Having just spent weeks (or months) with a content agency to script, film and edit a professionally polished video showcasing your product or local service offering, the best place to get a solid return on investment is to place it front and centre, above the fold ¹, on your business website’s front page – after all, you just spent all this money, the world should see it. Right?

Not so fast, because the nature of website visitors and streaming video technology, it could make the experience – and subsequent revenues as a result – much worse.

There are several reasons why this could be the case, the primary being that an embedded video is probably the single most expensive thing you can put on a website.

Even if you don’t think about the money a video costs to produce, it takes up storage on your web server and bandwidth on every single visit, as well as increasing page load times for every single visitor.

And have you ever checked whether visitors are actually watching it?

The reality is, most website visitors are impatient. Unless you’re running a media company, they’re not at your site for fun, they’re not there to stick around for a while to be entertained, they’re there for a specific reason: can this product or service provider solve my problem?

That’s the question worth considering, and asking how your video fits into their dilemma.

I’m sure your video is excellent, but does it answer the fundamental question your prospects are asking themselves? If you run a video production company, then the video is the product – but for everyone else, it’s a big maybe.

The problem is that placing a video in the hero section is usually a decision made once on the basis of how it looks in a design mockup. Meanwhile it sits at the top of your most important page, quietly taxing every visitor who arrives.

Here’s what it’s costing you, and how to find out whether it’s worth the price.

The video loads before your visitor decides anything

Before someone can judge whether your business can solve their problem, they have to see your page. Adding a video on top changes how long that takes.

For context on how much room there is to play with, the median mobile home page is roughly 2.6 MB ². That’s for everything on the page – every image, script, font and stylesheet combined. In 2015 the median was around 845 KB, so the web has roughly tripled in weight in a decade.

For comparison, in the 2024 edition of the same report, video was by far the heaviest resource type on the web, reaching roughly 8.6 MB at the 90th percentile. In other words, a single video will sometimes weigh more than three times an entire median home page.

Now, it is possible to optimise your video – compression and modern codecs can help, here – but you’re still forcing every visitor to download a substantial file before your page settles.

What the delay does to your bounce rate

There is a lot of data out there on the correlation between page load times and bounce rates ³, for instance Google’s mobile speed research ⁴ found that more than half of mobile visitors abandon a page that takes longer than three seconds to load.

Three honest caveats about those figures:

• Most of this data is now nearly a decade old, and mobile processors, as well as mobile internet speed, are much faster nowadays
• The research was mostly based on mobile advertising landing pages, where the user’s intent is different from someone who searches for your business by name
• It’s correlational – while slow pages and high bounce rates tend to happen in tandem, there could be other reasons for rapid abandonment that have nothing to do with site speed
• Funding for this research is done by companies selling performance services and Google (which indirectly makes money by improving performance)

That being said, the research clearly establishes that the penalty for being slow isn’t linear, it compounds. A page that takes six seconds to load is significantly worse off than a page that takes three, which is mildly worse than it if loaded in one.

While every second counts, every subsequent second counts more than the last.

When a video is placed into the hero section of your website, it can push your website into slow territory, getting noticed by search engines that place you lower in search results.

How video load speed affects search engine rankings

For several years now, load times have been taken into consideration when ranking web pages.

This is done by a metric called “largest contentful paint”, or LCP – which is one of the three “core web vitals” that feeds Google’s page experience signals – measures how long it takes for the biggest visible element to render.

Google’s threshold for a “good” LCP is less than 2.5 seconds, based on tracking field data from real website visitors over time.

Since 2023, a video element can be an LCP candidate, with the timestamp taken from when the first frame is painted.

So if you have a large autoplay video at the top of your home page, it is almost certainly the largest element in the viewport, which means it is what Google is timing. And autoplay tells the browser to start pulling video data immediately, which competes for bandwidth with the styles and fonts and scripts that everything else on the page is waiting for.

Fortunately in this instance the fix is usually very simple, either you could:

1. Move the video down the page so it’s no longer visible during page load
   • This doesn’t stop the file from downloading but it does remove it from the LCP contention
2. Disable the video’s autoplay attribute or set preload="none"
   • This makes the browser wait until a user actually wants to watch the video

But before that’s done, it’s important to check whether the video is even the problem.

When the expensive part of a video isn’t the video

Sometimes the fix isn’t always what you expect.

On one site I was working on, the home page video sat at the bottom of the initial viewport with only a sliver of it shown – not that that matters because the autoplay attribute starts the download regardless of how much of the video is visible during page load.

After analysing the page, I could see the actual culprit was not the video but the poster (this is an image reference that loads a picture before the video starts) which was a ~550 KB JPEG that could easily be optimised.

Due to the nature of images, there are different file formats that are more appropriate than others depending on the contents, and in this instance it was the wrong one – not a photograph but a black and white graphic with lots of solid areas and hard edges. Converting this image to a PNG reduced the file size to ~28 KB with no loss in quality.

It was the exact same image, exact same dimensions, with no visible difference, but about 95% smaller.

After applying the new image, the page’s performance score moved from concerning to near-perfect.

This is something that automatic image optimisation plugins routinely get wrong because they often maintain the same format that was uploaded, rather than understanding and applying what’s more appropriate for each picture.

This is an example of when a performance problem is not a technical issue caused by a developer making the site slow, but a content problem caused by a knowledge gap in the designer who uploaded and applied an unoptimised image to the front page video.

But even once it was fixed, the problems didn’t immediately cease and we eventually decided to move the video further down the page.

The cost of a video that has nothing to do with site speed

So here we are with a beautifully compressed video served from a fast CDN sitting behind a lightweight poster image, on a page that loads in under two seconds.

The immediately-visible video still might be the wrong call because it’s asking for too much, too soon.

Going back to an earlier point, think about what someone actually wants when they land on your home page – once they’ve either clicked a search result or a link from a directory – they’re considering one question: do these people do the thing I need, near where I am, at a price I can afford? They want to answer that question within a few seconds, by scanning, because like us they’re busy people.

A video, especially when controls are disabled, forces you into its own timeline – it plays at a fixed pace, in a fixed order, and it can’t be skimmed.

Someone who wants to know whether you bulk-bill or whether you service the northern suburbs can find that in a paragraph in two seconds. But in a 90-second brand video, that information might arrive halfway through or not at all.

You’re asking a stranger for 90 seconds of undivided attention before they’ve decided you’re worth 5.

Wistia’s 2026 State of Video Report ⁵ shows that while website home pages, video galleries and product pages do get the highest play rates of any page type, viewers generally only get about halfway through. Videos under a minute long average roughly 52% engagement.

Considering that half of a video’s viewers are gone by the midpoint, and the people who never pressed play are a much larger group again, whatever matters most needs to be in the first few seconds – or better, in text next to the video where everyone can quickly see it.

Ultimately, your website is unique

Everything above is a hypothesis about websites in general, and it might be wrong for you. A well-made video on a fast page in front of the right audience can be the single best thing on your home page.

The problem is that almost everyone argues about it from the perspective of taste rather than evidence.

Your analytics dashboard doesn’t settle it either. Google Analytics (and the other website tracking tools) will show you your aggregated bounce and engagement rates, but it can’t tell you why anyone left – it doesn’t distinguish whether someone left after watching part way through and decided you weren’t for them, or scrolled straight past without even registering the video existed.

Those are opposite problems – one means the video is doing its job but the messaging is wrong, the other means the video is not doing its job – and should be treated as such.

To tell them apart you need to see what people actually do – how far down the page they get, whether the play button is ever touched, what they scroll past to reach. That’s what behavioural analytics tools do. But even they can only show you what happened and not why it happened – for that you’ll need to show it to people and gather their reactions ⁶.

This article is not an argument against website videos, it’s an argument against the default beliefs that don’t tend to be questioned.

But we can say with confidence that, outside of specific business types, a video isn’t usually what a website visitor wants – they want answers to their questions, and if it a video delays those answers, then it’s competing for their attention.

If you happen to have a video and you’re not sure whether it should go on your home page or not, feel free to contact us and we’ll provide a free website review.

1. On a website, content “above the fold” refers to anything immediately visible when the page loads – the term comes from newspapers, where they are stacked with a physical fold halfway, so a front page story could be positioned below this line and only seen by the people who pick it up ↩︎
2. HTTP Archive’s 2025 Almanac, which analysed over 17 million sites ↩︎
3. A “bounce” is when a website visitor leaves the page before doing anything ↩︎
4. The research was from 2016 but republished through Akamai as the State of Online Retail Performance Report 2017 https://business.google.com/ca-en/think/marketing-strategies/mobile-page-speed-new-industry-benchmarks/ ↩︎
5. https://wistia.com/blog/video-marketing-statistics#benchmarks-and-roi ↩︎
6. But even when showing design assets to others, it is surprisingly common for them to lie when they think a lot of work has been put into producing it because they don’t want to hurt anyone’s feelings: https://ixdf.org/literature/article/how-to-get-more-honest-feedback-in-user-testing#1_people_dont_want_to_hurt_your_feelings-0 ↩︎