Inside Our New Free Audit Engine: 15 Live Checks, No Guesswork

Most free website audits are lead magnets wearing a lab coat. You type in a URL, wait for a progress bar, and receive a score that was never really measured: a generic checklist with your domain pasted on top. We rebuilt the NetWebMedia audit engine this month around the opposite rule. Every number in the report comes from a live measurement of your site, and when the engine cannot measure something, it says so instead of guessing. You can run it yourself on the free audit page. It takes about a minute and needs no login.

What does the audit engine actually measure?

The engine fetches your page the way a visitor or crawler would, then scores it on 15 dimensions. Each dimension returns a score from 0 to 100, a pass, warn or fail status, the evidence it found, and the specific fix.

• AI search visibility (AEO): whether your domain shows up in live search results for the questions customers in your industry and city actually ask, and whether an AI assistant recommends you or your competitors.
• Mobile speed: real Core Web Vitals from Google PageSpeed Insights, on mobile and desktop.
• Schema markup: which JSON-LD types you publish and whether your business listing has the fields that matter, such as address, phone and opening hours.
• Lead capture: working forms and CRM embeds.
• Review management: review and rating markup and visible star ratings.
• WhatsApp integration: chat links and widgets, including ones built by an external script.
• Mobile conversion: viewport setup, tap-to-call and sticky calls to action.
• Automation: chat widgets, exit-intent tools and email tracking.
• Analytics and tracking: GA4, Tag Manager and ad pixels, including tags loaded from your own script files.
• Social media presence: profile links, checked to confirm they resolve.
• Content depth: sitemap size, a blog, article markup and publish dates.
• Branding: Open Graph tags, favicon, title and meta description.
• Reputation: team, client, press and portfolio signals, plus a Google Business link.
• Crawlability: robots.txt, HTTPS, noindex tags and a declared sitemap.
• Trust and compliance: privacy and terms pages, cookie consent, a declared page language, image alt text and semantic page structure.

The dimensions are weighted by how directly they turn into leads, so mobile speed and lead capture count for more than branding. The result is one score out of 100, compared against benchmarks for small businesses.

How does it check whether AI assistants recommend you?

This is the dimension we get asked about most. Ranking on Google and being named by an AI assistant are no longer the same thing. The engine picks niche-specific questions for your industry and city, such as the questions a buyer would type when looking for a real estate broker or a clinic nearby, and asks them in the language of your report. It records whether your domain is among the recommendations and which competitors are named instead. If the AI sample cannot run, the engine leaves that part out and scores the rest. It never fills the gap with an estimate.

English and Spanish audits now use their own query sets. Earlier this month, English-language audits were still being scored against Chilean search queries, because the language setting never reached the engine. That was fixed, and cached reports built on the wrong market were thrown out.

What is new in this version?

1. A 15th dimension, trust and compliance. It has no benchmark yet on purpose: we have no audit history for it, and a made-up average next to real ones would be worse than no comparison.
2. Opportunity findings, kept separate from the score. A dimension answers how good something is. An opportunity answers what is costing you right now: a website builder subscription you may not need, a certificate that is actually about to lapse, a domain nearing expiry, no booking system in an appointment business, no way to pay in a business that sells, stale content, trackers running without consent, broken internal links, or an outdated WordPress install.
3. A downloadable PDF report in English or Spanish. The server rebuilds it from the stored audit, so a report cannot be forged with a score the engine never gave.
4. A side-by-side mode that runs the same checks on a competitor's site so you can compare.

Why does the engine print nothing it cannot measure?

A report that opens with a confident, checkable falsehood destroys trust in every true finding underneath it. So the engine treats "could not measure" as a real answer. A blank result means we could not measure it. It is never a guess.

We held the engine to that by running it on our own sites first. It flagged problems that were not there, and each one taught us something about how real websites are built:

• It said netwebmedia.com could not take payments, because the checkout lives on the cart page rather than the homepage. It now counts a linked store, cart or booking page.
• It said we tracked visitors without consent, because our cookie banner is drawn by JavaScript after the page loads. It now looks for consent scripts and a linked cookie policy before raising that alarm.
• It missed our WhatsApp button and our analytics tag, because both are loaded from separate script files. It now follows a small, capped number of same-site scripts before calling something missing.
• It flagged auto-renewing certificates from issuers such as Let's Encrypt as an emergency weeks before renewal, which is how those certificates normally work. It now reads the issuer and only raises the alarm when renewal has actually failed.
• It missed about, blog and terms pages on sites that use relative links. It now recognizes both styles of link.

Every one of those fixes shipped with a regression test, and the audit tests block our deploys. If a future change brings back a false finding, the change does not go live.

What should you do with your results?

Start with the top three priorities the report lists for the next 30 days. They come from your three weakest dimensions, each with its concrete fix, so you start where the gap is widest. Most fixes on the list, such as completing your business schema, adding a cookie notice or declaring a sitemap in robots.txt, are things your current developer can do from the fix text alone.

If you would rather talk it through, the results page offers a free 30-minute walkthrough, a deeper AI visibility scan written up by a strategist, or a message to us on WhatsApp. Or just fix what the report found and run it again to see the change. Run your free 15-dimension audit.

← Back to all articles