Question and scope
This protocol measures how often a cross-team request changes scope after work has started, and how often that change carries a recorded approval and a revised due date. Scope churn is easy to feel and hard to prove. A requester adds one more market, a manager raises the quality threshold, or a receiving team discovers that the original brief left out a required review. Each change may be reasonable, but together they move the completion date without changing the original promise.
Scope is one request type handled by one coordination team during a defined window, such as one quarter or the last thirty completed requests. Include changes that alter the deliverable, the acceptance criteria, the volume, or the level of review. Exclude changes that only correct a typo or clarify wording without changing the work. State the request type and the window in the report so another reviewer can reproduce the boundary.
What counts as a scope change
A scope change is a modification that changes the effort required to meet the agreed outcome. Adding a required approval, increasing the number of records to process, or raising a quality threshold all qualify. Renaming a field or fixing an obvious typo does not qualify because the work remains the same. The definition should be written before coding so a minor edit and a material addition are not counted as the same event.
The timing matters as much as the content. A change proposed before work begins is part of intake and should be handled there. A change proposed after the first production action is the event this protocol measures. Record both the date the change was requested and the date it was accepted, because a slow approval can add delay even when the scope itself is small.
Unit of analysis and exclusions
The unit is one request with one approved baseline and one receiving team. A bundled request should be split when each part can be approved independently. If the parts cannot be separated, keep one unit and record that the scope was shared. Use the baseline as it was accepted at triage, not a cleaned version written after the work was finished.
Exclude duplicates, test records, and requests that were cancelled before any work started. Exclude a change that was never accepted, but report it separately as a proposed change. Do not exclude a late change simply because it was approved by a senior person. Every exclusion needs a reason and a reference to the record so the denominator can be defended.
Start and stop clocks
Start the clock at the first production action on the accepted request, not at the first message. The first action might be a draft, a data pull, a system change, or a scheduled review. If the team cannot identify that moment, use the triage timestamp and mark the start as estimated. An estimated start is weaker evidence and should be reported as such.
Stop the clock when the work is accepted by the receiving team or when the request is closed without delivery. Count each approved scope change as a separate event, and record the days between the change and the revised completion date. Do not stop the clock at the first change, because the cumulative effect of several changes is often the real source of delay. Use the same stop rule for every unit in the sample.
Evidence and data dictionary
Each unit needs an identifier, the original request, the baseline date, the first production date, the requesting team, and the receiving team. Add each change with its request date, approval date, approver, reason, and revised estimate. Link to the message, ticket, or meeting note where the change was agreed. The evidence should let a reviewer rebuild the sequence without interviewing the team.
Define the terms before coding begins. Distinguish a clarification, a proposed change, an approved change, and a reversal. Define whether a change approved in a meeting counts when the minutes are published later. Include one example of a material change and one example of a clarification in the dictionary. A short shared vocabulary prevents two reviewers from counting the same event differently.
Sampling and test design
Use a census when the window contains fewer than roughly one hundred completed requests. Above that volume, draw a random sample from the request export and document the method. Stratify by requesting team or request type so one high-volume group does not dominate the result. Freeze the sample before reading the change history to reduce selection bias and make the audit repeatable.
Have a second reviewer recode at least ten requests and compare the results. Resolve disagreements by returning to the source record rather than by voting. If the disagreement rate is high, revise the definition of a material change and recode the full sample. A measure that depends on one reviewer's intuition will not survive a change of reviewer or a new quarter.
Decision rules before coding
Decide how to treat a change that was approved after the work was already completed. One option is to count it as a change and record the retroactive approval date. Another is to exclude it from the timing analysis but keep it in the approval share. The choice changes the result, so write it down before the first row is coded.
Decide how to handle a change that replaces one deliverable with another of similar effort. Record it as a change even when the total effort looks unchanged, because the receiving team still had to re-plan. Also decide whether a change requested by the receiving team counts the same as one requested by the requester. Both can cause churn, but the pattern may point to different fixes.
Analysis and reporting
Report the share of units with at least one material change after work started, the median number of changes per affected unit, and the median added days. Add the share of changes with a recorded approver and the share with a revised acceptance criterion. Show the count behind every percentage, especially when the sample is small. A share based on four changes should be presented as a count, not a rate.
Break the result down by requesting team, receiving team, and request type. Suppress groups with fewer than five units so a single unusual request does not look like a pattern. Pair the scope result with the original due date and the final due date. A team with few changes but long delays may have a different problem, such as slow approval or unavailable capacity.
Reporting template
Use a one-page template with blocks for scope, units reviewed, share with a material change, median added days, share with recorded approval, oldest open change, and one action for the next cycle. Put the raw counts beside every percentage and name the source of each date. Add a short note for every exclusion and every change to the definition. The template should be readable in a five-minute review.
Store the template with the request export, the coding sheet, and the sample list. A later reviewer should be able to rebuild the headline share without asking for context. If the request type or the system of record changes, start a new series instead of merging old and new data. A clean break is more honest than a trend line built from two different processes.
Limits and sources of error
The cited public sources describe project scope, internal control, and iterative delivery in general terms; they do not provide a benchmark for a specific queue. A recorded change can be a reasonable response to new information, and a low change count can reflect a brief that was too vague to reveal the real work. The measure also depends on whether the tracker preserves the original baseline. A system that overwrites the first estimate will undercount churn.
Backfilled approvals and reconstructed timelines are common sources of error. Check a sample of changes against the original messages and mark any row that was entered after the fact. Exclude a unit when the baseline or the first production date cannot be verified, but report the exclusion count. A precise number built on guesses is worse than a range with a clear caveat.
Review cadence
Review the result at a fixed point and end with one change, one owner, and one check date. Repeat the test after two full cycles before setting a target, because a single quarter may reflect an unusual project mix. Keep the definition and the sample method with the result so the next reviewer can compare like periods. Mark any process change on the same timeline as the data.
Retire the measure if it stops informing a decision. If the team keeps it, publish the share of approved changes and the median added days together, because approval without schedule impact can hide a slow handoff. The purpose is a clearer promise to the receiving team, not a perfect record. Stop when the team can show that outcome consistently across several cycles.
Sources
- U.S. Government Accountability Office, Standards for Internal Control in the Federal Government, reviewed November 9, 2026.
- U.S. Government Accountability Office, Agile Assessment Guide: Best Practices for Agile Adoption and Implementation, reviewed November 9, 2026.
- Project Management Institute, PMBOK Guide, reviewed November 9, 2026.
- U.S. Bureau of Labor Statistics, Occupational Outlook Handbook: Project Management Specialists, reviewed November 9, 2026.
