Why your category page looks inconsistent

The grid is the product, not the photograph

Product photography is nearly always reviewed one image at a time — does this shot look good, is the colour right, is the garment clean. That review cannot catch the problem this post is about, because the problem only exists in aggregate.

A shopper scanning a category page is not evaluating photographs. They are scanning for the item they want, and anything that breaks the rhythm of the grid costs them a beat. Twenty images that are individually excellent but mutually inconsistent produce a page that feels harder to shop than twenty mediocre images that match.

The five things that drift

Drift is not random. It happens along the same handful of axes every time, because those are the variables a shoot resets when it starts.

• Crop and framing — how much space around the garment, where the body is cut. The most visible, because it changes apparent size in a fixed tile.
• Model — a different person between shoots, or the same person a year apart.
• Lighting — warm versus cool, hard versus soft. Reads as a colour difference in the garment, which is worse than reading as a lighting difference.
• Background — a grey that is not the previous grey is more noticeable in a grid than either grey is alone.
• Scale — the garment occupying 60% of one tile and 85% of the next.

Why it accumulates in exactly the wrong place

New arrivals go up as they arrive. Best-sellers get reshot when there is budget. Returns-heavy items get new photos to fix a fit complaint. Each of those is a sensible decision made on its own, and together they guarantee that the most-viewed category pages are the ones assembled from the most separate shoots.

So the drift concentrates where the traffic is. A quiet corner of the catalogue shot once in one afternoon looks perfectly consistent; the page that earns the most is the one stitched together from two years of good intentions.

Fixing it without reshooting everything

The instinct is to reshoot the category, which is expensive and, done listing by listing over weeks, reproduces the drift it was meant to remove. The cheaper move is to treat the five axes separately, because they are not equally expensive to fix.

Crop, background and scale are post-production. They can be normalised across an existing set without re-photographing anything, and they are the three most visible in a grid. Model and lighting need the images regenerated — so do those only where the mismatch is severe.

• Standardise the crop first — cheapest fix, biggest visible gain in a tile grid
• Normalise the background — a background swap is a per-image edit, not a shoot
• Match scale — decide what fraction of the frame the garment fills, apply it to the set
• Regenerate for model and lighting — only where the mismatch actually reads
• Write the standard down — otherwise the next shoot re-introduces the drift

The standard is the durable part

Whatever you fix now will drift again unless the rules outlive the person who applied them. A one-page spec — crop ratio, garment fill fraction, background value, model treatment — is what makes the next shoot match this one.

It is also what makes generated imagery worth more than a shoot for this particular problem. A written standard is only as good as the compliance, and a render applies the same model, framing and lighting to the two-hundredth garment as it did to the first, which is the one thing a series of shoots cannot promise.

Where this stops applying

Consistency is a means, not a goal. Some catalogues are deliberately varied — vintage and one-off resale inventory cannot be made uniform and should not be, because the variation is the honest representation of the stock.

It is also a second-order problem. If a category page is not converting and the photos are inconsistent, the inconsistency is probably not the reason. Price, sizing information, stock availability and review count all move conversion more, and a tidy grid of the wrong products is still the wrong products.

• One-off and vintage inventory, where variation is the truth about the stock
• Editorial or campaign pages, where deliberate variety is the design
• Any page whose real problem is price, sizing or availability
• Very small catalogues, where the grid is too short for drift to register

Common questions

More from the blog

Related