E-commerce Product Video: What to Make and Where It Runs

Most online stores that "need a product video" are picturing one polished clip that will sit on the product page and also run as an ad and also go on Instagram. It never quite works in all three, and the usual conclusion is that video is expensive and does not do much. An e-commerce product video is not one asset. It is a small set of cuts, each built for one place and one kind of viewer, and most of them can come out of the same footage.

I spend my days on the timeline, so this guide is written from there: which cuts an online store actually needs, where the footage can come from if you have never done a shoot, how UGC-style and AI-assisted ads fit in, and what we can and cannot make for you. That last part matters, because some e-commerce briefs need a studio with a product photographer and a lighting rig, and we are not that. The e-commerce industry page is upfront about it too.

The four places product video earns its keep

The same product needs to be shown to four quite different people, and the mistake is cutting one video for all of them.

On the product page, the viewer already clicked. They are interested, they are comparing, and they want to know what the thing is really like: size, texture, how it opens, what is in the box. This video can be calmer and a little longer, and it can assume the viewer has read the product title.

In an organic social feed, nobody asked to see you. The video appears between other posts, muted, and has about a second to justify itself. It needs to be short, vertical and interesting from the first frame.

In paid social, the constraints are the same as organic but there is a second job: being testable. One ad teaches you whether that ad worked. Several genuinely different ads teach you which message works.

After purchase, the viewer owns the product and wants to use it properly. A short how-to-use clip in the confirmation email or on the product page's FAQ saves support replies and, for products with a learning curve, heads off returns caused by confusion.

Where the footage comes from

This is the question that stops most small stores, so it is worth being concrete. There are five realistic sources, and most good kits use two or three of them.

Footage you already have. Phone clips from the founder, an unboxing somebody filmed for a launch, old ads, customer videos you have permission to use. It is usually messier than you think and more usable than you fear. A proper edit with a grade, sound design, captions and motion graphics on top turns a folder of loose clips into something that looks deliberate. This is the most underused source in e-commerce.

Product photography. Good stills can be animated: camera moves, parallax, type and callouts built around them. It will never look like live footage, but for a feed cut or a product page loop it is often enough, and it is fast.

2D and 3D product animation. Where footage does not exist, or the product has an inside worth showing, a modelled and lit product render gives you total control of every angle. It costs more and suits products where precision matters, and it is a standard part of our catalogue.

AI UGC. Creator-style videos where the person, setting and voice are generated rather than filmed, and your product is built from clean reference plates so the label, shape and colour stay correct. It is fast and cheap per variation, which is exactly what ad testing needs. More on that below.

A live-action shoot. Models, locations, a studio product setup. This is the right answer for some brands, and it is the one source we do not supply ourselves.

UGC-style and AI-assisted ads

UGC-style ads work in feeds because they look like the rest of the feed. The bottleneck is people: finding creators, shipping product, waiting for footage, and paying again when a hook misses. Each round tests one or two ideas.

AI UGC videos change the speed of that loop. We design a small cast of AI creators for your niche, lock each one with a character sheet so they stay the same person in every clip, write hooks from your reviews and your category's winning angles, and edit each video like a native post. Takes where the product label warps or the shape drifts are regenerated, and when a hand interaction is too complex to generate reliably, a real product shot is composited in.

Two honest caveats. First, nobody can promise AI UGC will outperform real creators; what it changes reliably is how many angles you can test for the same time and budget. Many brands use it to find winning angles and then scale those with real creators. Second, our AI creators never pose as real customers describing personal results. The format works without that, and platforms increasingly expect realistic AI content to be disclosed.

If you are running paid social, the bigger idea is the batch, not the format. Our AI video ads service builds a set of distinct angles, each cut hook-first and exported for every placement, with filenames that encode the angle and hook so results read back cleanly. We wrote more about why winners wear out in our post on ad creative fatigue.

A scenario: a small skincare store with thirty products

Picture a Shopify store selling skincare. Two founders, thirty products, decent product photography, and a folder of phone clips of the founders applying things in their bathroom. No budget for a production company. Traffic comes mostly from Instagram and a small Meta ad spend.

Here is how I would plan it, from the edit outward.

The hero products first. Of thirty products, maybe four drive most of the revenue. Those get a proper product page video built from the phone footage and photography. The rest can wait.

Reels from what already exists. The same footage and photography become a batch of short vertical cuts for organic posting: a texture shot, a before-and-after of the routine, a "what's in the box". Cheap, because the raw material is already paid for.

A testable ad batch. For paid social, a set of AI UGC variations on the two best-selling products, each arguing something different: the problem, the result, the objection, the routine. The account tells you which angle holds attention, and the next batch doubles down.

One how-to clip. For the product with the most "how do I use this" questions in the inbox. Linked from the order confirmation email.

Nothing in that plan needs a studio. Almost all of it is editing.

Selling to shoppers is not selling to merchants

One more distinction, because it shapes everything. A shopper-facing product wants the product in use: hands, texture, scale, result. A merchant-facing tool, something sold to store owners, wants a business problem named in their language and shown as a number or a queue.

Our published e-commerce case study, CartResQ, is the second kind. It is an automated cart-recovery tool, and the explainer opens on the moment a sale is lost, a cart abandoned mid-checkout, before the recovery sequence kicks in. If you sell to merchants, that problem-first ordering is the lesson. If you sell to shoppers, lead with the product doing its thing.

Notes from the timeline

A few things I check on every e-commerce edit, whichever source the footage came from:

• Silent-first. Captions and on-screen type carry the argument. Music and voice are for the minority who unmute.
• Reframe, never centre-crop. A 16:9 master cropped to 9:16 loses its composition and often its captions. Each ratio gets re-blocked with type inside that platform's safe area.
• Show scale. A product on a white background could be any size. A hand or a known object tells the viewer what they are buying.
• Accuracy over polish. Colour graded to look good is not colour graded to look right. The shade on screen must match the box.
• Variations share one production. Ads that differ only in the hook should match in quality, or your test measures the edit rather than the message.

Who this is for, and who it isn't for

This approach suits online stores and DTC brands with some raw material (photography, phone footage, past ads) that need a steady supply of cuts, brands whose ad accounts burn through creative faster than they can book creators, and merchant-facing tools that need a problem-first explainer.

It is not for you if the brief is a styled studio shoot with models and a product photographer. Hire a production company for that, and send us the rushes if you want them edited and versioned afterwards.

What it costs with us

Our prices are published. For e-commerce, the usual starting points are:

• Reel Only, $100. Ten promotional reels under 20 seconds each, 5-day turnaround, one revision round per reel. For stores that already have core assets and need posting volume. See the Reel Only package.
• Solo Launch, $300. A product intro, a walkthrough, a tutorial and five reels, 5-day turnaround for the whole kit.
• Startup Boost, $550. Everything in Solo Launch plus a feature explainer, eight reels instead of five, and a ten-asset social kit, on a 7-day turnaround. Built for a first round of paid tests.
• Growth Bundle, $950. Adds five modular paid-social variations, a 2D/3D product intro and more, 10-day turnaround.
• AI UGC and AI video ads are quoted per batch against an approved list of hooks and creators. The first batch includes designing your creator cast; later batches reuse it.
• The monthly retainer, $2,450/mo, for always-on ad accounts: one request at a time from your queue, averaging about 48 hours each.

If you have a large pile of raw footage, full video editing and post-production on one footage set of up to ten finished minutes is included in the Scale Package at $1,600.

Frequently asked questions

How long should an e-commerce product video be?

For feeds and ads, 6 to 20 seconds, vertical, with the hook in the first second. For the product page gallery, 15 to 45 seconds is usually enough to show the product in use and at scale. How-to-use clips can run longer because the viewer wants the answer.

Do product videos increase conversions?

They can, but anyone quoting you a fixed uplift is guessing, because it depends on the product, the price and the page. What video reliably does is answer questions photos cannot, such as size, texture and how the product works, which is where hesitation usually comes from.

Can you make product videos without us shipping the product?

Often, yes. We can edit footage and photography you already have, build 2D/3D product animation, or produce AI UGC with your product built from clean reference plates. We still need good reference images: front, side and label detail.

Do you shoot physical products?

No. Studio product photography and live-action shoots are not what our team has shipped, and we will say so rather than take the brief. We edit footage others shoot, animate products in 2D and 3D, and produce AI UGC and AI ads.

How many product videos does an online store need?

Fewer than you think for the product pages, more than you think for social. Start with proper videos for the handful of products that drive most revenue, then a steady supply of short cuts for feeds and a batch of genuinely different ads to test.

The takeaway

An online store does not need one perfect product video. It needs a product page cut for interested shoppers, short vertical cuts for feeds, a testable batch for ads, and a how-to clip for new owners, and most of that can come from footage and photography you already have. Put the effort into the products that earn the most, and let the ads be many and different rather than few and precious.

The e-commerce video and motion page sets out what we make for online retail and where our experience starts and stops. For short-form volume, our product promo video service explains how one production becomes every ratio and length, and our guide to AI UGC videos for brands goes further into creator-style ads. When you know which products matter most, send us what footage you have and we will tell you what it can become.