Identify the variation that matters.
Start with a specific difference. One team sends requests directly to a reviewer; another routes them through a coordinator. One shift records exceptions in the system; another keeps a separate list. Those differences deserve examination when they affect quality, delays, rework, or the next person’s ability to act.
Some variation may be appropriate. Different request types can require different handling. Define the required outcome and the conditions under which the procedure should apply before treating every difference as a defect.
Compare the same kind of work from start to finish.
Select a few recent examples from each team or shift. Keep routine work separate from urgent or unusual cases. Map the steps, decisions, handoffs, and return loops with the people who do the work. Use the current procedure as a reference and record where practice differs.
ASQ’s flowchart guidance recommends checking the map with the people involved and walking through the process to confirm its accuracy. A polished diagram is less useful than a map that captures a missing handoff or an unofficial approval.
Ask what information was available at each decision, which instruction was used, and what happened next. If accounts conflict, look for records or observe another example rather than choosing the most confident account.
Find the point where the instructions stop helping.
The useful question is what makes the next action different. Check document versions, vague decision rules, system permissions, responsibility for exceptions, and the information available at a handoff. If the difference appears only during busy periods, examine workload and priorities as well.
Illustrative scenario, not a client result. A procedure says to send a request to a coordinator “when needed.” One team interprets this as every request; another uses the coordinator only when information is missing.
Repeating the instruction does not resolve its ambiguity. The process owner needs to define which conditions require coordinator review, what a ready request contains, and how unusual cases are handled. That proposed rule still needs a test in the actual workflow.
Use the comparison to understand the process. Variation alone does not establish that an individual is unwilling or unable to do the work.
Try the revised rule where the difference occurs.
Agree on the rule with the people who send and receive the work. Make the current instruction easy to find, retire conflicting versions, and identify the person who can resolve an exception. Provide practice if applying the rule requires a skill people have not yet learned.
Start with a defined part of the workflow and a review period. Record whether the rule is used, why exceptions occur, and whether work takes an unexpected route. A sample diagnostic shows one way to document the recommendation and its limits.
Check consistent use and the outcome it serves.
Measure use of the agreed step among eligible cases. Pair that with the outcome that prompted the work, such as repeat handoffs or request turnaround time. Report counts with percentages and keep the type and complexity of work visible.
More consistent use does not by itself show that the revised procedure is useful. Look for decision errors, extra preparation, and a queue building elsewhere. If outcomes improve, check whether other changes could account for the result before expanding the approach.
If people know the procedure but cannot use it reliably, read the training-transfer guide. If the main symptom is delay, start with workflow bottleneck analysis. Our process improvement consulting brings those questions together around one defined problem.