatically deny requests, change clinical criteria, or prove that a campaign caused the result.