Pareto analysis for workflows: a worked example

Pareto analysis helps you choose which part of a recurring work problem to investigate first. It requires comparable records, a consistent measure and understandable categories. A ranking shows where the recorded burden is concentrated; the cause and a suitable change still need separate investigation.

Define the problem and the unit

Choose one process, such as handling internal requests for documents. Make the question specific: which situations consume the most additional time? You are measuring delays, not the overall value of employees or projects.

The ASQ procedure starts by defining categories, a measure and a period before assembling the data. Keeping brief notes about each case also lets you investigate what a category label actually describes.

Do not add request counts to minutes or place the same minutes in two categories. Mark missing observations explicitly. An absent record does not establish that no delay occurred.

Rank the observations by size

Here is an invented example. While handling 50 comparable requests during a defined period, a team records 300 minutes of additional work. Each interval is counted once.

The two largest categories account for 210 of the 300 minutes, or 70%. Choosing an area to investigate does not require an exact eighty percent. The cumulative percentage expresses the combined contribution of the leading categories, as explained in Minitab’s chart interpretation guidance.

The table does not promise that all 120 minutes of file searching can be removed. Some searching may be necessary, and records may be incomplete. At this stage, the ranking justifies examining cases; it is not a complete automation plan.

Investigate the largest category

Open the records labelled as file searching. Were people distinguishing between versions, trying to locate a folder or waiting for a download? Those situations need different changes. Split a broad category by observable circumstances if its label conceals several distinct problems.

Examine consequences elsewhere in the list too. An occasional access problem might block urgent work despite accounting for relatively few minutes. A ranking by time does not replace decisions about urgency and necessary conditions.

If the investigation reveals a recurring error, use the guide to correcting a process after a mistake. If the difficulty occurs during a handover, examine whether the task was specified clearly enough.

When the selected category still has several possible explanations, a fishbone diagram can organise them and help assign checks. Drawing the diagram does not turn a proposition into an established cause.

Change one element you can check

Suppose the examined cases show that finding the current template version takes the most time. One possible trial is a clear link to that version at the point where a task arrives. Establish who maintains the link and how colleagues will learn about the change.

Record your expectation: the next comparable set of requests should require less time specifically spent choosing a version. Keep necessary document checks. Their value does not disappear because they are less visible than producing the final output.

For digital organisation, see the guide to matching tools to a work task. A new application is unnecessary if a clear storage location addresses the problem. The trial should follow the observed cause, rather than the appeal of a new tool.

Compare results under similar conditions

Use the same recording method and account for workload. A total of 300 minutes across 50 requests has a different average from 300 minutes across 60 requests. A change in the type of requests may also explain the difference.

Check whether work has shifted to someone else or corrections have become more common. If the link helps, retain the arrangement. If it does not, return to the records and reconsider the explanation.

For the limits of the method, read the Pareto principle overview. The useful outcome is a reasoned change that has been checked, not a chart that happens to display a familiar ratio.