Before writing this article, we did something almost no agency does publicly: we audited our own site, with the same tools and the same rigor we’d apply to a client. We didn’t use a plugin or a generic automated report — we reviewed the actual HTML sent to the browser, with curl, line by line.
We found three defects. They weren’t cosmetic — they were the kind that cost a site real rankings, every single day, without anyone noticing they’re there. We’re sharing exact numbers because that’s more useful than a generic list of “best practices.”
Defect 1: zero meta descriptions, period
We reviewed six main pages of the site (home, the three service pages, blog, and contact). None of them had a <meta name="description">. Not a poorly written one, not a duplicated one — completely absent, verified with grep against the raw HTML.
Why it matters: without a meta description, Google builds its own snippet from the page’s content. Sometimes it gets it right, sometimes it doesn’t — but either way, you lose control over what people read in the search result before they click. It’s the difference between controlling your first impression and leaving it to chance.
The fix isn’t complicated: a sentence of 120 to 160 characters, written for humans, describing that specific page’s real benefit. The hard part isn’t writing it — it’s making sure it exists on all six pages, not just one.
Defect 2: 5 of 6 pages with no H1 at all
This one surprised us the most. Searching for the <h1> tag in the HTML of five of the six pages, we found none. The text that visually functioned as the main heading was marked up as <h2 class="elementor-heading-title"> — a design-plugin class, not a real semantic tag.
Why it matters: the H1 is one of the signals a search engine uses to understand what a page is about. Not having one isn’t disqualifying by itself, but combined with the missing meta description (defect 1), it strips the content-understanding system of the two most basic signals that exist.
What this reveals: it wasn’t a one-off mistake on a single page — it was a pattern repeated across five different pages, generated by the same template. That’s the real problem with building sites on page builders (Elementor, Divi, Gutenberg): a template defect multiplies automatically across every new page, silently.
Defect 3: 1.9 seconds just for the first byte
We measured the TTFB (Time To First Byte — how long the server takes to start responding, before a single image loads) of the homepage: 1.9 seconds, confirmed across two independent measurements.
For context: Google considers an LCP (the moment the main content becomes visible) of 2.5 seconds or less to be “good.” With a TTFB of 1.9s, more than three-quarters of that budget is already gone before a single image has loaded.
Why it matters: this isn’t just user experience — Core Web Vitals is a direct ranking signal. A slow site competes at a disadvantage even when the content is excellent.
What we didn’t find (and it’s worth saying)
Not everything was broken: the site did have HTTPS with HSTS correctly configured, a self-referencing canonical with no contradictory signals, and meta robots with no improper blocks. This isn’t a case of “everything’s broken” — it’s a case of specific, targeted defects with an impact disproportionate to how simple they are to fix.
What we did next
We didn’t stop at diagnosis. We rebuilt the site fixing exactly these three issues — every page with its own H1, its own meta description, and an architecture without the weight of a page builder — and the resulting HTML went from 290 KB to 5.7 KB on the homepage. Same content, same navigation structure, 98% lighter.
What to check first on your own site
If you can only check three things today, make it these — in this order:
- Does every page have a unique
<meta name="description">? Open it, right-click → “View page source,” search fordescription. - Does every page have exactly one
<h1>? Same page source, search for<h1. - How long does your server take to respond? Any PageSpeed tool gives you the TTFB in seconds.
These are three questions you can answer in five minutes, and in our own case, they explained a significant part of why the site wasn’t performing the way it should have.
Want us to run this same audit on your site? Contact us and we’ll tell you exactly what we find, with the same kind of evidence shown here.