On this page
Research question and scope
How often does website work that the client already accepted return for correction after the handoff, and which handoff conditions make that return more likely? The scope is one outsourced design operation during a defined observation period, not an industry benchmark. The measure matters because a handoff that looks complete can still generate weeks of quiet repair work. This protocol does not judge whether the vendor or the client performed well. It describes how often accepted work comes back, what kind of defect returns, and where the handoff record failed to prevent it.
Population and unit of analysis
Start with a dated list of deliverables accepted during the observation window, such as a page template, a component, a content migration, or a release package. Record the project, the deliverable, the acceptance date, the named approver, and the production route or repository reference. Keep the original list stable and document any deliverable added later. The unit is one accepted deliverable that returns for correction within a defined follow-up period. A page set accepted as one package counts as one unit, while a single component used across several pages still counts once unless it was accepted separately.
Definitions and measures
A return is a request to change accepted work because the delivered result did not meet an agreed requirement, not because the client changed direction. Distinguish a defect return from a scope change, a content update, and a routine maintenance task. Record the return date, the reason category, the affected page or component, and whether the issue was visible in the acceptance evidence. For each deliverable, report whether it returned, how many times, and how long the correction took. Also record the handoff channel, because a written acceptance ticket provides different evidence from a verbal approval in a meeting.
A return also needs a severity label so a broken checkout is not averaged with a spacing correction. Use three levels: blocking, where the page cannot perform its main task; material, where the result differs from an agreed requirement but still works; and minor, where the difference is cosmetic or editorial. Record the time from acceptance to return and from return to correction, since a fast fix can hide a slow discovery. If a deliverable returns more than once, keep each return as a separate event rather than counting only the first. This keeps the measure sensitive to repeat failure without exaggerating the number of affected deliverables.
Evidence and data dictionary
Each return needs the original acceptance reference, the defect description, the reviewer who raised it, and the evidence supplied with the complaint. Record the acceptance version separately from the production version, since a later deployment can create a defect that the handoff did not cause. If the client reported the issue through chat, preserve the message timestamp and attach it to the deliverable reference. Keep screenshots of the accepted state and the returned state where the repository allows it. Use project references rather than client names in the analysis file, and keep unpublished designs inside the approved system.
Test design
Use a census when the number of accepted deliverables is small enough to review directly. For a larger portfolio, select a documented sample that includes large page sets, small component handoffs, and deliverables accepted near a launch deadline. Freeze the sample before analysis so the return rate cannot be adjusted after the first results appear. Have a second reviewer classify a subsample of returns and record disagreements before resolving them. Separate returns caused by the handoff from returns caused by a later platform change, and note any deliverable whose follow-up window was too short to observe.
Pair each sampled return with a control deliverable accepted in the same period that did not return. Compare the handoff evidence available for both, such as tested routes, breakpoint screenshots, and written acceptance notes. This comparison is more useful than a raw rate because it shows which evidence was present when the handoff held. Match controls on deliverable type and project size where possible, and record when no suitable control exists. The pairing does not prove causation, but it reduces the chance that a busy month explains the whole result.
Measures and limitations
The result can support a change to the acceptance form, the evidence pack, or the post-launch support window. It cannot prove that the vendor caused every return, because a content editor or a hosting change may have altered the accepted state. A low return rate does not always mean the handoff was strong, since clients may absorb small defects without reporting them. Report the follow-up period and the project mix beside the rate, and avoid ranking individual reviewers. The useful question is whether the handoff record gives the next person enough evidence to distinguish a defect from a new request.
Report the number of sampled deliverables separately from the number that returned, and show the follow-up window for each group. If some projects stopped reporting after a few weeks, state that the rate is incomplete rather than treating silence as success. Include a short list of returns that could not be classified and explain why the evidence was insufficient. A transparent unclassified group is more honest than forcing every event into a defect or a scope change. The reader should be able to see how much of the result rests on clear evidence.
Review and next action
Report returns by deliverable type, reason category, project, and month. Show the handoff evidence that was missing when a defect returned, and separate defects that were visible in the accepted screenshots from defects that were not covered by any acceptance check. Name one change to the handoff record, one owner, and one recheck date. For example, a team might add a required mobile screenshot for every accepted template or a short list of tested error and recovery routes. Keep the acceptance records and the return log with the finding so a later reviewer can repeat the comparison.
Reporting the result without blame
Present the finding as a property of the handoff system rather than a verdict on a person. Use counts and categories, and quote the missing evidence instead of describing a reviewer as careless. If a return came from a requirement that was never written down, label it as a specification gap rather than a delivery failure. If the same defect returns across several projects, show the pattern and the shared template that allowed it. A report that invites defensiveness will produce fewer honest classifications in the next cycle.
Review cadence
Run the review after each major release and summarize it at the monthly design operations review. Repeat it after a change to the acceptance template, the reviewer roster, or the support window. Archive the acceptance export, the return log, and the classification notes so the measurement can be repeated. A repeated return pattern should change the handoff checklist before the next project starts. Keep the observation window beside the result so the next review compares like with like.
Sources
The sources below provide general context for accessibility, security, and business operations. They do not supply a local return rate or a performance benchmark for this team, and they were checked on November 6, 2026. Use them for general context only, and keep the local observation separate from any external claim. The source list is not a substitute for the acceptance record or the return log.
- W3C Web Content Accessibility Guidelines , https://www.w3.org/WAI/standards-guidelines/wcag/
- NIST Cybersecurity Framework , https://www.nist.gov/cyberframework
- NIST Privacy Framework , https://www.nist.gov/privacy-framework
- Federal Trade Commission , https://www.ftc.gov/
- U.S. Small Business Administration , https://www.sba.gov/
- U.S. Government Accountability Office , https://www.gao.gov/
Turn the guidance into a reviewable handoff
Record the owner, input, acceptance condition, evidence, and escalation point before production begins. That small contract makes the work easier to inspect and approve.
Plan a website production handoff