What 'site health' actually means (and how to check yours)

A mechanic has a diagnostic computer. A doctor has blood tests. Your dentist has x-rays. They all solve the same problem: something's wrong, but you can't see it from the outside, so you need a tool that looks inside.

Your website has the same kind of diagnostics. Most business owners don't know they exist, or think they're something only developers care about. They're not. A site with "health issues" is costing you money in the same way a check engine light costs you money — slowly at first, then all at once.

Here's what a real site health check actually measures, what each number means in English, and how to tell what to fix first.

What a health check is actually measuring

Google has a tool called Lighthouse. It's what Chrome uses under the hood when you see "SEO score" or "Performance score" anywhere — including our free Site Audit, which runs Lighthouse and translates the output into English.

Lighthouse measures four things:

1. Performance — how fast your site loads and responds
2. Accessibility — whether people with disabilities can use it
3. Best Practices — whether the site follows basic web standards
4. SEO — whether Google can read and understand your site

You get a 0–100 score in each category. Here's what each one is actually looking at.

1. Performance: the math that decides if visitors stay

Performance is the big one. It's also the one most small business sites fail, because most small business sites were built on a template or WordPress theme that came with 15 plugins nobody audits.

Performance breaks down into three specific numbers Google publishes as "Core Web Vitals." I'll translate them:

LCP — Largest Contentful Paint

Plain English: how long until the main thing on the page actually shows up. If you land on a roofer's homepage and the big hero photo takes 4 seconds to appear, LCP is 4 seconds.

Target: under 2.5 seconds. Between 2.5 and 4 is "needs work." Over 4 is a real problem.

Why it matters: Google measures this on actual visitor devices — not the server. If it's slow for people, Google knows, and Google demotes you in search results. Visitors also leave. There are studies showing every extra second adds ~10% to bounce rate.

Common causes: unoptimized hero image (a 4MB photo that should be 400KB), slow hosting, too many third-party scripts loading before the main content, WordPress plugins stacking render-blocking CSS.

CLS — Cumulative Layout Shift

Plain English: how much stuff on the page jumps around while it loads. You know when you're about to tap a button and an ad loads above it and suddenly you tap the ad? That's CLS.

Target: under 0.1. Over 0.25 is bad.

Why it matters: it's maddening on mobile. Google hates it. Users hate it even if they can't name it.

Common causes: images without width/height attributes, ads loading late, fonts swapping in after page load, "hero banner rotators" that shift content.

INP — Interaction to Next Paint

Plain English: when you tap/click something, how long until the page responds.

Target: under 200ms. Over 500ms is bad.

Why it matters: nothing feels worse than a page that looks loaded but doesn't respond when you tap.

Common causes: heavy JavaScript running on every interaction, chat widgets, analytics scripts, rogue event handlers.

2. Accessibility: the score that secretly also helps SEO

Accessibility measures whether your site works for people using screen readers, keyboard navigation, high-contrast modes, or anything other than a standard mouse-plus-monitor.

The top things that drop this score:

• Images without alt text. A screen reader reads "image" instead of "photo of John fixing a furnace" — useless for a blind user, also useless for Google.
• Low color contrast. Light-gray text on a white background is a fail. Your designer thought it looked clean; a 60-year-old reading your site in sunlight can't see it.
• Forms without labels. Inputs with placeholder text but no proper label mean a screen reader announces nothing when the user tabs to the field.
• Buttons that aren't buttons. <div> with a click handler is not a button. Keyboard users can't tab to it.

The accessibility score is not legally binding in most jurisdictions (yet — this is changing), but it helps SEO because Google uses similar signals for its own crawler. A site that's hard for a screen reader is usually hard for Google's bot to understand too.

3. Best Practices: the "is this site broken" check

This is the odds-and-ends category. Top issues:

• HTTPS everywhere. If any part of your site still loads over plain http://, this fails. Chrome also shows a "Not Secure" warning to visitors, which tanks trust.
• Old JavaScript libraries with known security holes. Usually a WordPress plugin that hasn't been updated in 4 years.
• Console errors. Your site might look fine but be throwing 12 JavaScript errors on every page load. Breaks features quietly.
• Deprecated APIs. The web evolves; sites that use APIs Chrome is phasing out will stop working suddenly.

Best Practices is usually easy to fix. It's the category where "just update your WordPress plugins and check the console" gets you from 60 to 95 in an afternoon.

4. SEO: the basics Google expects

This is where we overlap with the broader SEO topic, but specifically at the technical level — not content strategy. A Lighthouse SEO check looks for:

• Page has a title tag (and it's not empty or a duplicate of another page)
• Page has a meta description (between about 120–160 characters)
• Page has an H1 (exactly one, describing what the page is about)
• Images have alt text (yes, same one from accessibility — it counts twice)
• Links have descriptive text (not "click here")
• Page is crawlable (not accidentally blocked by robots.txt or meta-noindex)
• Page is mobile-friendly (viewport meta tag, readable text, tappable targets)
• Valid hreflang if you have multiple languages
• Structured data (optional but heavily rewarded — tells Google "this is a LocalBusiness, here's the address, here's the rating")

If you want to go deeper on any of these, our post on what SEO actually is breaks down the bigger picture.

What the numbers actually mean (when to panic, when to chill)

The 0–100 scale throws people off. Here's how I actually read it:

The catch: a 70 isn't an F. The scale isn't a school grade. Most mid-tier small business sites live in the 60–80 range and function perfectly well for visitors. The goal isn't a perfect 100 — it's catching the specific issues that are affecting your business.

For most contractors, dentists, and local service businesses: if you have a green Performance score on desktop and your phone number is visible, you're fine. If you're red on Performance mobile, that's actually losing you leads.

Triage: what to fix first

When our Site Audit returns results, it groups findings by severity:

Critical — broken things. Site won't load, 500 errors, HTTPS not configured, robots.txt blocking everything. Fix today.

Major — real money on the table. Performance score under 50 on mobile, no meta descriptions, broken internal links, no H1s, missing Google Business Profile. Fix this month.

Minor — polish. Alt text on decorative images, minor console warnings, opportunities to defer non-critical scripts. Fix when you have time or when you're already in the codebase.

The trap is fixing minor stuff because it's easier while ignoring major stuff because it's harder. A site going from 85 to 95 on desktop is a nice-to-have. A site going from 30 to 60 on mobile is a revenue event.

How to actually check yours

Three options, in order of effort:

1. Run a free audit (30 seconds). Our Site Audit runs Lighthouse plus local-business-specific checks (GBP status, NAP consistency, competitor context). Gives you a prioritized list of fixes in English. No signup.

2. Run Chrome Lighthouse directly. Open your site in Chrome → right-click → Inspect → Lighthouse tab → Analyze. Same data, raw. You'll need to know what to do with the numbers.

3. Google PageSpeed Insights. Paste your URL into pagespeed.web.dev. Same Lighthouse underneath, plus real-world data from actual Chrome users visiting your site (if you get enough traffic).

What you want to come away with: a handful of specific, named fixes in priority order. Not a vague "we need to work on SEO." A site health check that doesn't produce a to-do list isn't a health check.

Get your site's diagnostic in 30 seconds →

Free, no account, real numbers with plain-English fixes. If something looks off in the results — like a metric that seems wrong for your business — run it twice. Some of these measurements have natural variance, and the second run usually confirms or disproves the first.

The worst thing you can do with site health is ignore it until you lose rankings and don't know why. The second worst thing is pay someone $500/mo to watch numbers move around without telling you what they did. That's the topic of the next post.