Rinse Mobile Detailing Job-Time Pilot Protocol

Proposed protocol, version 0.1, October 18, 2026. This document is a research design, not a report of findings; no participant data have been collected or analyzed for it.

Purpose

Measure how participating mobile detailers spend time across travel, setup, inspection, cleaning, closeout, and job-related administration. The study should help readers distinguish booked service time, actual elapsed job time, and total person-hours.

The initial output should be a transparent pilot field report, not a national “state of the industry” benchmark. Its value will come from clear definitions and honest observations, including observations that do not support a favorable story about Rinse.

Research questions

• Time allocation: What portion of recorded job-attributed crew person-minutes falls outside hands-on service?
• Estimation: How does actual site elapsed time compare with the site time estimated before the job?
• Variation: How do observations differ across service categories, vehicle sizes, crew sizes, and participant businesses?
• Operational friction: Which recorded reasons accompany delays, scope changes, callbacks, or rescheduling?
• Collection quality: Can operators record the required fields consistently without unreasonable burden?

Do not use this observational pilot to claim that software causes time savings, that one operating policy causes better outcomes, or that participating businesses represent all detailers. Those are different research questions requiring different designs.

Proposed scope and recruitment

Start with a small instrument test involving approximately three to five businesses, then recruit a planning target of 15–25 independent businesses for a four-week observation period. These are feasibility targets chosen for this project, not statistical power calculations or thresholds proving representativeness.

• Eligibility: Businesses performing paid mobile passenger-vehicle detailing, with an operator able to consent to participation and record job timing.
• Recruitment mix: Seek both Rinse users and non-users, several regions, solo and multi-person crews, and new and established businesses. Record the mix actually achieved instead of asserting diversity from the recruitment plan.
• Observation rule: Ask participants to record all eligible scheduled jobs during their agreed window, including cancellations and incomplete jobs. Consecutive recording reduces cherry-picking but does not eliminate self-selection or recording bias.
• Incentives: If offered, compensate participation or collection effort on the same terms regardless of results. Keep any testimonial request separate and optional.
• Disclosure: Explain that Rinse sponsors the work and sells software in this field. Participation does not require purchasing the product.

Keep an invitation and participation ledger so the report can disclose recruitment channels, businesses invited, enrolled, and retained. Do not call a convenience sample a random sample.

Protect participants and customers

Use a written participant information and consent process before collecting records. Have the final process and storage arrangements reviewed for applicable privacy and contractual requirements before launch.

• Exclude customer identifiers: Do not collect customer names, exact addresses, phone numbers, email addresses, VINs, plate numbers, card data, inspection photos, signatures, or customer message contents.
• Use coarse location: Collect broad region or climate category where needed, not an address or exact route. Do not collect precise GPS traces for this pilot.
• Separate records: Store business identity and consent records separately from study rows linked by a study-specific ID. Limit access to the people doing approved study work.
• Minimize financial information: The default study is about time, not revenue. Do not collect invoices or payment screenshots to verify timing.
• Publish aggregates: Do not promise raw job-level public data. Even de-identified timestamps and service patterns can make a small business recognizable.
• Retention and withdrawal: Specify a deletion date, withdrawal cutoff before analysis lock, permitted uses, and a contact before recruitment begins. Obtain separate permission for any attributed quote or case example.

Use coarse categories and remove unnecessary exact dates from the analysis extract. Suppress identifying small cells; a proposed rule is not to publish a subgroup contributed by fewer than five businesses, with additional review where combinations remain recognizable.

Data dictionary

The structure below is deliberately broader than a single check-in/check-out record. The instrument test should determine whether participants can supply it reliably before freezing the main version.

Elapsed time and labor time are different

Two people working together for 90 elapsed minutes contribute 180 person-minutes if both are actively assigned for that entire interval. Do not add overlapping phase durations and label the result elapsed job time.

Within each person's record, minutes should belong to only one phase at a time. Where two people perform different phases simultaneously, add their person-minutes for labor analysis but use arrival-to-departure timestamps for site elapsed time.

Travel attribution

For the primary analysis, assign a base-to-first-job leg to the first job, an inter-job leg to the destination job, and the final return-to-base leg to the last job of the day. Count the person-minutes of the participating crew on each leg once; record the policy and any deviations.

This rule deliberately makes first and last jobs different from intermediate jobs. Report a sensitivity analysis excluding base legs, and do not imply the job-level travel distribution is independent of the attribution convention.

Work outside a particular job

Capture general marketing, bookkeeping, training, equipment maintenance, and unassigned idle time separately if feasible. Exclude them from job-attributed metrics and state that exclusion prominently; the primary metric is not the total burden of running the business.

Proposed metrics

Freeze formulas, inclusion rules, and missing-data handling before reviewing the main results. Keep canceled and completed jobs in distinct analyses with clear denominators.

• Total attributed person-minutes: Travel + setup + inspection/approval + service + closeout + job-specific administration + waiting, with no double counting.
• Non-service share: For eligible completed jobs with usable phase data, divide total attributed person-minutes outside hands-on service by total attributed person-minutes. Call this “non-service time,” not “waste”: travel, inspection, and setup may be necessary work.
• Site estimate error: Actual site elapsed minutes minus the pre-job estimated site elapsed minutes. Report signed error and absolute error separately.
• Relative site estimate error: The signed error divided by the pre-job estimate, limited to records with a positive usable estimate. Disclose how many records lack one.
• Delay prevalence: Completed jobs with a recorded delay divided by completed jobs with known delay status. Also show unknown status separately.
• Scheduled-job outcomes: Counts and proportions for completed, canceled, rescheduled, no-show, and partially completed records, using the explicitly defined scheduled-job denominator.
• Time to collect data: A brief participant burden measure, because an unusable logging process is itself a pilot finding.

For a rescheduled job, retain the original scheduled occurrence and link the replacement occurrence without calling it a second independent customer job. Publish both occurrence-level scheduling outcomes and completed-job timing with their respective units.

Analysis and quality control

Data checks

• Sequence: Flag negative durations, departure before arrival, impossible crew totals, and duplicate job IDs.
• Consistency: Compare phase person-minutes with crew attendance and elapsed site time where records permit. Investigate discrepancies rather than silently forcing them to reconcile.
• Missingness: Report missing data by field and business. Never replace an unrecorded phase with zero.
• Outliers: Do not automatically remove long jobs; verify the record and classify its scope. Document exclusions and report a sensitivity view when their effect is meaningful.
• Instrumentation: Keep timer-based and recalled records distinguishable. Repeat key summaries using only higher-quality timing if enough observations remain.

Weighting and summaries

Report both a pooled job-level result and a business-level summary so readers can see whether one high-volume participant dominates. For example, show the median job-level non-service share, the pooled ratio of total non-service minutes to total attributed minutes, and the median of business-specific shares as distinct statistics.

Every table should include the number of jobs and contributing businesses, the observation period, and the relevant missing-data count. Use medians and interquartile ranges for durations where appropriate; treat subgroup comparisons as exploratory.

Do not calculate a conventional survey “margin of error” as if recruitment were random. Any uncertainty interval should state its method and should not be used to disguise selection bias or clustering within businesses.

Hypothetical calculation, not a finding

Suppose a fictional solo operator records 30 minutes of travel, 15 setup, 10 inspection and approval, 120 hands-on service, 15 closeout, and 10 later administration, with no waiting. Total attributed time is 200 person-minutes, of which 80 are outside hands-on service.

For that invented job, the non-service share is 40%. This demonstrates the formula only; it is not a Rinse observation, an industry average, or evidence that the 80 minutes were avoidable.

Proposed release package

The first report should present what was actually measured, with methods available alongside the findings. If enrollment or completeness falls short, publish a smaller pilot and its lessons or delay the findings rather than filling gaps with assumptions.

• Main report: An accessible explanation of the sample, observed time allocation, estimation error, variation, and limitations.
• Methods document: The frozen protocol version, changes, recruitment information, data dictionary, validation rules, exclusions, and analysis formulas.
• Aggregate data: A privacy-reviewed summary table containing units, denominators, and contributing-business counts. Do not automatically release job-level logs.
• Practical worksheet: A free blank log readers can use, with the same definitions used in the research.
• Corrections log: A dated record of material changes and a route for participants or readers to report errors.

Suggested title after data collection: “How Participating Mobile Detailers Spent Their Job Time: A Four-Week Pilot.” The subtitle must give the actual number of businesses, jobs, locations at an appropriate level, and collection dates.

Claims the study may and may not support

• Appropriate: “Among participating businesses during the collection window, we observed...” followed by the actual statistic, denominator, and limitations.
• Not appropriate: “The average detailer loses X hours,” unless a design truly supports that population claim.
• Not appropriate: “Rinse saves X hours,” because an observational timing pilot is not a software-impact experiment.
• Not appropriate: “Documenting approvals prevents disputes,” because this protocol does not measure a causal effect on disputes.
• Not appropriate: “More non-service time means worse operations,” because job mix, geography, crew structure, and required checks can explain differences.

Launch checklist

Do not recruit until consent, storage, retention, and the collection instrument are ready. Do not publish findings until validation and privacy review are complete.

• Owner: Name a research lead and someone responsible for reviewing calculations.
• Feasibility test: Test definitions and burden with the instrument-test businesses and document resulting changes.
• Protocol freeze: Version the main protocol before its collection window and record deviations afterward.
• Consent: Approve participant information, withdrawal rules, incentives, and any optional quotation permissions.
• Collection: Use the same definitions for every participant and provide a way to ask questions.
• Analysis: Check formulas independently, retain excluded-record reasons, and compare business-weighted with job-weighted results.
• Release review: Verify every headline against the underlying statistic and every table against its denominator.
• Separation from sales: Do not make a favorable product quote or trial signup a condition of participation.