How to properly audit a page in 2026
Page audits used to be a checklist. In 2026, with AI Overviews, entity grounding, and shifting click behaviour, the checklist isn't enough. Here's a tighter framework.
Page audits used to be a tidy checklist: title, H1, internal links, schema, done. In 2026, that checklist still runs — but it falls short of the most important stuff. AI Overviews rewrite the SERP these days. Entity grounding decides whether your page even qualifies. And the click you're optimising for might never happen. In this article, we'll walk through a four-layer audit you can run on any page in roughly an hour. Layers 1 and 2 we'll automate with our free 2026 SEO Audit — it crawls up to 50 pages and grades the technical and entity layer. Layers 3 and 4 are still hand-rolled, because nobody can tell you what your page is for except you.
Think about a page audit as a stack, not a checklist
Four layers, in order. Skip one and the layers above it stop making sense.
Layer 1 — Crawl and render
Open the page in Search Console's URL Inspection tool. Confirm it's indexed, confirm the rendered HTML matches what you wrote, confirm canonical and hreflang resolve cleanly. If JavaScript is hiding your main content, nothing below this layer matters.
Our SEO audit covers most of this in one pass — HTTPS, robots.txt and sitemap.xml validity, broken internal links (4xx/5xx), canonical tags, AI crawler permissions, and Core Web Vitals from Google PageSpeed Insights. The one gap\\: we fetch server-rendered HTML\\, so JavaScript-only content still needs Search Console's URL Inspection to confirm Googlebot sees what your users see.
Layer 2 — Entity clarity
Paste the URL into a model (Claude, ChatGPT, whatever) and ask: "What is this page about, who is it for, and what entity does it represent?" If the answer is vague, the page is vague. Check that your primary entity is named in the title, the H1, the first 100 words, and the schema. Entities, not keywords.
Our audit handles the structural half: title length and uniqueness, H1 presence and count, Schema.org type coverage across the site, and whether you've shipped llms.txt for AI search engines (nah, lets just say we might need, its 2026!). It won't tell you whether the page is clear — that's the model conversation above. But it will flag the structural gaps that make clarity impossible.
Layer 3 — Query coverage
Pull the page's queries from Search Console (last 28 days). Group them by intent. Then run the top 3 queries through Google and look at the AI Overview. Is your page cited? Could it be? If the Overview answers the question and your page doesn't add anything past that answer, you have a depth problem — not a ranking problem.
This one we don't automate — Search Console data lives behind your login, and AI Overview presence has to be eyeballed. Our keyword volume checker and SERP rank checker help once you've identified the queries, but the grouping by intent and the Overview comparison are still manual.
Layer 4 — Outcome
What is this page supposed to do? Rank, convert, get cited, feed an internal link graph? Pick one primary job. Then check GA4 (Reports → Engagement → Pages and screens) against that job. A page ranking #3 with a 4% scroll depth is not winning.
No tool can audit the job for you, ours included. The closest we get is the prioritised action plan at the end of the audit — but that ranks fixes by severity, not by whether they serve the page's job. That decision stays with you.
Concluding thoughts: audit the job, not the page
Most page audits in 2026 fail because they grade the page in isolation. Pages don't exist in isolation anymore — they exist inside an AI-mediated SERP, an entity graph, and a user journey. Audit the job the page is doing. The checklist follows.
If you want a head start on Layers 1 and 2, run our free audit — it'll surface the technical and entity gaps in roughly 45 seconds. Layers 3 and 4 are on you, but you'll know exactly what to focus on by the time you get there.
Written by Khmer. Start a project →