Corrective & Preventive Actions (CAPA)

Why findings die in reports

Every assurance activity a property runs produces findings. Inspections, audits, incident investigations, guest complaints — all of them generate a list of things that are wrong. And without a loop that carries a finding through to a verified fix, the list is where they stop.

So the alarm device that failed at the annual test is still dead when the authority arrives. The water-temperature exceedance appears quarter after quarter, logged each time. The guest complaint marked resolved recurs, because the fix addressed the symptom and nobody asked what produced it.

CAPA is borrowed from formal quality management — ISO 9001’s Clause 10.2, and the regulatory backbone of the pharmaceutical industry — for exactly this reason. It is not extra process. It is the closed loop those industries were forced to build when the cost of a finding that quietly died in a report was measured in lives.

The distinctions that make it real

• Correction, corrective action, preventive action — and closing a CAPA with a correction is the most common counterfeit there is. The device failed. The correction is to fix the device. The corrective action asks why devices like this fail here — is it the procurement specification, a gap in the maintenance regime, the environment they sit in? The preventive action takes that answer to the six identical devices at the sister property, before any of them fail. Do only the first and you have a closed ticket, an unchanged system, and a paper trail that says you handled it. Auditors probe this distinction precisely because it is where the honest programmes separate from the theatrical ones.

• Root cause, proportionate to consequence. Five Whys is enough for most findings, and pretending otherwise turns the method into a ritual. Where the consequences warrant it, go deeper — the same doctrine the incident investigation process uses, for the same reason: the depth of the analysis should match what is at stake, not what is in the template.

• Effectiveness verification is what closes the loop, and it is the step that gets dropped. ISO 9001 explicitly requires reviewing whether the action worked — after enough time has passed for the answer to mean something. A CAPA closed at “action taken” is not a control; it is an assertion. And the organisations that skip this step do not find out that their fixes are not working, which is precisely why their problems keep coming back with a full file of closures behind them.

• Auto-raise from failures, or the whole thing is discretionary. Every failed inspection line item, every audit finding, every defined complaint pattern should raise a tracked CAPA with an owner and a due date — automatically, by rule rather than by mood. This single mechanism is the entire difference between an inspection programme and inspection theatre. Discretionary raising is not a lighter-touch version of the same control. It is how findings escape.

• And the CAPA data itself is a management signal. Ageing open actions by property. Repeat root-cause categories. Effectiveness-failure rates. Rising repeat findings alongside fast closure times is the signature of a fake-closure culture, and it is legible from the data long before it is legible from the operation. This tells an owner more about the health of a property than most KPI packs manage.

What it costs to run without one

Findings are logged and never actioned, so the same finding arrives next cycle — and when the authority or the insurer finally acts on it, your own log is the document proving you knew.

Corrections masquerade as corrective actions, so failures recur, each one trailing a tidy history of closures. With no verification step, the actions that did not work are counted as wins. And with no preventive propagation, the same failure is discovered and paid for separately at every property in the portfolio, as though each one were the first.

How Zepth runs CAPA

CAPAs raise automatically from failed inspections and audit findings, carrying an owner, a due date and an escalation path. Closure requires evidence rather than assertion, and every CAPA carries a scheduled effectiveness check — so the loop cannot be closed on a promise.

The dashboards show open, ageing and repeat CAPAs by property, asset class and root-cause category, across the portfolio. Which is where the same root cause appearing at four of your eleven hotels stops being four separate problems.