Outsourced ProgrammersPowered by StealthAgents

What share of offshore defects are reopened after a fix?

A protocol for measuring reopened defects, the reason, the time to reopen, and whether the first fix addressed the cause.

What share of offshore defects are reopened after a fix?
WORK_LANE / RESEARCH BRIEF / QUALITYREVIEW QUEUEACCEPTANCEPASSowner review requiredEVIDENCE TRAILTICKET → BRANCH → REVIEW → CLIENT RELEASE
Table of contents
  1. Research context
  2. Practical application
  3. Operating controls
  4. Sources
  5. Frequently asked questions

Research context

This protocol asks what share of offshore defect fixes are reopened and why. The unit is one defect that was marked fixed during the window, identified by the defect identifier and the fix build. The study measures the share reopened within thirty days, the reason, and the time between the fix and the reopen. It does not judge the engineer's skill or assume that every reopen is a failed fix.

Record whether the reopen was reported by the same customer or by a different one. A different customer may indicate that the fix addressed only the reported symptom. A same-customer reopen may indicate a communication gap. The source changes the interpretation and the next action.

Set the scope to one product, one defect process, and one definition of fixed. Record the environments, the acceptance evidence, and the reopen window at the start. If the process changes during the window, keep the earlier definition for the full sample or analyze the change as a separate period. A new verification step can look like more rework when it is only a different gate.

Measure the time from the first fix to the reopen in business days, not only calendar days. A defect reopened after a weekend may have been waiting for a customer test rather than for the engineer. Report both when the gap is large. The reader should see the customer's experience, not only the timestamp difference.

Define the population as every defect marked fixed in the window, including fixes that were later reopened or reverted. The unit is the defect, not the commit, because one fix can touch several commits. A duplicate defect that is linked to the original is a separate unit with a link. Test the definition on two recent fixes before collection begins.

Check whether the fix included a regression test that reproduced the original defect. A fix without a test is more likely to be reopened because the team cannot prove the behavior stayed fixed. Record the test identifier and whether it ran in the release pipeline. That field turns the reopen rate into a test-coverage finding.

Record the defect severity, the area, the fix size, the verifier, the evidence attached, and the time to close. Add the reopen date, the reason, and the second fix. Keep the original report and the fix record together. A second reviewer should be able to classify the reopen from the evidence without reading a chat thread.

Ask the verifier what evidence they used before closing the defect. A verifier who relied on a developer's note is less independent than one who ran the reproduction. Record the verification method and the person. The reopen review should distinguish a verification gap from a coding gap.

Practical application

Report the number of fixed defects, the share reopened within thirty days, the median time to reopen, and the share reopened more than once. Show the counts by severity, area, and verifier. Present the share of reopens caused by an incomplete fix, a regression, a misunderstood requirement, or a test gap. Use the same denominator for every rate.

Separate a reopen caused by a new defect from one caused by the same defect. A new defect in the same area is a regression risk, while the same defect is a failed fix. Report the two categories separately. Pooling them hides whether the team is fixing the cause or chasing new symptoms.

Inspect the reopened defects and describe the pattern without naming the engineer. A cluster in one area may point to a fragile test or a missing environment, while a cluster in one severity may point to a rushed fix. Compare the result with the fix size and the evidence attached. Report defects reopened after a customer report separately.

A useful additional view is the share of reopened defects that were closed again without a code change. A documentation or configuration fix may be legitimate, but a pattern of no-code reopens suggests the first fix was incomplete. Record the second resolution and its evidence. The pattern is often more useful than a single case.

The record cannot show whether the first fix was reasonable given the information available. It can show only that the defect was reopened and what reason was recorded. A reopen is not proof of poor work, because new evidence can change the expected behavior. The protocol also cannot capture a defect that was never reported again but remained broken.

Do not use the reopen rate as an individual performance metric. A team that discourages reopening will leave defects marked fixed and lose trust with the customer. The measure is useful at the process level, where it can reveal a weak verification step or a missing test. Keep individual data out of the summary.

Give the result to the QA lead and the client owner and ask them to choose one observable change. The change might be a required reproduction before closing, a second verifier for high-severity fixes, a regression test for every reopen, or a shorter reopen window. Record the owner, the effective date, and the measure that will show whether the change worked. Repeat the protocol with the same definitions.

Compare the reopen rate with the defect severity and the fix size. A small fix in a critical path may deserve more scrutiny than a large cosmetic change. Report the groups separately and describe the cases. A single overall rate can make a safe area look risky or a risky area look safe.

Operating controls

Review the defects that were reopened after the customer reported them again. That sequence means the customer experienced the same problem twice, which is worse than a late first fix. Record the customer impact and the apology. The owner should see the cost of a weak verification step.

End the review with one change and a date to re-measure. The change might be a required reproduction, a second verifier for critical defects, a regression test, or a shorter review window. Record the owner and the expected effect. Repeat the protocol with the same definitions so the comparison is honest.

A useful additional view is the share of reopens that happened after the fix had already passed a code review. A review is not a substitute for running the reproduction, and a reopen after review points to a verification gap rather than a coding gap. Record the review and the verification separately. Check whether the fix changed more than the reported behavior. A broad fix can introduce a regression in a nearby path, which then appears as a reopen. Record the lines changed and the tests that cover the area. A smaller fix is easier to verify and easier to reverse.

Ask the customer who reported the reopen whether the second fix was faster than the first. A slow second fix compounds the frustration and may damage the relationship more than the original defect. Record the time to the second fix. The owner should see the full customer cost, not only the internal reopen rate.

Do not pool a reopen caused by a failed deployment with one caused by a coding error. A deployment that did not include the fix is a release problem, not a fix problem. Record the build that reached production and the build that contained the fix. The pattern can reveal a pipeline gap.

A useful additional view is the share of reopened defects that were caught by an automated test before the customer reported them. A test that catches the reopen is a success even though the defect returned. Record the test and the time it ran. That view separates a detection win from a quality failure.

Ask the customer support team whether the reopened defect produced a second complaint or a refund request. A reopen that reaches support again has a customer cost that the engineering record does not show. Record the customer contact and the remedy. That view helps the owner prioritize the verification work by impact rather than by ticket count alone.

Sources

  1. ISTQB glossary
  2. Google Testing Blog
  3. NIST SP 800-53 security and privacy controls
  4. CISA Secure by Design

Frequently asked questions

What should happen first?

Begin with a bounded ticket, approved access, explicit acceptance evidence, and a named reviewer.

Who approves production changes?

The client's technical owner keeps final merge and release authority.

Related research

How often do offshore releases need a rollback rehearsal?

How often do offshore support escalations miss their response target?

Build a programming lane your technical owner can review.

Bring the stack, backlog, acceptance checks, and release boundaries. We will help scope the role around the work.

Discuss your programming needs