Website & App Audit

Buy the diagnosis before you buy the operation

The standard way this goes: a merchant senses the store is underperforming, asks three agencies to take a look, and receives three free audits. Each one is between four and nine slides. Each one concludes that a rebuild is required. Each one is followed by a proposal for that rebuild, from the agency that wrote it.

None of those agencies were lying. They were doing unpaid sales work, and unpaid sales work has exactly one possible conclusion. Nobody spends two days investigating for free and then recommends the cheap fix.

A paid audit changes the incentive, and that is the entire product. We are paid for the diagnosis. We are paid the same whether the answer is “rebuild this” or “you have four apps doing two jobs, remove two of them and re-encode your hero images”. Roughly the second of those is what we conclude a great deal of the time, and we can afford to.

It also changes what you get. Not slides — a written report, with the evidence attached to each finding, in an order you can start at the top of, that can be handed to any competent developer including one who has never met us.

What gets measured

Performance. Core Web Vitals from field data — what real visitors on real networks experienced over a rolling window — alongside lab traces that explain why. Largest Contentful Paint under 2.5 seconds, Interaction to Next Paint under 200 milliseconds, Cumulative Layout Shift under 0.1. Then the part that makes it actionable: every third-party script, app and image attributed to the template it loads on and the milliseconds it costs there. “The site is slow” is not a finding. “The review widget costs 380ms on the product template and is used on nothing else” is.

Technical SEO and AEO. Crawl and index coverage, canonicals, faceted navigation, internal link structure, and structured data. Then the question almost nobody’s audit asks: can a retrieval system extract anything useful from these pages? Is there an entity here it could resolve? Is there a passage it could quote? Does the page render without JavaScript?

Conversion and UX. Template by template, on a real phone, with the checkout completed end to end using a real card. Where the journey leaks. What the forms ask for that they do not need. What moves under a thumb half a second after the page paints. This part is deliberately not a heuristic checklist — it is someone attempting to buy from your store and writing down where it got annoying.

Accessibility. Contrast, focus order, heading structure, form labelling, target sizes, against WCAG 2.2 AA. Worth noting that the criteria most often failed — small tap targets, unlabelled inputs, focus you cannot see — are the same ones costing you orders from people who are not disabled at all.

Code and architecture. The theme, the app, or the repository, read rather than skimmed. What is maintainable. What is load-bearing and undocumented. What would have to be rewritten before anything new can be added safely. This is the section that tells you whether the next project is a three-week job or a three-month one, and it is the section that is missing from every free audit.

Apps and integrations. Everything installed, what it costs per month, what it loads on which template, and which of them are doing the same job twice. This section frequently pays for the audit on its own.

What the report looks like

Findings, not observations. Each one has evidence attached — a trace, a screenshot, a crawl export, a line of code — so that anyone who disagrees can check rather than argue.

Each one is scored on impact against effort, and the list is ordered by that score rather than by the section it came from. You should be able to start at the top and work down until the budget runs out, and know that what you did was the most valuable available.

And the raw data comes with it. Traces, crawls, exports. Not because most clients will open them, but because a report whose evidence cannot be checked is an opinion with formatting.

Who this is for

Merchants who have been quoted for a rebuild and want to know whether they need one. Teams who have inherited a store and do not know what is in it. Businesses whose traffic is fine and whose revenue is not. Founders about to acquire a store who want to know what they are buying. And technical leads who already know something is wrong and need it written down in a form a budget holder will act on.

It is the right first engagement for almost anyone, which is why it is priced to be one — a fixed fee, a fixed timeline, and no obligation attached to either end of it.

The honest limits

An audit does not fix anything. It is a diagnosis, and if nobody has capacity to act on it this quarter then it will be out of date before it is read. Buy it when someone can act.

It will not tell you what your customers want. That is research — interviews, session recordings, testing — and it is a different engagement with different methods. An audit tells you where your store is failing on its own terms.

And it will not always tell you what you were hoping to hear. Sometimes the finding is that the platform choice was wrong two years ago. Sometimes it is that the store is fine and the marketing is the problem, which is outside our scope and we will say so rather than inventing findings to fill the section.

After

The report is yours. Act on it with us, with your own team, or with somebody else entirely — there is no clause in it that says otherwise, and it is written to be handed over rather than to be interpretable only by the people who wrote it.

If you do want us to do the work, the audit becomes the scope document, which is a considerably better starting point than a discovery phase billed at the front of a project. And if the conclusion was that you do not need a project at all, that is a good outcome too. It is the one that gets us recommended.