Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Metric

Rework Rate

The share of items sent back from a review or launch gate to an earlier stage; each return carries a reason, so the pattern points to the step that failed.
AnalyticsmetricUpdated

Formula

Items sent back from a gate to an earlier stage ÷ Items that reached the gate

Unit: % · lower is better

Definition

Rework rate counts returns at a gate, one of the checkpoints in a workflow where work is approved or sent back. The hub on marketing process design defines rework as the share of items sent back from a gate to an earlier stage, and pairs it with cycle time as the measures that show whether a process is improving.

Items sent back from a gate to an earlier stage ÷ Items that reached the gate(%)

  • Items that reached the gate are the pieces of work submitted to one gate, such as brand review, legal review or the launch check, in the period being read.
  • Items sent back are those the gate returned to an earlier stage, such as the brief or the build, rather than approving them.
  • The gate is part of the definition. A rate pooled across every gate hides which one is returning work and why.

Write down how items are counted and keep it: the guide to measuring AI's impact on team output counts the share of outputs sent back at least once.

How to interpret it: by gate and by reason

Two readings turn the figure into a diagnosis.

By gate. The process hub reads each gate on its own terms: returns at approval most often trace to a weak brief rather than a hard reviewer, and returns at the launch check to a build procedure with a step left out.

By reason. Every return should carry a reason chosen from a short list. The campaign workflow guide suggests reasons such as brand standard, off-brief, legal, technical or an inaccurate claim, and observes that a gate returning work without reasons teaches the team nothing. Counted by reason, the returns show a pattern, and each cause in it has an owner upstream. The same guide traces most rework at review to a brief nobody signed, or one party altered after signing.

A low rate is not automatically good news. Read it beside errors found after publication, which the AI measurement guide uses to show whether review is keeping up with volume.

What moves it

  • Brief quality and sign-off. A brief signed by the requester and by the campaign owner, with changes going back through the same signature, removes the most common reason for returns at review.
  • What reviewers see. Reviewers given the signed brief alongside the work review against intent rather than taste, so fewer returns rest on preference.
  • Written procedures with checks. A build task with a written procedure, ending in a check of its own, sends fewer items back from the launch check.
  • Inputs at intake. Work started without its facts, assets or decisions is more likely to be built on assumptions that a reviewer then rejects.
  • AI-assisted drafting. New tools change the rate in either direction. The AI measurement guide counts rework beside defects found in review and time per accepted output, so that faster drafting paid for in returns shows up.

What good looks like

There is no reliable benchmark for rework rate, and the site does not cite one. How much a gate returns depends on what it checks, how strict its standard is, how risky the work is and how the team counts, so two teams' rates are seldom counts of the same thing. Compare it with the team's own history, gate by gate and reason by reason, counted the same way each period. And judge it by what returns cost: each one adds rework time and another wait in a review queue, and the reason codes show which upstream fix would remove the most of them. Zero is not the aim. A gate exists to stop work that is not ready, and some returns are the gate doing its job.

Illustrative example — not a real result

A software firm finds why its landing pages bounce

Tulkaix Payroll, a payroll software firm, began coding every return from brand review. In one quarter 60 landing pages reached the gate and 27 were sent back, a rework rate of 45%. The reviewer was assumed to be the problem, but 18 of the 27 returns carried the same reason: the trial length or the discount had changed after the brief was signed, so the page no longer matched the offer sales had agreed. The firm added an offer decision, owned by the head of sales with finance, as the step before any brief could be written, and froze the offer once the brief was signed. The next quarter 58 pages reached brand review and 11 were sent back, 19%, with 3 citing the offer.

Cycle Time is rework's partner on the process hub: each return repeats stages and rejoins queues, so rework shows up as elapsed time, and faster delivery bought with more returns is a gate being hurried rather than a process improving. An Approval Matrix decides which gates each type of work passes and who sits at each, which is what makes a rate per gate possible to read. The Creative Brief is the input most returns at review trace back to. Human Review is the gate at which AI-assisted work is accepted or sent back, and the rework rate at that gate is one of the measures of whether a tool is helping.

The tutorial on setting service levels for marketing requests counts rework after review among the places where time goes when a request type misses its level, reported for the team rather than by person. In a KPI tree, defect and rework rates sit beside the drivers as guardrails, so that a gain in speed or volume is not counted as success while returns climb.

Common mistakes

  • Blaming the reviewer. A high rate at a gate usually starts upstream, in the brief, the inputs or the build procedure.
  • Returning work without a reason. The rate then shows that work bounces but never why.
  • Pooling every gate into one figure. Returns at brand review and at the launch check have different causes.
  • Changing the counting rule between periods. The rule is written down once and kept.
  • Celebrating a falling rate alone. Read it beside the errors found after publication.
  • Making it a personal target. Items get closed early, rounds of feedback are skipped and difficult work is avoided, and the number stops describing the process.

More terms to look up

Where this term is used

Related