How to Run a Website Audit That Actually Sells the Work

The usual audit is forty pages generated by a tool, full of red warnings the client cannot interpret, delivered with "let me know if you'd like to discuss". The client reads two pages, feels vaguely criticised, and does nothing. That is not an audit problem — it's a translation problem.

A finding is not a reason to buy

"Missing alt attributes on 34 images" is true and entirely inert. The client has no way to judge whether it matters, what it costs them, or what happens if they ignore it.

Every finding needs three parts: what's wrong, what it costs in their terms, and what fixing it involves. Without the middle part, you've handed over homework rather than a decision.

Lead with the one thing, not the list

Length signals thoroughness to you and overwhelm to them. The audit that sells opens with the single most expensive problem, explained in two sentences, with everything else beneath it.

"Your booking form doesn't work on an iPhone. Roughly two-thirds of your visitors are on a phone, so most people who try to book can't." That's the whole opening. The other eleven findings can wait — they are supporting evidence, not the argument.

What's actually worth checking

For a small business site, the findings that matter commercially are few:

• Does the main action work on a phone? Form, booking, call button. Test it, don't assume
• How long does it take to load on mobile data? Not on your fibre connection
• Is it HTTPS? The browser warning alone costs trust
• Is there a page for each thing they sell? One page cannot rank for six services
• Do page titles differ, and do they name the town?
• Is the phone number a tappable link? Astonishingly often it isn't
• Is the content current? Old prices, departed staff, a copyright year from four years ago

Seven checks, most of them visible in a few minutes, and all of them explainable without jargon.

Record it rather than writing it

A three-minute screen recording beats a document for this. The client watches you try to use their own site on a phone and fail — and seeing it is categorically different from reading that it happens.

Keep it short, show rather than narrate, and stop before you start listing minor issues. The video's job is to make one problem undeniable.

Charge for it, or give it away?

Both work, for different situations.

Free suits cold outreach, where you need to earn attention before anyone will pay for anything. Keep it short — a free audit that took you four hours is a bad trade.

Paid suits warm prospects who approached you. A few hundred for a proper audit filters out tyre-kickers and funds the work, and it's usually credited against the project if they proceed.

What doesn't work is a long free audit for a prospect who never asked. That's unpaid consulting with a thank-you at the end.

End with a decision, not an offer to discuss

"Happy to discuss" puts the next move on the client, and the next move is usually nothing. End with something they can answer in one word: "Shall I fix the form first? It's about half a day."

Small, specific, immediate. The rest of the findings become the second conversation, which is far easier than the first.

Auditing enough businesses to make it worthwhile

The economics only work if most of the businesses you audit turn out to have a real problem. Audit twenty at random and you'll find fourteen sites that are basically fine.

Scoring a whole trade's websites first — mobile behaviour, speed, HTTPS, metadata — tells you which ones are worth the manual half-hour. The audit stays personal and specific; the screening in front of it doesn't have to be.