Introducing Webperf Audit, a full-site website audit

Most website checks look at one page, usually the homepage, and assume the rest of the site is fine. It rarely is. The slow page is three clicks deep. The cookie set before consent hides on a campaign landing page nobody re-tests, and a broken heading order sits in a template copied across four hundred articles.

A Webperf Audit starts from the opposite place, your whole site at once. It's a new full-site audit inside Webperf Cloud that crawls every page, runs Lighthouse across performance, accessibility, SEO and best practice, checks the privacy side (cookies, third parties, trackers, TLS), and hands you one clear, quantitative picture. Then it lets you drill from that overview straight down to the exact pages, and the exact issues, that need work.

It's the difference between our homepage scores 92 and 614 of our 2,000 pages ship the same render-blocking script, and here they are. One is a number. The other is a decision.

The mechanics are simple. Drop in a domain. Webperf Audit follows your sitemap.xml to discover URLs where there is one, and renders pages to find links where there isn't, then runs each page through the full pipeline. That makes it a whole-site website audit covering hundreds or thousands of pages, not a spot-check.

Every page is scored on all four Lighthouse categories under controlled, repeatable conditions. Consistency matters here. When you re-audit later, a change in the score comes from a change in your page, not from a busier network that day.

A distribution, not a single score

When you audit thousands of pages, one average hides everything that matters. So the first thing Webperf Audit gives you is the shape of your site. It shows how your pages are distributed across each category, from the median page through the slow 25% to the worst offenders, rather than a single number that buries the tail.

It also counts, for every issue, exactly how many pages it affects. This is the number that changes how teams work. A problem on 3 pages is a ticket. The same problem on 900 pages is a template fix, one change that makes nine hundred pages better. You can't tell those apart from a homepage score, and they call for completely different responses.

For CrUX, we go past the pass/fail badge and read the real numbers out of every page. Largest Contentful Paint (LCP), Cumulative Layout Shift (CLS), Total Blocking Time (TBT), First Contentful Paint (FCP) and Speed Index all come out as percentile distributions (p50, p75, p90, p95) against Google's official thresholds. We have an LCP problem becomes our 75th-percentile LCP is 4.1 seconds, and here are the pages dragging it over the line.

From the number to the exact page and the exact element

A quantitative overview is only half the value. The other half is being able to click a number and land on the work.

Every run produces deep-dive dashboards, one per category, that go from the aggregate straight to the specifics. Not improve your LCP, but which page, and what on it:

The overview tells you where to spend the week, and the deep-dive tells you what to open first. Together they turn an audit from a report you read into a backlog you work.

Lab scores and real-world Core Web Vitals (CrUX)

Lighthouse is lab data, every page measured under identical conditions, which is ideal for comparing pages and re-testing after a change. But lab data tells you how a page can perform, not how it does for real visitors.

So where it's available, Webperf Audit pulls in the Chrome UX Report (CrUX), Google's dataset of real, anonymised measurements from actual Chrome users, for desktop, phone and tablet, with the trend over recent months rather than a single snapshot. Seeing field data and lab data side by side tells you whether a problem shows up only on the test bench, or is genuinely costing real visitors real time. Core Web Vitals are also a Google ranking signal, so it costs you organic traffic too.

A privacy check on every page, covering cookies, trackers and TLS

Speed and accessibility are only part of what a page does. It also sets cookies, contacts third parties, and negotiates a TLS connection, usually before the visitor has agreed to anything. Webperf Audit checks all of it on every page, using our headless-Chromium Webbkoll fetcher. It amounts to a full-site privacy and GDPR audit that most performance tools simply don't run:

A privacy problem hiding on one page nobody re-tests doesn't get to stay hidden.

You own every number, dashboards and raw JSON

The audit dashboards are self-contained, HTML you can save, share and archive with no server and no login. But the visuals aren't the whole delivery.

Every run also hands you the complete underlying data as JSON. That includes the per-page results, the metric distributions, the cookie classifications, the third-party inventory, the CrUX series, and the full list of crawled URLs. The dashboard is one way to read your audit. The JSON is your audit. Feed it into your own pipeline, diff two runs, chart it your way, or draw conclusions we didn't think to surface.

That's deliberate. You should own your web data outright, not rent a view of it behind a black box. The open-source foundation and the freedom to take your data are the same principle applied to auditing. Nothing here is a number you can't inspect or take with you.

Who it's for

To get one, you order an audit as a job inside Webperf Cloud's tools. You choose a site and how many pages to cover; it crawls, runs the full performance, accessibility, SEO and privacy pipeline, and delivers the dashboards and JSON when it's done, with an email to let you know. It's also the engine behind the whole-site baseline report, so if you've read what's in the baseline, this is how it's produced.

Every plan includes Webperf Audit. Compare the plans to see how many pages each one covers, or ask us at hello@webperf.cloud what your own site's overview would show.