How to audit your product image catalogue in 5 steps

Two notes before you start. There are no conversion percentages in this article, because the figures that circulate about product imagery rarely trace back to research we would put our name to. And this audit does not end at your own webshop. We have deliberately kept that part for last.

How a catalogue drifts

Imagery is almost never made in one go. There was a shoot at launch. Then a colour was added and somebody quickly produced something in-house. A supplier delivered files for two articles. The latest collection went to a different photographer. Every one of those decisions was reasonable at the time.

You do not notice the result product by product. You notice it when you put twenty product pages next to each other and see that the white is not the same white, that one product fills the frame while the next one floats in the middle, and that some articles have six images while others have a packshot and nothing else.

That is why a page-level audit will not find this. If you want to know whether one page convinces, run our visual checklist for a product page, which takes about ten minutes. If you want to know whether your range reads as one brand, you need rows rather than pages.

Step 1: get every product page in front of you

The audit opens with a dull step you cannot skip: everything in one place. One line per product, with enough detail that you can judge it without opening the page again.

Here is how we do it. We load a brand's website into our tooling and pull in all of its product pages. Webshops already publish their product data for Google, and that feed carries the SKU, the name, the category and the alt text of the main image. Our website scanner reads that data and turns it into one line per product in the Products module, which is currently in beta. It runs in the background and emails you when it is finished.

What is missing then stands out immediately: a product with no image, a main image with no alt text, a SKU without a proper name, a row where the product information is only half filled in. Those are exactly the gaps a manual round misses, because they are invisible on the page itself. From that same overview you fill in what is missing, or you ask us to produce the imagery that is not there.

A few things no scanner can know. Note those yourself, because you will need them to judge:

• •Which types are live: packshot, close-up, scale, product in use, lifestyle
• •The long edge in pixels of the main image
• •When the image was made, and by whom
• •Where the original file lives
• •Which channels show this product

If you are not working in tooling yet, an export and a spreadsheet will do the same job. The work is identical, it just takes longer and you are more likely to skip a row.

You do not have to cover the whole catalogue at once. Take five to ten products per group, always including the bestsellers and the newest items. If the same gap keeps turning up in that sample, it is a group problem and it gets solved as a group.

Step 2: score six checks, not taste

The overview shows you what is missing. It says nothing about the quality of what is there. That part needs your eyes, and six fixed checks so it does not turn into a conversation about what looks nice. Give each check a rough verdict: fine, borderline, gap. Anything more granular is wasted, because you will act per group anyway.

1. Set

Does the set of images answer the questions buyers actually ask you? For most physical products there are three: how big is it really, what does the surface look like up close, and what exactly is included. A dimension drawing does not solve scale; the product beside something familiar does. Do not count images. Ask whether the last one removes the last doubt.

2. Style

Put four or five products from the same group side by side. Same background, same light, same angle on the first image, roughly the same distance? This is the check that most often surfaces a problem nobody reported, because each image on its own is perfectly good.

3. Resolution

Enough pixels to zoom into, and enough for the largest channel the file has to serve. What each channel needs is set out in product image resolution per channel.

4. Currency

Does the image show the product as it is now? Packaging changes, a component is replaced, a colour leaves the range. This is where a catalogue ages most quietly, because the image still looks good.

5. Channels

Does the image exist in the crops your channels ask for? A landscape file that only exists in landscape will be cropped to square or vertical somewhere by a system that does not know where your product sits in the frame. Whoever makes that crop decides what the customer sees.

6. Findability

Is the right file easier to find than the wrong one, for everyone who uses your images? Your own team, your dealers, your resellers. This check earns its place because it explains why good imagery often still goes unused.

Step 3: rank by cost, not by lowest score

Once the scores are in, the temptation is to start at the bottom. That is exactly the wrong order. A product that barely sells and has three gaps costs you less than a bestseller with one weak first image.

So plot your list against revenue or units and split it into four groups:

• •Fix first: sells well, images fall short. This is where the return is.
• •Batch later: big gap, little revenue. Handle these in bulk, not product by product.
• •Keep an eye on: sells well and the images hold up. Mainly, do not break anything here.
• •Leave it: small gap, little revenue. Not every gap needs closing.

Add next quarter's launches straight into the top group. Imagery that is right at launch never has to be caught up on later.

Step 4: write the standard down, or it drifts again

An audit without an agreement is a snapshot. A year from now you will be in the same place with different products. So write down what applies, and keep it short enough that someone actually reads it. One page will do:

• •What a new product needs before it can go live: which image types, how many, in which crops.
• •What the first image looks like: background, angle, how tightly the product sits in the frame.
• •Which resolution you keep as the master, and what you export per channel.
• •Where the files live and who is allowed to add or replace them.
• •What you expect from suppliers who deliver imagery, and what you send back when it does not meet the standard.
• •What dealers and resellers receive, and how they can tell a newer version exists.

Those last two points are often the difference between a catalogue that stays tidy and one that has drifted again within two seasons. They are also the bridge to the final step.

Step 5: run the same check outside your own webshop

This is the step most brands skip, and it is the one where we find something most often. If you sell through dealers, resellers or marketplaces, the majority of your product pages are not yours. That is where your customer actually sees your product, and that is where the imagery is most often out of date.

We run this comparison as a matter of course before we talk to a manufacturer, and it works like this:

• •Pick two or three bestsellers or featured products from your own site and note the number of images, the type and the quality for each.
• •Look each product up by EAN, article number or product name, and pick three to five parties that sell it.
• •Confirm it is genuinely the same product first. If you are in doubt, leave the match out.
• •For each party, note the number of images, the type, the sharpness, whether the images are current and whether colour variants are missing.
• •Summarise it in a few lines per product. No more than that.

What you will usually find is the same product presented in three different ways inside a single market. The images came from whichever file someone happened to have when the listing was built: a press shot from two seasons ago, a crop out of a catalogue PDF, sometimes a photograph of the box. And sometimes the imagery at your partners is simply good. Then there is nothing to win, and we say so.

What matters is the kind of problem this is. It is not a page problem and you cannot fix it page by page, because those pages are not yours. It is a supply problem: the right file has to be easier to find than the wrong one. How that plays out on marketplaces is in standing out on marketplaces, and the layer underneath it in digital asset management: the last mile.

What we would skip

Chasing a perfect score. Arguing about the ideal number of images per product, when the real question is whether the last image removes the last doubt. And above all, booking a new shoot for everything that came up red.

A good share of the gaps you find do not call for a new photograph at all. A missing square crop, an export that is too small, an outdated file still sitting with a partner: those are solved with material you already own. New imagery is for the gaps that are genuinely missing, such as scale, texture or the product in use.

Where LensLab fits

You already met our side of the audit in step 1: the scanner that pulls in your range and shows you where the gaps are. What happens next is the part that counts.

From that same overview you request visuals in bulk for the products at the top of your list, using Visual Sets as a reusable combination of image types. Our creators produce those visuals with AI against your Brand DNA, so background, light and angle are decided once and still hold for the products you add next quarter. You only pay for the visuals you accept.

Per product you decide who sees it: all connected dealers, a dealer group or individual dealers. They download the approved files in their own portal, which is exactly the part that makes step 5 of this audit smaller over time. Integrations with Shopify and WooCommerce are in development.

If you want the practical sequence, you can see how it works, or look at what this means for brands like yours in our use cases.

In closing

Pull in your range, score six checks, rank by cost, write the standard down, and only then look outside. The first three steps fit into an afternoon if you work from a sample. Step 4 is the only one that stops you from repeating all of this next year. And step 5 takes about an hour and usually produces the longest list.