Layered Process Audits
In one sentence
A layered process audit (LPA) is different management tiers re-observing critical process conditions on the shop floor and responding to findings within their own authority.
By the end of this lesson
- Calculate the total number of audits and the time workload from 12, 6 and 3 observations at three tiers.
- Separate the concepts of number of observations, coverage of process points and separate observation evidence.
- Keep the shared critical process questions and define additional checks and acceptance evidence suited to each tier's authority.
- Correct an arrangement that counts repeated signatures as new audit evidence.
Why it matters
A layered process audit is not set up to have the same checklist signed by more people. It makes visible the difference between how the work should be done and how it is actually done, and why that difference persists. While the tier closest to the shop floor observes how the standard is applied, the next tier up can examine whether the necessary material, competence and support are available. The tier above that tests the ownership of obstacles that cannot be resolved and the effect of the management decision. The tiers can re-observe the shared critical questions on the shop floor at different times; the upper tier can add questions about support and resources within its own authority. Instead of merely reading the same answer, new observation evidence is produced; this way the audit workload turns into information usable for management.
If the standard is not clear, the audit result stays dependent on personal interpretation. Expressions such as looks orderly, works carefully or good practice are not acceptance criteria by themselves. It must be described which check will be made for which product, in which order and based on which record. The criterion must be observable on the shop floor and understood by the worker. That a document has been signed does not automatically prove that the condition in the document has been met. The observer builds the link between the record and the actual practice. When an incomplete standard or an unclear instruction is found, instead of closing it as the worker's fault, the observer treats it as a finding of the process that governs the standard.
Number of audits, coverage and effectiveness are separate dimensions. A large number of visits can concentrate on the same point; another critical job may never be seen. A point having been visited also does not mean that all of its shifts and products have been verified. For this reason the plan explicitly covers the risky process points and the observation times. Results are interpreted as limited to the observed condition. When audits are selected purposively, the finding percentage is not the defect rate of all production. A repeat finding requires the effect of the earlier action or the design of the control to be re-examined. Merely increasing the number of new audits does not answer this question.
The audit must not postpone safe work or the immediate handling of a problem to the day of the meeting. When a critical situation is observed, the predefined protection and support arrangement is run. Other findings are tied to an action with an owner, a date and a verification condition. The hierarchical level of the person auditing does not replace the accuracy of the finding; evidence is sought with the same clarity. The upper tier, while repeating the work of the lower tier, must not neglect the resource obstacle it needs to resolve itself. The success of the audit is seen less in producing many conforming boxes than in detecting the disconnect between the standard and management support and tying it to lasting learning.
A process audit must also be separated from the acceptance check of the finished product. A product coming out conforming does not by itself prove that all the defined process conditions were monitored. Likewise, finding an instruction record conforming does not mean that the characteristics of every part produced at that moment were measured. A layered audit produces selected evidence that the necessary conditions are being sustained; it does not automatically take the place of product inspection. This distinction prevents more confidence being placed on the audit result than it can carry. When the plan is reviewed, it is explained which risk is managed by which control. When unnecessary repetition is reduced, two different assurances are not accidentally merged. If a control is to be removed, where the information it provided will come from and how the decision will be protected is shown in advance. This way a lighter audit plan is not confused with weaker assurance.
Formula
Total audits (observations) = First tier + Second tier + Third tier
Total workload (min) = Sum of (Number of observations × Time per observation) over the tiers
Point coverage (%) = Distinct process points observed ÷ Process points planned × 100
Finding rate (%) = Observations with nonconformity ÷ Observations completed × 100
Unlock the rest of this lesson with an Academy account
Create a free Academy account with your work email address; it can be used once it is approved.
- Step-by-step instructions
- Worked example with real numbers
- Where the method misleads
- Common mistakes