In a layered process audit, which layer should find the most?

In February 2012, in an Elsmar Cove forum thread about taking layered process audit results to management review, a member suggested “reporting both number of non-conformances (audit findings) and well as repeat issues. Break it out by layer.” Then came the rule: “It’s no good if the 1st layer group leader finds none, 2nd layer supervisor finds 10, and 3rd layer quality finds 50. Should be the other way around.”

Earlier in the same thread, another member had laid out the structure: “supervisors audit certain items daily, managers audit items weekly up to the plant manager who audits items monthly.” The first layer is next to the process every shift. If it sees the least, the place to look is what it is being asked to check.

What a clean sheet means

The argument over what a clean sheet means goes back further. In May 2009 a member posted: “I had three audits done from different levels for the same area today and three different answers.” The first reply didn’t treat that as a failure, “so long as they aren’t contradictory,” and asked: “Did the three auditors get together and discuss their results?”

Then the thread split. One member blamed the design, “the whole concept of keep asking the same thing over at such a high frequency (per shift, etc) and expecting people to ‘go along with it’.” Another posted seven warning signs, “Program is ran by quality only” among them, and minutes later answered the skeptic: “I completely disagree.” A third had “seen both effective and marginal LPAs” and put the difference in “expectations from the top and corporate culture.” Two days into the thread, the same member named what people watch: “How seriously the upper levels take their portion of the audit, and how seriously the action items are reviewed and fixed.”

The case for zero

There is a fair case for zero. “Now that they are doing the LPA I am finding less nonconformities during internal audits of their processes,” a member wrote in 2008. If the line catches its problems first, the auditor has nothing to write. A rule passed along in 2005 points the other way: “If your layered audits are always compliant 100% the audit is doing you no good anymore. It’s time to re-evaluate your process and ‘up the bar’ on the requirements.” On a monthly report, those two plants can look the same.

One large count comes from a vendor’s customers. EASE, which sells LPA software, reported in May 2026 on more than 2.3 million audits run by its users: “While nearly half of upper management audits result in at least one finding, 84% of frontline audits result in zero findings.” It covers EASE users only.

If your layered audits are always compliant 100% the audit is doing you no good anymore.

What goes upstairs

Then there is what goes upstairs. The 2008 member reported LPA activity at the monthly quality meeting “by % completed by each level.” In 2011, the member who opened the management review thread wanted better metrics than “audits completed to schedule,” which “does not give any indication of the effectiveness of the LPA in identifying manufacturing process weaknesses.” Counting findings has its own trap. “Measuring ‘findings’ as a KPI will drive that number, probably not the best idea,” a member wrote in 2016, and offered another: “How about ‘findings corrected’?”

The checklist ages too. The 2012 member suggested keeping the core items fixed and adding checks “based on Customer return, failed calibration or higher scrap seen on a particular machine,” then letting them drop off once a verification period passes. Michael Ballé, answering a reader in the Lean Enterprise Institute’s Gemba Coach column in 2018, was writing about maturity audits, and the warning carries over: “standardizing and routinizing such audits is plain silly as people will habituate to it.” On process audits: “Absolutely, but never twice the same.”

Questions for your own records

If you run the audit program across plants, the forum’s argument turns into questions for your own records:

• What does the monthly report count: audits completed, findings, or findings corrected?
• When two layers audit the same area in the same week, does anyone compare what they found?
• Do the line’s scrap and customer returns agree with its clean sheets?
• When did a question on the checklist last change, and what triggered it?
• Who runs the program: quality alone, or the managers whose layers are on the schedule?

Start where the 2012 post did. Break last quarter’s findings out by layer. Which one found the most?

Sources

• Elsmar Cove forum, “Consolidating results of LPA (Layered Process Audit) for Management Review”: the 2012 “other way around” post, #10; the 2011 description of the layers, #3; the 2011 “audits completed to schedule” post, #4.
• Elsmar Cove forum, “Having faith in your Layered Process Auditors - Pencil whipping audits,” May 2009: opening post, first reply, #2, the design objection, #3, the warning signs, #4, “I completely disagree,” #5, “expectations from the top,” #6, what people watch, #8.
• Elsmar Cove forum, “What is a Layered Process Audit (LPA)?”: 2005, the “always compliant 100%” rule; 2008, fewer nonconformities and “% completed by each level”.
• Elsmar Cove forum, “KPI (Key Performance Indicators) for the Audit Process,” 2016, post #3.
• EASE, “3 Surprising Results from Our Layered Process Audit Benchmark Report,” May 6, 2026. Vendor data, from EASE users only.
• Lean Enterprise Institute, Gemba Coach, “Should we do lean maturity audits?,” October 29, 2018.