Process Improvement Techniques for Auditors

Every audit at one bakery ends with the same finding: incomplete sanitation records on the night shift. Each time, staff are retrained. Records improve for two weeks, then slip again. Retraining treated the symptom. Process improvement asks a harder question: what in the process makes the error likely, and how should the process change so the error becomes difficult?

As an auditor, you do not redesign the auditee’s processes. You do evaluate whether the organisation can improve its own processes, and you can sharpen your own audit process using the same techniques. Both uses matter for lead auditors.

Root Cause Analysis Done Properly

Clause 10.1 requires the organisation, after a nonconformity, to evaluate the need for action to eliminate root causes so the problem does not recur elsewhere. Weak root cause work stops at human error. Human error is a starting point: why was the error possible, why was it not detected, and why did supervision and system design allow it?

Common structured methods include asking why repeatedly to reach a systemic cause, grouping possible causes by category such as people, process, equipment, materials, environment, and management, and testing each candidate cause against evidence. Whichever method the organisation uses, audit the logic. A root cause claim must explain all the evidence, including why similar shifts or lines did not fail, if that is the case. Corrective action must match the cause: a form redesigned to fit the sanitation sequence, a checklist moved to the point of use, an interlock added, staffing adjusted. Action effectiveness must be reviewed after implementation, over a period long enough to prove stability.

Process Mapping and Waste in Audit Trails

Improvement begins with seeing the process as it is. Flow diagrams required by Clause 8.5 already map production. Improvement mapping adds decision points, queues, rework loops, and record creation steps. Many recurring nonconformities live in those additions: product waiting unlabelled, rework re-entering without identification, records completed retrospectively because the form lives in an office away from the task.

During audits, map the evidence trail the same way. If verification records exist but analysis never happens, the break is a process step missing an owner and a deadline, not a diligence problem. Recommend nothing as an auditor, but frame findings so the missing step is visible. Precise findings enable improvement; vague findings invite cosmetic fixes.

Using Data: Pareto and Trend Thinking

Improvement needs prioritisation. Grouping nonconformities, complaints, and deviations by type, line, shift, and supplier over time usually shows concentration: a small number of causes producing most events. Organisations should be able to show such analysis under Clause 9.1 and Clause 8.8. As an auditor, ask to see it. If data exists only as unreviewed lists, evaluation requirements are not effective even if records are complete.

Trend thinking also protects against overreaction. A single deviation inside a stable process calls for correction and evaluation. A slow upward drift in environmental positives calls for systemic action before a limit is breached. Distinguishing signals from noise is a competence to test in the food safety team and to practise in your own audit conclusions.

PDCA and Small Controlled Changes

Plan-Do-Check-Act disciplines improvement. Plan a change with a clear problem statement, baseline data, and success criteria. Do it on a limited scale where food safety allows. Check results against criteria, including unintended effects on hazard control. Act by standardising the change, updating documented information, hazard analysis, and training, or by abandoning it.

Clause 6.3 requires changes to the FSMS to be planned and carried out considering purpose and consequences, system integrity, resource availability, and allocation of responsibilities. Uncontrolled improvement is a food safety risk: a new cleaning chemical, a faster line speed, or a new supplier introduced as an improvement without validation and updating violates Clause 8. Audit improvements as rigorously as original controls. Ask for validation evidence, updated flow diagrams, and revised monitoring where a process changed.

Improving Your Own Audit Process

Apply the same discipline to auditing itself. After each audit, review what sampling missed, which questions produced strong evidence, and where time was wasted. Lead auditors should run brief team retrospectives and feed lessons into the audit programme under ISO 19011 thinking: auditor competence development, better risk inputs for planning, and checklist revision. An audit programme that never changes its sampling despite repeated surprises is exhibiting the same failure it reports in auditees.

Continual improvement under Clause 10.2 judges the organisation’s suitability, adequacy, and effectiveness over time. Your evidence sources are trends in internal audits, verification analyses, management review decisions, and whether yesterday’s corrective actions stayed effective. Improvement is a pattern across cycles, not a slogan in a policy.

Standardisation locks in gains. When an improvement proves effective, procedures, forms, training material, flow diagrams, and verification plans must all move together to the new method. Partial standardisation, where the procedure changes but training and verification stay on the old method, creates a fresh nonconformity while solving an old one. In follow-up audits, test standardisation directly: ask staff trained after the change to demonstrate the task, check that forms in use carry the new revision, and confirm verification now checks the new control. Improvement that survives staff turnover and shift change is real improvement.

Key points

• Root cause analysis must go beyond human error to the process conditions that made error likely and undetected.
• Corrective action must match the proven cause, and its effectiveness must be reviewed over a sufficient period.
• Mapping queues, rework, and record steps exposes where recurring nonconformities originate.
• Use grouped and trended data under Clauses 9.1 and 8.8 to prioritise action and separate signals from noise.
• All process changes affecting the FSMS must be planned under Clause 6.3 and validated before standardisation.
• Apply PDCA to your own audit programme and sampling as well.