Merchentia.

Why your complete-the-look widget is empty on some product pages

A troubleshooting guide for Shopify merchants · updated September 2026

The short answer, true far more often than the other four combined: a look needs a photo of a person wearing or carrying the items. If a product’s photos are packshots or flat lays, no outfit can be detected and nothing is built — by design, not by failure. Everything else is a second-order cause.

1. The photo has no person in it

When the app analyses a product image it is looking for a real model wearing or carrying the products. A flat lay, a packshot, a ghost-mannequin shot or a garment on a hanger returns no items at all.

That is a deliberate refusal. Handed a lone jacket on white, an image model will cheerfully “detect” the jacket’s own belt, zip pull and shadow as separate garments and assemble an outfit out of one product. Returning nothing is the honest answer.

Why shoes and bags go missing specifically: the frame has to contain the item. A waist-up hero shot has no feet in it, so the footwear row has nothing to detect however good the photo is.

Gallery order matters. Photos are read in order and merged until the outfit looks complete, so leading with a full-length shot of a person is both the best result and the cheapest one. If a model shot already exists further down the gallery, reordering is usually the entire fix — no reshoot.

2. The matched product isn’t buyable

A card can only show a product a shopper can actually buy: Active, in stock, and still published to the Online Store sales channel. A draft or unpublished product has no live product page at all, so recommending it would dead-end the shopper.

This one is usually invisible rather than empty — the row silently shows the next-best candidate instead. Our guide on stock-aware looks covers that substitution behaviour in full.

When a row genuinely can’t be filled, the reason falls into one of three buckets: there is nothing buyable in that category at all (only stock fixes it), nothing resembled the garment closely enough (loosen the Similar threshold or widen the match scope), or the detector couldn’t describe the garment well enough to match it (usually a photo problem again).

3. A category is switched off

The detected-category allow-list in Settings decides which kinds of item may appear on the storefront. Untick “Bags” and every bag row disappears instantly, with no rebuild.

Note this doesn’t necessarily blank the widget: if the allow-list empties an AI look, the app falls through to a hand-built look for that product if one exists, then to the similar-products fallback if that setting is on.

4. Your plan’s look capacity is used up

Only the most recently built looks within your plan’s capacity serve to the storefront. If you have generated more looks than your plan covers — or downgraded — the older ones stop serving without being deleted. They come back when capacity does.

5. Both widgets are on the same page

This one is deliberate and surprises people. If the similar-products fallback would render on a page that also has a “You may also like” block, the fallback is suppressed — because the two would compute and show the same items, and a shopper would see one row twice.

So “the widget shows nothing on products without a look” can simply mean you have both blocks placed and the fallback is correctly standing down.

Telling them apart

  1. Does the hero photo show a person? No → cause 1. This is where most investigations end.
  2. Is it empty on every product, or just some? Every product → look at placement, capacity, or the allow-list. Just some → almost certainly the photos.
  3. Is a specific row missing on otherwise-good looks? That’s the allow-list, framing (cause 1), or nothing buyable in that category.
  4. Did it work and then stop? Capacity or stock, not photography.

The two fixes worth doing this week

Reorder your galleries so a full-length model shot leads on your best-selling products. It costs nothing and fixes the most common cause outright.

Add one full-length frame to hero products next time you shoot. You don’t need a new campaign — one frame with the whole outfit in it is enough to make a product eligible.

Do it with Shop The Look

Merch - Shop The Look is deliberately honest about this: it builds a look only where the photography supports one, and shows you where it didn’t rather than inventing an outfit. The readiness card tells you up front how much of your catalogue is usable, the similar-products fallback covers pages where no look can be detected, and the “you may also like” widget is included on every plan for catalogues that are packshot-only. It installs as a no-code theme app block. Live on the Shopify App Store, with plans from a free tier up to Enterprise.

Install on the Shopify App Store →

About Shop The Look →

Frequently asked questions

Why is the widget empty on one product but fine on another?

Almost always the photo. A look is only built from an image showing a real person wearing or carrying the items, so a product whose hero shot is a packshot or flat lay produces nothing while its neighbour with a model shot works.

Does it work with flat-lay photography?

No. A flat lay, packshot or garment on a hanger is deliberately skipped and returns no items, so the app doesn’t invent an outfit from one product’s own straps and trim.

Why are shoes and bags missing from my looks?

Usually framing. If the hero shot is cropped at the waist, the feet aren’t in the picture, so the footwear row has nothing to detect. A full-length frame recovers it.

Does the widget always show something?

No. With the similar-products fallback enabled it usually does, but that fallback is deliberately suppressed when a “You may also like” block sits on the same page, because the two would show identical items.

Do I need to reshoot my catalogue?

Usually not. Photos are read in gallery order, so if a model shot already exists further down the gallery, reordering it to the front is often the whole fix.

More from Merchentia

See all apps →

Questions?

Email ask@merchentia.com — we read everything.