How the audit actually works

Your overall lead-readiness score is a weighted combination of four sub-scores. The weights reflect our judgement about what costs a trades business the most calls. They are our judgement, not a law of nature.

We run live queries against ChatGPT with web search, Perplexity and Google Gemini, phrased the way a homeowner would actually phrase them, for your trade and your service area. We record which businesses each assistant names, in what order, and what sources it cites.

Then we compare: the businesses that got named have evidence on their sites and listings. We check what that evidence is and whether you have it. The gap between the two is the finding.

Mobile Call Now speed. We load your site under a throttled mobile connection and measure how long until the Call Now button is actually tappable. Not how long until something appears. How long until a panicking homeowner could press it.

Local SEO and Google Business Profile. We fetch your site, read the structured data in it, and cross-check your name, address and phone number against your Google Business Profile and the directories that carry weight in your area.

Trust signals. We scan your public pages for the things a homeowner checks before letting someone into their house: licence numbers, insurance, warranty terms, recent reviews, verifiable certifications.

A measurement of a website and its listings, taken on a specific day. It is not a prediction of revenue and it is not a grade of your workmanship. Plenty of excellent trades businesses score badly here, which is the entire reason the product exists.

This is the section most audit tools leave out.

Why an AI chose someone. We can see who got named and what evidence they have. We cannot see inside the model. Nobody outside those companies can. When we say a factor is associated with being recommended, we mean we observed it on the businesses that were recommended, not that we proved it caused the recommendation.

What happens next week. AI answers are not deterministic. Run the same query twice and you can get two different answers. Platforms change their models and their rules without notice. A finding is true when we ran it.

Whether fixing it will work. Closing an evidence gap makes your business easier to verify and easier to cite. It does not guarantee anyone will cite it. We do not promise rankings, leads or revenue, and you should be suspicious of anyone who does.

Sample size. We test a set of representative queries, not every query a homeowner could type. A different phrasing can return a different set of names.

Any figure we publish has to come from a named, checkable source, or from our own audit data. If we cannot point at where a number came from, it does not go on the site. We removed several statistics under exactly this rule rather than keep claims we could not stand behind.

Where we cite a study, we will tell you whose it is and what it actually measured, including when it measured something adjacent to your situation rather than your situation itself.

During early access, audits are run by hand by the founder. That means slower turnaround (two business days) and more judgement applied to your specific business. When the automated scanner ships, the checks stay the same and the turnaround gets shorter. We will say clearly on the report which one produced it.