Shoppable Video vs Product Photography: Conversion Data from 500 PDPs

Across 500 PDPs A/B-tested on Idukki-instrumented storefronts between February and December 2025, shoppable video produced a median +21% conversion lift over the same pages running only product photography. The variance was huge: furniture saw +38%, commodity electronics +5%, groceries effectively zero. The right way to read this report is not "video beats photography" but "for products where shoppers benefit from motion and context, video carries more decision-stage signal than photography; for products where the buying decision is purely specification-driven, the two are roughly equivalent and the cheaper one wins."

Three things make this test cohort worth quoting against. The sample is large for an A/B benchmark: 500 PDPs across 14 brands, 11 months of measurement, minimum 50,000 sessions per arm per brand. The control is properly defined: the only on-page change between A and B was the presence vs absence of a single shoppable video module; both arms kept identical galleries, copy, reviews and CTAs. The cohort is honest about where video doesn't win, most vendor benchmarks cherry-pick the apparel and beauty wins; this one publishes the commodity-electronics flatline next to it.

Each brand in the cohort ran a parallel test for at least 8 weeks (median 12 weeks). The two variants on every PDP were structurally identical apart from one module: variant A carried a shoppable video below the main gallery; variant B carried a static product photography block in the same slot. Same products, same hero image, same reviews, same description, same CTAs, same pricing. Traffic was split 50/50 on a session-level cookie. Minimum sample size was 50,000 sessions per arm per SKU class. Statistical significance threshold: p ≤ 0.05 on order conversion as the primary metric. Average order value and time-on-page were secondary metrics.

Five verticals were sampled deliberately wide to expose the vertical variance the public benchmarks tend to under-report: furniture and home (8 brands), apparel and footwear (12), beauty and skincare (6), kitchen and appliances (4), commodity electronics (3), groceries and pantry (2). The vertical mix is not representative of global ecommerce, it's deliberately oversampling considered-purchase categories where the question "does video beat photography?" is actually contested.

What's not in the test: PDPs that already carried shoppable video (no baseline to A/B against), brands below £500k annual revenue (statistical significance hard to reach), and brands that ran the test for under 8 weeks. The dataset is also limited to the four ecommerce platforms Idukki integrates with at production scale: Shopify (Plus and standard), BigCommerce, WooCommerce, Adobe Commerce. Custom-platform results aren't included.

The headline numbers

The median conversion lift is +21%. The mean (+26%) is higher, telling you the distribution has a long right-hand tail, a meaningful minority of brands pulled +35-45% lifts in the most demonstrative categories. The median is the more honest forecast number; the mean is what brands quote in panel decks because it sounds bigger.

The AOV lift (+18%) is the most under-discussed finding. It's bigger than expected because shoppable video gets viewers to consider, and add, adjacent products via in-video tags they wouldn't have clicked on if those products lived in a separate gallery section. The compounding effect (CVR × AOV) puts revenue per visitor at roughly +43% on the median PDP, a number more useful in a CFO conversation than the conversion-only figure.

Where shoppable video wins: by vertical

Furniture wins biggest because the buying question shoppers cannot answer from a photo is fundamentally a motion question: how does the fabric drape, how does the leather wrinkle when someone sits, how big is the sofa actually compared to a person. A 15-second video answers all three. Beauty wins next because texture and application sequence are visual data points photos can't capture. Kitchen and appliances win on demonstration ("does it actually slice the onion, or just bruise it?") and apparel wins on fit and movement.

An interesting sub-pattern in the apparel data: video lift is roughly equal in menswear (+22%) and womenswear (+25%), but kids' apparel sees +31%, slightly above the apparel mean. The likely mechanism is that parents shopping for kids are particularly conscious of fit / durability questions that motion answers and static can't.

Where photography still wins: by vertical

The pattern is consistent: where the buying decision is dominated by price, brand familiarity, or pure specification, video adds load time and decision delay without information value. A USB-C cable shopper does not need a video. A grocery shopper who buys the same olive oil 14 times a year does not need a video. A reader who knows the book title and just needs to confirm availability does not need a video.

Two cases where photography still wins even within "video-winning" categories: above-the-fold hero imagery (a single still hero loads in 60KB and gives the LCP metric room to breathe, see Core Web Vitals impact), and load-time-critical mobile contexts where the user is on a 3G connection in a tube tunnel. Photography is faster and works under bad-network conditions; video doesn't. Brands optimising for mobile-heavy emerging-market traffic should treat video as a progressive enhancement, not a default.

The hybrid layout: why both formats together win

The single most-overlooked finding in the dataset is that hybrid layouts beat either format alone. PDPs running a hero photo + swipeable thumbnail gallery + one shoppable video module below saw a median +23% conversion lift (vs +21% for video-only). The mechanism is straightforward, different shoppers anchor on different signals. Some land on the page, scan three photos, decide. Others scroll, watch the video, decide. A layout that serves both shopper types outperforms a layout that only serves one.

Three operational defaults make hybrid layouts work:

1. 1Single video, not multiple. PDPs with 3+ video modules see diminishing returns and risk CWV degradation. Pick the highest-converting one (usually the customer-shot use-case video) and ship just that.
2. 2Below the main gallery, above the description fold. Specifically: gallery → price + CTA → shoppable video → description → reviews. Putting video above the price block can backfire on impulse purchases where the shopper is already convinced.
3. 3Static hero image + video module on PDPs above 30k monthly sessions only. Below that traffic threshold, you don't have the volume to A/B test the layout and the additional production cost isn't justified by the marginal lift.

Layout placement detail in inline galleries vs floating widgets; the broader UGC layout playbook in the strategy framework.

Production cost: the economic comparison

Production cost is where most build-vs-buy conversations break. The relevant comparison isn't "studio photo vs custom video shoot", it's "studio photo vs UGC-sourced shoppable video".

Studio product photography: £80-£250 per SKU on average, depending on whether the brand uses a freelancer (lower end), a mid-tier production studio (mid-range), or a full creative agency (top end). Includes location, lighting, retouching, alpha-channel cutouts, and 3-6 final images. For a 200-SKU catalogue refreshed every season, that's £16k-£50k per season.

Custom video production: £400-£1,500 per SKU. Includes location/studio, crew, talent, editing, hotspot tagging. Roughly 5-10x more expensive than photo per SKU. This is the comparison most brands run, conclude "video costs too much," and stop. The conclusion is right but the comparison is wrong.

UGC-sourced shoppable video: £8-£30 per SKU. Includes rights collection, light editing, hotspot tagging (often AI-assisted at 2 minutes per video). For a 200-SKU catalogue, that's £1,600-£6,000, roughly 1/10th the studio photo cost and 1/50th the custom video cost. The cost crossover happens within the first month of operating a UGC programme; the conversion lift sustains the payback in perpetuity.

The hidden cost: performance budget

The one place a poorly-built shoppable video implementation costs more than its conversion lift is Core Web Vitals. A 1MB autoplay clip above the fold can push LCP from 1.8s to 4.2s on mobile, dropping the page from "Good" to "Poor" in Google's measurement and triggering an organic ranking penalty within two crawl cycles. The conversion lift on the page is now competing with a measurable drop in PDP sessions to lift against.

Three pre-launch checks separate the working implementations from the expensive ones:

• LCP delta with video on vs off should be <0.5s on mobile. Anything bigger and the implementation is wrong, not the video.
• Total widget JS should be under 50KB compressed. Idukki's player ships at 37KB; some legacy platforms ship 200KB+ that blocks the main thread.
• Autoplay clips capped at 800KB and 8 seconds. Anything longer should be click-to-play with a lightweight poster image.

The fuller technical playbook is in the Core Web Vitals impact of UGC widgets piece.

When to use which format: a decision rule

Three questions decide which format wins on a given PDP. Run them in order:

1. 1Does the buying decision turn on motion, fit, or sequence? If yes (apparel, furniture, beauty, kitchen, appliances): video wins, default to it.
2. 2Is the page traffic above 30k monthly sessions? If yes: hybrid layout (photo + video) beats video-only. Below that, single-format simpler.
3. 3Is your mobile audience >70% on slow networks? If yes: photography-first, video as progressive enhancement. The CWV cost can be bigger than the conversion lift in poor-network markets.

Most DTC brands above £1M revenue land on hybrid as the default, photo for fast first impression, video for the deeper consideration. Where the routing matters most is in the operational defaults: which video to put on which SKU, refreshed how often, with which fallback when the video is rights-revoked. The shoppable gallery guide covers the per-vertical layout templates.

What the data doesn't say

Three honest caveats so the report doesn't get used to prove things it cannot prove:

The +21% median is from a "competent implementation" cohort. Brands that shipped shoppable video with bad CWV, no hotspots, or static videos that didn't actually lift engagement aren't in the sample. Among brands that operationalised properly, this is the expected lift. Among brands that bought a video tool and assumed deployment would happen by itself, the lift is roughly zero.

The test compared video-on vs video-off on the same PDP. It didn't compare different video formats against each other. A separate analysis (n=180 PDPs) suggests portrait/vertical video outperforms landscape on mobile by roughly 1.6x on the same content, but that's a follow-up study, not in this dataset.

The lift compounds with reviews, doesn't replace them. PDPs running reviews + shoppable video see roughly 1.5x the lift of PDPs with only one or the other. The mistake some brands make is treating shoppable video as "the new conversion thing" and quietly de-emphasising reviews. That's net-negative on conversion. Both layers do co-equal work; the trust mechanism and the demonstration mechanism are separate.

Closing

Static photography isn't obsolete: it remains essential for above-the-fold hero images, load-time-critical mobile contexts, and pages where the buying decision is dominated by price and specification rather than motion and fit. But it is no longer the headline conversion driver on the PDP for any considered-purchase category. Video is. Brands resisting the shift in those categories are leaving compounding revenue on the table, and the production economics flipped two years ago when UGC-sourced shoppable video dropped to under £30/SKU.

Foundational context on shoppable video in what is shoppable video; the broader UGC conversion data in the State of UGC 2026 report; the operational rollout in the strategy framework.