Adding Video to Your Website Without Wrecking Your Page Speed

Video is the heaviest thing most business owners add to their website, and almost none of them realise they have done it.

A photo you can see is obviously big. The video you embed hides its weight behind a small thumbnail and a play button, so it feels light. It is not. A single YouTube embed can pull more than half a megabyte of code onto your page before a visitor has clicked anything at all, and on a phone over patchy rural coverage that is the difference between a page that loads and a page someone abandons.

Consider a typical case, the kind we see regularly during site audits. A wedding venue in Westmeath pays a videographer for a beautiful two-minute cinematic film, and the obvious place to show it off is the homepage, set to play automatically the moment the page opens. The film is gorgeous. The homepage now takes the guts of ten seconds to become usable on a mid-range Android phone, which is exactly what half their couples are browsing on from the sofa at night. The bounce rate climbs and nobody connects the two things, because the video was the best part.

This is a solvable problem, and you do not have to choose between having video and having a fast site. You have to add it the right way. What follows is how the weight actually works, what it does to the scores Google measures, and the four ways to put video on a business website ranked from worst to best for speed.

Why Video Is the Heaviest Thing on Your Page

Start with the raw numbers, because they are worse than most people expect. Google's own Lighthouse documentation puts a full YouTube embedded player at around 540KB, against roughly 3KB for a static placeholder that only loads the real player when someone clicks it ^[2]. That is not a rounding error. It is close to two hundred times the weight, sitting on your page whether a single visitor ever presses play or not.

Download size is only half of it. According to HTTP Archive data referenced by web.dev, a YouTube embed blocks the browser's main thread for more than 1.7 seconds on the median site ^[4], though that figure swings a lot with the device and the connection. Blocking the main thread means the browser is too busy to respond. In plain terms, your customer taps a button and nothing happens, because the page is still wrestling with the video player's code instead of listening for the tap.

The trend is going the wrong way too. The HTTP Archive Web Almanac found the median web page shipped about 315KB of video in 2025, up from roughly 246KB the year before, an increase of around 28% in a single year ^[5]. Video is becoming the default. Page weight is climbing with it. The businesses that notice are the ones watching their load times, and the ones that do not notice are the ones losing visitors quietly.

What Video Actually Does to Core Web Vitals

Google measures real-world experience through three numbers it calls the Core Web Vitals, and video pushes on all three. The targets, measured at the 75th percentile of your visitors, are a Largest Contentful Paint under 2.5 seconds, an Interaction to Next Paint under 200 milliseconds, and a Cumulative Layout Shift under 0.1 ^[1]. Miss them and you join the large share of Irish WordPress sites that fail Google's mobile performance tests.

Largest Contentful Paint is the big one for video. If your autoplaying background video, or its poster frame, is the largest thing in view when the page loads, it becomes the element Google times. A heavy video that paints slowly drags your score straight past the 2.5 second line. Web.dev notes that a video poster image now counts as a Largest Contentful Paint candidate once painted ^[3], which cuts both ways: a well-optimised poster can score for you, and a bloated one can score against you.

Then there is layout shift. A video that loads without its dimensions declared in advance shoves everything below it down the page when it finally appears. The reader is halfway through a sentence, or reaching for a button, and the whole page jumps under their thumb. That is Cumulative Layout Shift, and it is one of the most irritating things a website can do to someone on a phone.

Interaction to Next Paint is the quiet one. All that player code has to be parsed and run, and while it runs, taps and clicks queue up unanswered. The customer who wants to ring you, or open your menu, waits a beat longer than they should. On a good day they do not notice. A poor connection is where they give up.

Four Ways to Put Video on a Business Website

There is no single right way to add video. There is a right way for your situation, and it depends on where the video lives and what it needs to do. This is the trade-off, ranked from heaviest to lightest.

Approach	Weight on load	Best for	Watch out for
Raw YouTube or Vimeo embed	Heaviest, around 540KB, blocks the main thread	The quickest copy-paste job	Loads fully even if nobody plays it
Facade or click-to-load	Almost nothing, around 3KB until clicked	Most business videos	Autoplay will not fire on load
Self-hosted video, done properly	Moderate, and fully under your control	Short background or hero clips	You must compress and configure it
Poster image, no video on the page	Lightest of all	When the video is optional	The video lives one click away

The Raw YouTube or Vimeo Embed

This is the copy-paste option: grab the embed code, drop it in, done. It works, and for a video buried far down a page that most visitors never scroll to, the convenience can be worth it. The problem is that the browser does not know the video is unimportant. It loads the entire player, all 540KB of it, as eagerly as if it were the main event. If you use nothing else from this article, at least add the lazy-loading attribute so an offscreen embed waits until the visitor scrolls near it. On WordPress that is often handled for you, but verify it rather than assume.

The Facade, or Click-to-Load

A facade is a static element that looks like the embedded player, a thumbnail with a play button, but carries almost none of the weight ^[2]. The real player loads only when someone actually clicks. For a YouTube video that is roughly 3KB instead of 540KB until the moment of interaction, and if the visitor never plays it, that code is never downloaded at all. Lightweight components such as lite-youtube and lite-vimeo do exactly this, and several WordPress performance plugins offer it as a checkbox. For most business videos, a promo film, a testimonial, an explainer, this is the correct default. The person who wants the video gets it on click. Everyone else, who came for your prices and your phone number, is not made to pay for it.

Self-Hosted Video, Done Properly

If the video is short and central, a looping hero clip behind your headline, self-hosting can beat an embed, because you are not loading a third party's entire player framework. Done properly means a few specific things, and web.dev's video performance guidance sets them out clearly ^[3]. Export both WebM and MP4 versions and list WebM first, so modern browsers take the smaller file. If it autoplays, it must be muted and carry the playsinline attribute, or mobile browsers will refuse it. Set preload to none or metadata so the full file does not download until it is needed, and give it a poster image so there is something to see immediately. Done carelessly, self-hosting is worse than an embed: a raw 40MB clip straight off the videographer's drive will punish every single visitor. Get it right, though, and it is often the fastest option of the lot.

The Poster Image That Is Not a Video at All

The lightest video on your homepage is the one that is not there. A high-quality still image with a play button, linking to the video on its own page or in a lightbox, gives you almost the entire visual benefit for a fraction of the cost. For a venue showing off a film, or a restaurant with a kitchen tour, the homepage job is to make someone want to watch. The watching itself can happen one click later, on a page built for it. Not every video needs to greet people at the front door.

How to Add a Video Without Slowing Your Page

If you already have a video dragging a page down, or you are about to add one, this is the order of operations. It works whether you self-host or embed.

Compress the source first. Before anything touches your site, get the file down to a web-appropriate size and format, WebM and MP4, not the enormous master file the videographer delivered.

Choose the lightest approach that does the job. A facade or a poster-and-link for most videos, self-hosting only when the clip is short and genuinely central to the page.

Set the loading rules. Add lazy-loading for embeds, and preload none with a poster for self-hosted clips, so nothing downloads before it is actually needed.

Verify on a staging environment. Test the change on a staging copy of the site, on a real phone, before it reaches your live pages. A staging environment exists precisely so you break the copy and not production.

Deploy and re-measure. Push it live, then run the page through Google's PageSpeed Insights and confirm your Core Web Vitals held up. Do not assume. Verify.

What Even the Best Setup Cannot Do

Two honest caveats, because the right expectations matter more than the tricks.

First, the facade trade-off. A facade cannot autoplay, because the real player does not exist until someone clicks it. If your design genuinely depends on a video playing the instant the page opens, a facade will not give you that. You trade the autoplay for the load time. For most business homepages that is the right trade, but it is a real one, so make it with your eyes open.

Second, and this is the one people forget: none of this removes the job of actually serving the bytes quickly. A perfectly optimised, lazy-loaded, compressed video still has to travel from a server to your visitor's phone. If that server is slow, or sitting in another country, or crawling because another site on the same shared box is having a busy morning, your clever video setup is undermined before it starts. This is where the hosting underneath earns its keep. A managed stack that serves your files from enterprise-grade Irish infrastructure with Nginx, Redis and FastCGI page caching means the optimised video you worked on actually arrives fast, close to the people looking at it. Optimisation on top of slow hosting is a fast car with the handbrake on. You want both, which is the whole point of getting the rest of your performance setup right rather than fixing one thing in isolation.

And the concession, because it is a real one. If your business is video, an online course platform pushing hours of HD, or a production studio whose showreel is the product, a dedicated video host built for streaming genuinely earns its place, and trying to self-host all of that on a general website would be a mistake. For everyone embedding the occasional promo clip or testimonial, that is machinery you do not need. Match the tool to the job.

Conclusion

Video is worth having. A film of your venue, a tour of your kitchen, a testimonial from a happy customer, these are among the most persuasive things you can put in front of someone deciding whether to trust you. The mistake is not adding video. It is adding video the heavy way and never noticing the visitors it costs you.

So the next time you or your web person drops a video onto a page, ask one question before it goes live: does this load only when someone wants it, or does every visitor pay for it whether they watch or not? Get that single decision right and you keep the video, the speed, and the customer who was only ever one slow homepage away from leaving. The choice of how to add it is yours, and now it is an informed one.

Frequently Asked Questions

Does embedding a YouTube video slow down my website?

Yes, if you use the standard embed code. Google's Lighthouse documentation puts a full YouTube embedded player at around 540KB, and it loads whether or not a visitor ever presses play. A lighter approach, such as a facade that only loads the real player on click, can cut that to a few kilobytes until someone actually interacts with the video.

What is a video facade or click-to-load?

A facade is a static placeholder, usually a thumbnail image with a play button, that looks like an embedded video player but carries almost none of the weight. The real player only loads when a visitor clicks. For most business videos this is the best default, because visitors who never play the video never download its code.

Is it better to self-host video or use YouTube?

It depends on the clip. For a short, central video such as a looping hero clip, self-hosting a properly compressed WebM and MP4 file can beat an embed, because you avoid loading a third party's entire player framework. For longer videos, or if your business is built around hours of video, a dedicated video host is the better tool.

Does video affect my Google Core Web Vitals score?

Yes. A heavy or slow-painting video can push your Largest Contentful Paint past the 2.5 second target, a video that loads without declared dimensions causes Cumulative Layout Shift, and a bulky player delays interactivity. All three are Core Web Vitals that Google measures from real visitors.

Should I autoplay video on my homepage?

Only with care. If you autoplay, the video must be muted and carry the playsinline attribute or mobile browsers will block it, and a heavy autoplaying video can badly hurt load time. For many business homepages a still image linking to the video is a better trade, giving you the visual impact without the weight.

How do I check if a video is slowing my page down?

Run the page through Google's free PageSpeed Insights tool, which reports your Core Web Vitals and flags heavy third-party embeds. Test on a phone as well as a desktop, ideally on a slower mobile connection, because that is where a heavy video hurts most.

Sources