This is the method behind our ten-day operations audit, written out in full. It is not a secret. The value of a paid audit is the time, the access, and the accountability, not the checklist. If your team can run this yourselves, you should.
The core mistake we see in WordPress audits is starting with the software. Plugin lists, PageSpeed scores, and PHP versions are easy to collect and mostly beside the point. A content operation fails in the space between people and systems, so that is where the audit has to start.
Step one: map the workflow as it actually runs
Sit with the people who do the work, not their managers, and trace one real item end to end: one article, one product, one order. Write down every tool it touches, every person who handles it, and every place it waits. The waits matter most. A publishing workflow that takes four days of calendar time usually contains about forty minutes of actual work.
Two questions expose most of the cost:
- Where does anyone re-type or copy-paste information that already exists in another system?
- Where does an item wait for a person whose only contribution is forwarding it?
Every yes is a cost you can estimate: people multiplied by minutes multiplied by frequency. In our audits this table, not the technical findings, is what changes the buying decision. When a client sees that their team spends roughly ten hours a week moving order data from WooCommerce into a spreadsheet by hand, the conversation stops being about technology.
Step two: review the platform against the map, not against best practice
Now look at the WordPress install, and only now. The question is never “is this set up well” in the abstract. It is “does this setup explain the waits and the re-typing we just mapped”. A technically messy install that supports a smooth workflow is a lower priority than a clean install that forces staff through five tools to publish a page.
What we actually check, in order of how often it turns out to matter:
- The editing experience. Can a non-technical person compose a real page from the blocks they are given, or do they file a ticket? Count the tickets.
- The integrations. For each connected system: what happens when it fails, who notices, and how long has it been silently broken before? Almost every operation has one integration nobody trusts.
- The content model. Structured content in fields the team understands, or one giant page builder blob per page? This decides how expensive every future change is.
- Release safety. Can they deploy or update without fear? A team that avoids updating anything is telling you the platform owns them.
- Performance and reliability, measured, not assumed. Field data where it exists, lab data where it does not, and never a score without the metric behind it.
Step three: rank by cost, and show your reasoning
Every finding gets the same treatment: what it costs today, what fixing it costs, and how confident we are in both numbers. Then sort. The output is deliberately boring, a ranked table with the reasoning visible, because the reader has to be able to disagree with it. An audit whose ranking cannot be argued with is hiding its assumptions.
Honesty rules we hold ourselves to, and you should hold any auditor to:
- When a plugin or an off-the-shelf tool is the right fix, the plan says so. Custom code has to earn its place.
- When a number is an estimate, it is labeled as one, with the basis shown.
- When something is fine, the report says it is fine. A finding count is not a quality score.
Step four: write the plan for a team that is not you
The last deliverable is a 90-day implementation plan written so that any competent team could execute it, including one that is not ours. That constraint is not generosity, it is quality control. A plan that only its author can execute is a proposal wearing a costume.
Sequence it by dependency and by payback: the first two weeks should remove a visible, daily irritation, because momentum is an engineering resource like any other. End with the measurement: for each change, the number you expect to move and when you will check it.
That is the whole method. Map the work, review the platform against the map, rank by cost with the reasoning showing, and write a plan you would be comfortable handing to a competitor. It takes us ten business days because the conversations take what they take. The thinking, as you can see, fits on one page.