The 'video everywhere' mistake
Somewhere around 2023 every ecom deck started carrying the same slide: product pages with video convert X% better. The number varied, the conclusion didn't. Put video on everything. So brands did. A looping clip on the hero, a UGC carousel halfway down, a founder talking-head near the reviews, on every product from the £120 hero SKU to the £6 replacement filter.
Here's the thing nobody says out loud: a lot of those pages didn't need it, and a few got worse. Video isn't free. Every one you add costs production time, slows the page unless it's hosted properly, pushes the add-to-cart button further down on mobile, and gives the shopper one more thing to watch instead of buy. On a page where the person already knows exactly what they want, that's not persuasion. That's a speed bump.
The 'video converts better' stat is real, but it's an average across pages where video answered a question the shopper actually had. The lift lives on those pages. The pages where there was no question to answer dragged the average down and nobody noticed, because 'we added video and conversion went up overall' is a nice thing to report.
So the useful question isn't 'should we have video on product pages'. It's 'which product pages have a question that a photo and three bullet points can't answer?' Those pages need video. The rest need faster load times and a bigger button.
This piece is a way of telling them apart, what kind of video each one wants, how to check it's actually working, and how to do the whole thing with Videowise so it isn't a six-month project.
The test: does the shopper have a question a photo can't answer?
Five questions. If a product page gets a clear yes on two or more, it needs video. One yes, probably worth testing. None, leave it alone and spend the money elsewhere.
- Does it move, or do something? A blender, a stroller folding, a jacket's hood packing into the collar, a serum absorbing. If the product's whole pitch is what happens when you use it, a still photo is a description of a verb.
- Is scale or fit the worry? Bags, furniture, anything worn, anything that goes in a room. 'How big is it really?' and 'what does it look like on someone my shape?' are the two questions that kill conversion on these pages, and they're nearly impossible to answer with product-on-white.
- Is it new, unfamiliar or hard to explain? A product people haven't seen before, or one with a mechanism, routine or setup. If your copy has a 'how it works' section longer than three lines, that's a video.
- Is the price high enough that people hesitate? Above roughly £80 to £100, shoppers want more reassurance before they commit. Seeing a real person use the thing, or the founder explain why it costs that, does what a fourth paragraph of copy won't.
- Is trust the blocker? Supplements, skincare, anything with a claim, anything where 'does it actually work?' is the real question. Customer video answers that in a way reviews and badges don't.
Run that over your catalogue and you'll get something like this:
| Page type | Moves or does something | Scale or fit | Unfamiliar | Price hesitation | Trust | Verdict |
|---|---|---|---|---|---|---|
| Hero product, considered purchase (£100+ appliance, bag, device) | Yes | Yes | Often | Yes | Yes | Needs video, probably two kinds |
| Apparel and footwear | Sometimes | Yes | No | Sometimes | No | Needs video: on-body and movement |
| Furniture, homeware with a footprint | No | Yes | Sometimes | Yes | No | Needs video: scale in a real room |
| Skincare, supplements, anything with a claim | Sometimes | No | Sometimes | Sometimes | Yes | Needs video: real customers, routine |
| New or unusual product, any price | Often | Sometimes | Yes | Sometimes | Yes | Needs video: the explainer |
| Consumables, refills, replenishment | No | No | No | No | No | No video. Make it fast and easy to reorder |
| Accessories, add-ons, low-price basics | Rarely | Rarely | No | No | No | No video, unless it's the demo of the main product |
| Bundles and gift sets | No | Sometimes | No | Sometimes | No | Usually no; a good unboxing photo set does it |
| Variants of a page that already has video | Same as parent | Same | Same | Same | Same | Reuse the parent's video, don't make more |
The honest result for most brands is that somewhere between a fifth and a third of product pages need video, and they're the pages that drive most of the revenue anyway. That's a very different project from 'video everywhere'.
What kind of video, and where on the page
Once a page passes the test, match the video to the question. One video answering the right question beats four answering nothing in particular.
'What does it do?' gets a demo. Fifteen to thirty seconds, product in use, no music needed, no intro. Hands, product, result. This goes in the image gallery as the second or third slide, where someone swiping through photos finds it without having to scroll.
'How big is it, what does it look like on me?' gets a scale or on-body video. For apparel, someone walking, turning, sitting, with their height and size stated on screen. For furniture and bags, the product next to a person or in a real room. Also in the gallery, because it's doing the job a photo would do if a photo could.
'How does it work?' gets an explainer. Thirty to sixty seconds, a person talking or a clear voiceover, the mechanism, the routine, the setup. This sits below the fold, under the description, where the person who's scrolled that far is the person who needs it.
'Does it actually work?' gets customers. Short, unpolished clips from real buyers: unboxing, first use, a week later. Three to six of them in a row, each under twenty seconds, as a shoppable carousel near the reviews. The lack of polish is the point; it's what makes them believable.
'Why does it cost this much?' gets the founder. Or whoever made it. A minute of them explaining the materials, the choice, the thing they refused to compromise on. Below the fold, near the price justification, for the person who's still hesitating.
Two rules that apply everywhere. Silent by default, with captions, because most of this is watched on a phone in public with the sound off. And every video that's near a product should be shoppable: tap the clip, add the thing to cart, without leaving the page. If the shopper has to go hunting for the button after watching, you've just spent production money on a distraction.
And one thing not to do: don't put a video in the hero slot that doesn't answer a question. A moody brand film on the first slide of a product gallery is the single most common way to make a page slower and less clear at the same time.
How to know it's working
'Conversion went up' isn't enough, because conversion goes up and down for twenty reasons a week. You want to know whether the video did it, on that page, for the people who watched it.
Four numbers, in order of how much they tell you:
- Watchers versus non-watchers, same page. This is the one that matters. Of the people who landed on the page, what share watched the video, and did the watchers convert at a higher rate than the people who didn't? If watchers convert at 2x and the watch rate is decent, the video's working. If watchers convert the same as everyone else, it's decoration. The better tools report this out of the box.
- Watch rate and completion. What share of visitors pressed play (or saw it autoplay) and how far did they get. A demo that most people drop at five seconds is answering the wrong question or answering it too slowly. Cut the intro.
- Add-to-cart from the video. If the video's shoppable, how many people added from it. This is the cleanest signal that a specific clip is doing the selling.
- Page speed before and after. Largest Contentful Paint on mobile. If adding video pushed it past about 2.5 seconds, the video is costing you conversions somewhere else on the page, and it needs to be hosted and lazy-loaded properly before you judge it.
Before rolling video across a whole category, test one page. Pick a product with enough traffic to get an answer in two weeks, add the one video the test says it needs, and run it as an A/B if you've got a testing tool, or as a before-and-after if you haven't. Watch the four numbers. Then do the next page.
And check the other direction too. Take one page that's got video but failed the test, pull the video, and see what happens. A surprising number of brands find the page gets faster, the button gets higher, and nothing else changes. That's a production budget you just got back.
Doing it with Videowise
The reason this used to be a six-month project is that every step was manual: shoot, edit, upload to a host, embed, tag products, check it didn't wreck page speed, repeat for every page, then try to work out in GA whether any of it mattered. Videowise collapses most of that, and more usefully, it gives you the watcher-versus-non-watcher numbers that tell you which pages actually needed the video in the first place.
What it is: a shoppable video platform for Shopify (and Shopify Plus, headless, Adobe Commerce, Magento and Salesforce) that lets you drop video onto product pages, collections, the homepage, landing pages, pop-ups and quizzes, with the product browsable, the variant selectable and the add-to-cart happening inside the player without leaving the page. It's one of the more established names in the category, with a 4.8 rating from a couple of hundred Shopify reviews and brands like Skullcandy on it.
The bits that matter for the approach in this piece:
- Finding out which pages need video, fast. It imports what you've already got from Instagram, TikTok, Drive, Dropbox and your Meta ads into one media library, so you can test a page with existing content before you shoot anything. Its AI Clips tool cuts long-form content (a ten-minute review, a live replay) into short shoppable moments, which is how you get a demo for a page without a production day.
- Bulk publish and the right widget for the job. Sixteen-odd widget types (carousels, stories, galleries, sliders, backgrounds) and bulk publishing across pages, so once you know a category needs on-body video, you can roll it out across the whole category in one go rather than page by page. And pull it off the pages that failed the test just as quickly.
- The numbers from the measurement section. Analytics on engagement, clicks, conversions, revenue and product performance, so you get watch rate, completion and add-to-cart from the video per page. Built-in A/B testing for placements, formats and individual videos, plus AI Experiments and AI Playlists that rotate videos on a page and work out which one's winning. This is the part that turns 'should this page have video' from an opinion into a two-week test.
- Speed. Built around lazy loading and Core Web Vitals, which is the whole reason video stopped being a page-speed tax. Check LCP on mobile anyway; it's your page.
- Beyond the product page. Shoppable images with product hotspots, the same videos pushed into Klaviyo and Postscript flows, the Shop app, TikTok Shop and Instagram, and a Meta comment-to-buy feature for Facebook and Instagram posts. Useful once you've worked out which videos sell, because then you want them in more places.
- Sourcing. An influencer network and a UGC hub for getting customer and creator video in, which is the input for the 'does it actually work?' pages.
Two things to know going in. Pricing scales with widget impressions, so a high-traffic store with video on every page pays for every page, which is one more reason to be choosy about where it goes. And live shopping is in there too; it's a different job from what this piece is about, so don't let it set the agenda for the pilot.
The practical workflow: run the test over your catalogue, pick the pages that pass, import or clip the video each one needs, publish it to just those pages, let the experiments run for a fortnight, and read the watcher numbers. Then, and only then, decide whether the category gets rolled out.
The first 30 days
You don't need a video strategy. You need a shortlist and a fortnight of data.
- Week one: run the test. Go through your top 50 products by revenue with the five questions. Mark each one: needs video, worth testing, leave alone. Be honest about the 'leave alone' pile; it'll be bigger than you'd like.
- Week one, same time: audit what's already there. List every product page that currently has video. Check it against the test. Note any page where the video's in the hero slot and doesn't answer a question. Check mobile LCP on your five busiest pages.
- Week two: pick three pages and one tool. One hero product that clearly needs a demo, one apparel or furniture page that needs scale or fit, one claim-led product that needs customers. Set up Videowise, import what you've got, clip or shoot the one video each page needs, publish to those three pages only. Pull video from one page that failed the test while you're at it.
- Weeks three and four: let it run and read it. Watch rate, completion, watchers versus non-watchers, add-to-cart from the video, LCP. At the end, you'll know which of the three worked and why. Roll out to the matching category for the ones that did. Rethink the video for the ones that didn't before you make more of it.
Things to resist. Resist shooting before you've tested with existing clips; half the time what you've already got answers the question well enough to prove the page needs it. Resist the brand film in the gallery. Resist 'while we're at it' additions to pages that didn't pass. And resist reading an overall conversion lift as proof; it's the per-page watcher numbers that tell you where the money is.
The pages that need video are the ones where the shopper's stuck on a question. Find those, answer the question, measure it, and leave everything else fast and simple. That's the whole strategy.
A useful tool starts with a clear brief. Explore the profiles below to find the questions worth asking.