Does a Shoppable Video App Slow Down Shopify? What Changed in 2026
Video is the heaviest thing most Shopify stores ship, but since July 2026 Shopify lazy-loads it automatically and the measured effect was large. Here is what the platform now does for you, what it does not, and the five checks to run on a video app before you install it.
On this page
Key takeaways
- Yes, video can slow a Shopify store, but it shows up in Speed Index rather than LCP. Shopify's own data shows median LCP barely moves between a page with 500KB of video and one with 50MB.
- Since July 2026 the video_tag Liquid filter sets loading=lazy automatically for videos in sections at position 3 or later that are not the header. Shopify measured the share of page views that downloaded a video file falling from 18.5% to 4.6%.
- Never lazy load a hero video above the fold. Pass loading: 'eager' explicitly, because an explicit value always overrides the automatic rule.
- The automatic rule only covers video rendered through video_tag. A third-party widget that injects its own <video> or its own player does not get it for free, so ask the vendor.
- Judge an app on whether it loads its script on pages where no widget renders. That is the cost you pay on every page, and it is the one you can test in two minutes.
A shoppable video app can slow your store down, but probably not in the way you have been told, and considerably less than it did a year ago. Video weight hurts Speed Index rather than Largest Contentful Paint, and since July 2026 Shopify has been lazy-loading below-the-fold video automatically, which removed most of the up-front cost without anyone touching their theme.
That leaves a narrower, more useful question than "is video slow". The question is what the platform now handles for you, what it still does not, and which parts of a third-party widget sit outside the automatic behaviour entirely. This post answers all three, with the numbers Shopify published rather than the ones app vendors quote.
The real number: 12.5MB of video is about 75% of a median mobile page
Shopify's performance team ran the HTTP Archive April 2026 crawl, roughly 670,000 mobile pages, and found that about 20% of Shopify storefronts use a <video> element on their homepage. When a video is present it dominates everything else: the median store with homepage video serves 12.5MB of video, which is about 75% of its total page weight on mobile. More than 80,000 mobile pages ship over 5MB of video.
That is the scale of the thing. Your fonts, your CSS, your app scripts and your product images added together are typically the minority of the bytes on a page that has a video on it. Which is why the single most effective performance lever on a video-heavy store is almost never minifying something.
Where these numbers come from, checked 28 October 2026
Every figure in this post comes from Shopify's own performance blog post on video_tag lazy loading, published 13 August 2026 by Dan Gayle, and from the video_tag filter reference on shopify.dev. The field data is Shopify's real-user monitoring; the page-weight data is the HTTP Archive April 2026 mobile crawl. We have not used any app vendor's published performance claims here, because in this category they are almost all self-measured and none of them are independently reproducible.
Speed Index, not LCP: where video weight actually shows up
This is the part that surprises people, and it changes what you should be optimising. Shopify's crawl data groups mobile pages by how much video they ship and compares median Speed Index. The pattern is clean and steep.
Median mobile Speed Index by video weight (HTTP Archive, April 2026)
| Video on page | Mobile pages | Median Speed Index |
|---|---|---|
| No video | 565,145 | 7,243ms |
| Under 500KB | 1,125 | 10,954ms |
| 500KB - 2MB | 5,387 | 10,962ms |
| 2MB - 5MB | 17,007 | 12,030ms |
| Over 5MB | 81,270 | 14,582ms |
A page shipping more than 5MB of video has a median Speed Index of 14.6 seconds against 7.2 seconds for a page with no video at all. That is roughly double.
Now notice what is missing from that table. LCP is absent, and deliberately so: across every video-weight bucket, Shopify found median LCP stays roughly flat at about 4.2 to 5.1 seconds. A page with 50MB of video has about the same LCP as a page with 500KB.
There are two reasons, and understanding them tells you what to fix. First, on most video pages the video is not the LCP element at all. The heavy video is below the fold and the LCP element is a hero image or a heading, whose paint time barely moves regardless of how many megabytes of off-screen video the page is also pulling. Second, even when the video is the LCP element, LCP records when the largest element paints its first frame, not when the file finishes downloading. A video paints as soon as enough data arrives to decode one frame, so a 5MB clip and a 50MB clip reach that point at roughly the same moment.
Speed Index is different because it measures how quickly the page visually fills in. A multi-megabyte video download hogs the network and starves the above-the-fold images, fonts and CSS the shopper is actually waiting on. The page looks unfinished for longer, and Speed Index climbs. So the damage video does is real, it is just not the metric most merchants are watching. If your store's symptom is shoppers leaving before anything renders, our guide to reducing Shopify bounce rate with video covers the other half of that equation.
What video_tag lazy loading does automatically since July 2026
Chrome added native support for loading="lazy" on <video> elements in Chrome 148, with an important bug fix in Chrome 150. Shopify shipped code to take advantage of it in July 2026. If you render video through the video_tag Liquid filter and do not set loading yourself, Shopify now sets it to lazy automatically for videos in sections further down the page, specifically sections at position 3 or later that are not the header. No theme change is required. It mirrors the behaviour image_tag has had for years and reuses the same section-position logic.
The measured effect, from Shopify's real-user monitoring across pages that contain a video file:
Shopify RUM, after the July 2026 rollout
Read that first line carefully, because it is the one that explains the mechanism. Before the change, 18.5% of page views on pages containing a video actually pulled the video file down. After it, 4.6% did. The difference is videos that shoppers never scrolled to, which now never download at all.
There is a second, quieter benefit worth knowing about. Lazy-loaded videos no longer hold up the window load event, per the HTML spec. If you have scripts or behaviour gated on load, they now fire earlier on video pages.
When to override with loading: eager
The automatic rule only applies when you do not specify loading yourself. Any explicit value passes through verbatim and takes precedence, which is exactly what you want for the one case where lazy loading is actively harmful: a hero video above the fold.
If a video is visible when the page loads, and especially if it is your LCP element, lazy loading delays the start of its download until after the page has already rendered. That makes the page feel slower, which is the opposite of the goal. Shopify's own guidance is blunt about it: never lazy load a video above the fold.
Forcing eager for an above-the-fold hero video
{% comment %} Force eager for an above-the-fold hero video {% endcomment %}
{{ section.settings.video | video_tag: loading: 'eager', autoplay: true, muted: true }}
{% comment %} Or drive it from section position yourself {% endcomment %}
{%- liquid
if section.index > 2
assign loading = "lazy"
else
assign loading = "eager"
endif
-%}
{{ section.settings.video | video_tag: loading: loading }}That second pattern matters if your theme's layout does not match the default position rule. A one-section landing page with the video in section four is still above the fold on a tall desktop viewport, and the automatic rule will lazy load it. Drive loading yourself using section.index and section.location when that is the case.
One more thing that bites: a hard-coded loading="eager" left in your theme from an earlier optimisation attempt will override the new automatic behaviour everywhere it appears. If you added one in 2024, go and find it.
Self-hosted MP4 vs YouTube embed vs app-delivered CDN video
These three behave very differently and the automatic lazy loading only helps one of them.
Shopify-hosted video through video_tag
This is the one that gets the automatic treatment. The file sits on Shopify's CDN, the filter emits a <video> element with a poster image and preload="metadata", and position-based lazy loading applies. It is the cheapest option in bytes and the least flexible in presentation. Note that Shopify's upload ceilings apply; the current Shopify video size limits are worth checking before you export at a high bitrate.
A YouTube or Vimeo embed
An iframe embed pulls in a third-party player, its own JavaScript and usually a set of tracking requests, and none of it is governed by Shopify's video_tag behaviour. The player script is frequently heavier than the video poster. It is the worst option for a product page and a defensible one for a long-form page where the shopper came specifically to watch.
App-delivered video
A shoppable video widget usually renders its own markup through its own script, which means the video_tag automatic rule does not apply to it. That is not automatically bad. A good widget implements intersection-observer lazy loading itself, serves adaptive bitrate so a phone on a weak connection gets a smaller file, and only loads on the pages where it renders. A bad one does none of those and ships a player bundle to your cart page. The difference is entirely the vendor, which is why the next section exists.
Auditing a video app before you install it: five checks
All five take under ten minutes on a trial install, and they are far more informative than the app's own performance page.
Five checks on any shoppable video app
- Open a page with no widget on it, open the network tab, and look for the app's script. If it loads there, you are paying for the app on every page of your store.
- Scroll a page that does have the widget and watch when the video bytes arrive. They should arrive as you approach the widget, not at page load.
- Check what is served to a throttled mobile connection. If you get the same file as desktop, there is no adaptive bitrate and your mobile shoppers are paying for your desktop quality.
- Look at the poster images. A feed of eight clips that loads eight full-size posters up front has moved the problem rather than solved it.
- Run the storefront before and after install on the same device and connection and compare Speed Index, not the Shopify admin speed score. The admin score is a lab number on one page.
Measuring it properly: 75th percentile field data, not a lab score
The Shopify admin's online store speed score is a lab measurement of a small sample of pages under simulated conditions. It is useful for catching a regression and close to useless for telling you what shoppers experience, because it does not know what devices or networks your shoppers actually use.
What you want is field data at the 75th percentile, which is the threshold Core Web Vitals is assessed against. If your store has enough traffic, the Chrome User Experience Report gives you that for free through PageSpeed Insights, and it reflects real sessions on real hardware. Look at it on a product page template, not the home page, because that is where most video widgets live and where most shoppers land from search.
Measure before you install, then again two weeks after, and compare the same page template on the same metric. Two weeks is the minimum for field data to catch up. Anything shorter and you are reading noise.
And separate two questions that get conflated constantly. "Did the page get slower" is a performance question answered by Speed Index at the 75th percentile. "Did the store make more money" is a commerce question answered by a holdout test. A widget that costs you 200ms and earns a meaningfully higher conversion rate on the pages it appears on is a good trade; one that costs you 200ms and earns nothing is not. If your video is not playing at all on some devices, that is a third question entirely, and autoplay failing on mobile and video not playing in Safari on iPhone are the two places to start.
Frequently asked questions
- Does Shopify lazy load videos automatically now?
- Partly, and the detail matters. Since July 2026 the video_tag Liquid filter sets loading=lazy automatically, but only for videos in sections at position 3 or later that are not the header, and only when you have not set loading yourself. Video rendered any other way, including by a third-party app's own script or a YouTube iframe, is not covered by this rule at all. Shopify measured the change in real-user monitoring and found the share of page views that downloaded a video file fell from 18.5% to 4.6%.
- Should I lazy load a hero video above the fold?
- No. Shopify's guidance is explicit that you should never lazy load a video above the fold. If the video is visible on load, lazy loading delays the start of its download until after the page renders, which makes the page feel slower rather than faster. The automatic position rule already avoids the header and the first two sections, but if your layout puts a visible video further down the section order, pass loading: 'eager' explicitly. An explicit value always overrides the automatic behaviour.
- Why does my Shopify speed score drop after installing a video app?
- Usually one of two causes. Either the app loads its JavaScript on every page of your store rather than only where a widget renders, which taxes pages that show no video at all, or the widget loads poster images and video metadata for every clip in the feed up front rather than as the shopper approaches. Both are testable in two minutes: open a page with no widget on it and look for the app's script in the network tab, then scroll a page that has the widget and watch when the video bytes actually arrive.
- Does video weight hurt LCP or Speed Index more?
- Speed Index, by a wide margin. In Shopify's analysis of the April 2026 HTTP Archive mobile crawl, median LCP stayed roughly flat at about 4.2 to 5.1 seconds across every video-weight bucket, while median Speed Index climbed from 7,243ms on pages with no video to 14,582ms on pages shipping over 5MB. LCP measures when the largest element paints its first frame, and a video paints as soon as one frame can be decoded, so extra megabytes never move it. Speed Index measures how fast the page visually fills in, and a large video download starves the resources that do that.
- How do I check if a video app is loading scripts on every page?
- Open a page on your store where the widget definitely does not appear, such as your cart or a policy page, open your browser's developer tools to the network tab, and reload with the cache disabled. Filter by the app vendor's domain. If their script appears there, it is loading everywhere. Some apps do this deliberately to detect where to render; a well-built one uses Shopify's app embed scoping or conditional loading so that the cost falls only on pages that need it. Ask the vendor directly, and treat a vague answer as an answer.
- Is a self-hosted MP4 faster than an app-delivered video widget?
- A Shopify-hosted MP4 rendered through video_tag is the lightest option, because it ships no extra JavaScript and it now gets position-based lazy loading for free. But it is only a fair comparison if you do not need what the widget does. A plain MP4 has no product tagging, no swipeable feed and no analytics, so the honest comparison is against a widget that lazy-loads properly and only loads on pages where it renders. A good widget costs you some JavaScript; a bad one costs you a player bundle on your cart page. The difference is the vendor, not the format.



