Two weeks before a customer audit, a familiar sequence begins. An area gets repainted. Records get "tidied". Staff are briefed on what to say if asked. A manager decides who will and will not be available on the day.
That is not preparation. It is rehearsal, and it fails for two reasons. The first is that experienced auditors recognise it immediately. The second, and more costly, is that it works against the only thing a customer audit is genuinely useful for — finding out what is actually wrong while someone competent is standing in your factory looking for it.
What rehearsal looks like from the other side of the clipboard
Auditors are not clairvoyant. They are pattern-matchers who have walked several hundred sites, and staged preparation leaves consistent traces:
- A record set completed in one sitting. Six weeks of daily checks in the same pen, the same handwriting, the same spacing, with no variation in ink or pressure. Genuine daily records look untidy because they were made on different days by different people.
- Selective freshness. A newly painted wall in the area the last audit criticised, with the adjacent area untouched.
- Answers that arrive too smoothly. An operator delivering a rehearsed sentence about a critical limit, then being unable to answer a follow-up question one step to the side of it.
- A manager who answers for everyone. The single most common tell, and the one that most reliably escalates scrutiny.
- Absent recent records. The current week missing while the previous six months are immaculate, because the tidying stopped before it caught up.
- Unnaturally new PPE on people who are visibly unaccustomed to wearing it.
The consequence is disproportionate. An auditor who concludes that the picture has been arranged stops sampling and starts digging. Every subsequent answer is weighed against the assumption that presentation is being managed. A site that would have received three minor findings on an honest day receives a deeper audit and a worse outcome — and, more damagingly, loses the presumption of good faith that makes the next audit easier.
Real preparation starts six weeks out and is mostly internal
The genuine version of preparation is not about the customer's visit. It is about knowing what they will find before they arrive.
Audit yourself against their standard, not a generic one. Customer audits are conducted against the customer's specification, which routinely contains requirements beyond any certification scheme — allergen thresholds, specific testing regimes, packaging or labelling requirements, ethical or provenance clauses. Obtain the actual document you will be audited against and audit against that. A site well prepared for a GFSI-scheme audit can still fail a customer audit on the twelve clauses unique to that customer.
Find your own gaps deliberately. Assign someone other than the person responsible for each area to look at it critically. The purpose is to arrive at a written list of known weaknesses, in your own hands, before the audit.
Then do the thing that feels counterintuitive: keep the list.
Disclosure beats concealment, reliably
A known gap with a dated action plan and evidence of progress is not a weakness in a customer audit. It is the strongest evidence available that your management system works.
Raising it yourself at the opening meeting — "we identified this in our internal audit in August, here is the root cause, here is the action, here is where it currently sits" — does three things a concealed gap cannot. It demonstrates that internal detection functions. It removes the finding's power to surprise. And it establishes, in the first ten minutes, that what you say can be relied upon, which changes the character of everything that follows.
The alternative is that the auditor finds it. Now you have the same gap plus a demonstrated failure to detect or disclose it, and every other area gets examined with fresh suspicion.
There is a limit to this, and it is worth being clear-eyed: disclosure works when the gap has a credible plan attached. Disclosing an unaddressed problem you have known about for a year is not transparency, it is confession, and it will be treated as such.
Get the mechanics out of the way
A surprising proportion of audit friction is logistical rather than technical, and all of it is avoidable:
- Document control current. Correct versions in use, obsolete copies removed from the floor, the document register matching reality. Uncontrolled documents at workstations are among the most common findings and among the easiest to prevent.
- Records findable. Decide in advance where each record type lives and who retrieves it. Time spent hunting reads as disorganisation and eats audit time that then comes out of areas you would rather have discussed properly.
- Traceability exercise recently completed, genuinely, with the result documented — including anything it exposed.
- Previous findings closed with evidence. The first thing most auditors check is the status of what they or a predecessor raised last time. An unclosed prior finding sets the tone for the entire day.
- Organisation chart, scope and process flow current, because the opening meeting starts there.
Brief staff on honesty, not on answers
The briefing that helps is short and contains no scripted content:
- Answer only what you are asked. Do not volunteer a tour of adjacent topics.
- If you do not know, say you do not know, and say who does. A guess that turns out to be wrong is far more damaging than an honest "I'd have to check."
- Do not speculate about why something is the way it is.
- If you disagree with something an auditor says, say so politely and explain why. Auditors do get things wrong.
Coaching people to recite answers produces exactly the brittleness auditors probe for. Telling people it is safe to say "I don't know" produces confident, credible staff — and it is also true, which matters, because they will be able to tell the difference.
On the day
Do not over-escort. One host, not a delegation. A crowd around an auditor reads as management of the narrative.
Let operators answer their own work. The impulse to rescue someone mid-answer is strong and almost always harmful. If an operator genuinely cannot answer something they should know, that is a real finding, and interrupting does not remove it — it just adds a second observation about management behaviour.
Do not argue findings in the moment. Ask two questions instead: what evidence did you observe, and which requirement does it fail against? Either the answer is sound, in which case arguing wastes goodwill, or it is not, in which case the questions expose that far more effectively than disagreement.
Take your own notes of every observation raised, so your corrective action work starts from your record rather than from a report that arrives a fortnight later.
The audit does not end at the closing meeting
The corrective action response is part of the impression. A response with a genuine root cause, a system-level action and defined verification criteria tells the customer their audit was taken seriously. A response saying "staff retrained" tells them it was processed.
For a customer relationship, the response is frequently more influential than the findings themselves. Findings are expected. A weak response is what gets remembered.
Practical takeaways
- Audit against the customer's actual specification, not a generic standard.
- Build a written list of your own known gaps, with actions and dates, before the audit.
- Disclose them at the opening meeting — but only where a credible plan is attached.
- Never back-fill records. It is detectable, and it converts a minor finding into a credibility problem.
- Brief staff to be honest, not fluent. "I don't know, but Sarah does" is a good answer.
- One escort, and let operators speak for their own work.
- Treat the corrective action response as part of the audit, because the customer does.
The most valuable thing a customer audit can give you is an accurate picture of your own operation, produced by someone with no stake in flattering you. Rehearsing for it is paying for that and then refusing delivery.