Callback Cohorts and Return-Cost Review
Reconcile original jobs and linked returns, compare matched follow-up horizons, and calculate affected-job frequency, repeat visits and observed cost.
Method and interpretation
All dollars are USD. A job is mature if completion date + selected calendar-day horizon ≤ observation cutoff. Listed jobs must fall within the cohort’s completion dates. Only mature jobs and their returns from completion through horizon end enter matched-horizon rates and costs. Events for incomplete jobs and events after the horizon stay in separate buckets with their cost totals. The horizon is a user assumption, not a warranty term.
Rates and cost
Affected-job rate = distinct mature jobs with included returns ÷ mature jobs. Events per 100 jobs = included events ÷ mature jobs × 100. Additional repeated events = included events − distinct affected jobs. Zero denominators are unavailable; zero observed events in a mature population is a measured zero for that population and window.
Gross event cost = paid person-hours × loaded hourly cost + parts + other nonlabor cost + refunds. Net observed cost = gross − credits actually received. Pending recovery is not deducted. Credits above gross remain a flagged negative result. Loaded labor allocation does not establish a cash saving or future reserve. Do not count the same travel, labor or parts twice.
Worked example
100 mature jobs, eight affected jobs and twelve included returns produce an 8% affected-job rate, twelve events per 100 jobs and four extra events beyond each first return. Eighteen hours × $60 + $600 parts + $240 other cost = $1,920 gross. Received credits of $300 give $1,620 net: $16.20 per mature job, $202.50 per affected job or $135 per event. All dates, rates and costs are hypothetical.
Validation and limits
Use valid input dates from 1900–2200, a whole horizon of 1–730 days, up to 1,000 listed jobs and 2,000 events. References are unique ignoring case; each event must match a listed job and occur between completion and observation. No negative/nonfinite costs are accepted. A complete-population control difference is displayed, not silently allocated. Cost completeness is your assertion. CSV requires the exact example header; malformed or oversized files retain the existing rows. Import needs review and recalculation.
Historical observed cost does not predict future warranty claims, establish contractual liability, prove workmanship or rank employees. Context: ASQ cost-of-quality categories. Use the companion guide for evidence and closing treatment.
Primary references
This numerical framework is explained in the companion guide. Contract, accounting and transfer conclusions require the actual evidence and appropriate review.
Make the next step a considered one.
Talk through the business, your objectives and the records that can make the next decision clearer.