Growth audit

A growth audit follows the route a buyer takes from first contact to a useful commercial outcome: a search, an AI answer, a post or a referral, then the website, the form, the follow-up and the sales call. It starts from one commercial question, because too few of the right visitors, visitors who leave and enquiries that go nowhere each need a different fix.

The evidence is channel activity and cost for a set period, conversion events with their definitions, enquiries by source and what sales hears about fit. Findability checks come first because each has a yes or no answer: is the page open to indexing, in the sitemap, linked from somewhere a visitor goes. Reading the pages as a buyer would, on a phone as well, takes longer.

Each finding at the end names the problem, the evidence, the proposed change and the measure that would show progress, with the effort and any dependencies.

Each team reports on its own stretch, and the faults sit at the handovers. An ad earns clicks while the landing page leaves the offer unclear. A form produces leads and the follow-up arrives too late to be useful. Sales turns away the audience the campaign was asked to attract. The usual response is more ad spend or a new website, and neither repairs a handover. The audit finds the constraint before the money goes anywhere.

If you would like to try working with us before committing, our growth audit runs one week for $2,500, and the amount comes off work that starts within 60 days. It covers your website, search presence and content in four steps: find where your attention comes from, read the site as a buyer would, rank the fixes by effort and effect with the first drafted, and write a plan for search and content that you keep.

The audit we can show is our own. In November 2026 we audited refract.group and found signed-off service pages still marked noindex, case-study titles that were the client’s name alone, and old addresses returning not found. Those were repairs, each checked by requesting the URL. Slow text on phones, held back by the hero’s entrance animation, was recorded as an open design decision with one option written down.

Say a data infrastructure company has steady search traffic and few demo requests. The yes or no checks find its integrations pages marked noindex since a redesign: a repair, fixed and checked that week. Read as an engineer would, the homepage never names the databases it works with: a hypothesis, measured by demo requests over two months. Sales notes say most demos go to students, and whether to gate the free tier is a decision for the founder.