Nobody on the team could explain why the homepage had an autoplay video of a warehouse forklift driving back and forth, when we first raised it on a screen-share call and asked the straightforward question of who had put it there and why. It had been added eighteen months earlier by a web designer who'd since moved on, meant to convey "scale and professionalism," and left running in the background ever since. It was also, we discovered, the main reason a Maldon retailer's Core Web Vitals scores had been failing for over a year, quietly, with nobody connecting the two.
What Core Web Vitals Actually Measure and Why They Matter
Core Web Vitals are Google's three headline measures of real-world page experience: how fast the main content loads, how quickly the page responds to the first interaction, and how much the layout shifts around while loading. They've been a confirmed ranking factor since 2021, though a relatively modest one on their own — content relevance still matters more. Where they really bite is on borderline rankings, the pages sitting just outside page one where every small signal can tip the balance either way, for better or worse.
An autoplay video hits two of the three metrics at once. It competes for bandwidth with the actual page content during load, delaying the largest visible element from rendering. And if it loads before its container has a fixed size, it causes the layout to jump as it appears, which is precisely what the "layout shift" metric is designed to penalise, sometimes quite severely depending on how far down the page the jump happens.
The third metric, interaction responsiveness, wasn't badly affected here, mostly by luck rather than design. Had the video also been set up to pause and resume based on a scroll listener, which we've seen elsewhere, it would likely have dragged that score down too, turning one forgotten element into a clean sweep of failures across all three measures at once.
Why Nobody Noticed for So Long
On a fast office broadband connection, sitting a few feet from a router with fibre running straight into the building, the video barely registered as a delay. On a customer's phone, on patchy 4G somewhere between Maldon and the coast, it was an entirely different experience: several seconds of a half-loaded page before anything usable appeared. Nobody at the business had ever tested their own site on mobile, on a slow connection, which is an easy thing to never get around to when the office wifi makes everything look instant and nobody has a reason to think otherwise.
This gap between how a site feels to the people who built it and how it feels to an actual customer is one of the most common blind spots we run into. It isn't laziness, usually. It's just that testing your own site properly requires deliberately simulating conditions you personally never experience day to day in a well-connected office. We now make a point of testing every client site on a throttled connection as standard, specifically because office conditions flatter almost every site equally and tell you almost nothing useful about the experience that actually matters.
The Numbers, for Anyone Who Likes Numbers
Google's PageSpeed Insights flagged a Largest Contentful Paint of 6.2 seconds on mobile, well outside the "good" threshold of 2.5 seconds. Cumulative Layout Shift sat at 0.34, against a target of under 0.1. Both failures traced back almost entirely to that one video element once we isolated it using Chrome's performance profiling tools, with everything else on the page scoring comfortably within range on its own. We confirmed this by temporarily disabling the video in a staging copy of the site and rerunning the test; scores jumped into the "good" range within minutes, which removed any doubt about where the problem actually sat.
- Largest Contentful Paint: 6.2 seconds (target under 2.5s)
- Cumulative Layout Shift: 0.34 (target under 0.1)
- Mobile bounce rate on homepage: roughly 68 percent
That bounce rate number is the one that actually got the business owner's attention, more than any technical score ever could, because it translated directly, plainly, into real people leaving the site before they'd even had a chance to see a single product on the shelf.
What We Did, and What We Didn't Do
We didn't remove the video entirely, because the business genuinely liked having some sense of motion on the homepage and there wasn't a strong reason to fight that preference, and frankly a fully static homepage would have been a harder sell to a client who'd specifically asked for something that felt lively and modern. Instead we replaced autoplay with a static poster image and a play button, compressed the video file by roughly 70 percent without visible quality loss, and set an explicit fixed-size container so nothing shifted when it eventually did load, whenever a visitor chose to press play.
Scores improved within a fortnight of deployment. Largest Contentful Paint dropped to 2.1 seconds. Layout shift fell to 0.04. The video is still there for anyone who wants to watch a forklift drive around a warehouse. It just no longer loads itself uninvited and drags the whole page down with it for every single visitor whether they wanted to see it or not. Mobile bounce rate on the homepage fell from that 68 percent figure to just under 50 percent within six weeks, which is the kind of change that shows up clearly in a conversion report, not just a technical one.
This Pattern Shows Up Everywhere, Not Just Video
Autoplay video is simply the most dramatic, easiest to spot example, but the same underlying mistake, loading something heavy and entirely unrequested before the page's actual content, shows up constantly in smaller, quieter forms too: oversized hero images never compressed, embedded social media widgets loading before anything else, pop-up newsletter forms firing mid-load and shifting everything beneath them just as someone tries to read the headline. None of these feel like SEO problems to the person adding them. All of them are, in roughly the same way, just at a smaller scale than one forgotten forklift video. A cookie consent banner that loads late and pushes the whole page down by eighty pixels causes exactly the same kind of layout shift penalty, and we see that specific issue almost as often as we see oversized images.
Making This Someone's Actual Job
Core Web Vitals aren't a one-time fix; new elements get added over time by well-meaning people who've never heard of layout shift, and scores quietly degrade again unless someone checks periodically. A fresh pair of eyes, checking every quarter rather than once a year, catches most of these before they've had time to properly bed in and start costing real traffic. We run this as a recurring check for clients rather than a one-off report, because the alternative is finding another forgotten forklift video eighteen months from now, on a different site, with a different animal or vehicle involved. It's a core part of the technical SEO monitoring we provide for Essex businesses, precisely because this kind of thing never announces itself until the traffic numbers do. We'd encourage anyone reviewing their own site performance to start with a genuine mobile test on an average connection, not the office wifi, before assuming everything's fine.
The Bottom Line
A single forgotten element, added with entirely good intentions by someone long since gone from the business, can quietly cap your rankings for well over a full year without anyone ever properly noticing the connection. Test your site the way a real customer genuinely experiences it, on mobile, on an average connection, not on the office wifi, and you'll usually find the culprit much faster than you'd expect, hiding in plain sight the entire time.
Looking for SEO agency in Essex?
Let's talk about what your business actually needs
Digital Impact Ventures provides SEO, AI visibility and digital marketing support for businesses across Colchester and Essex. Get in touch and we will talk through where your business is now and where it needs to go.