Catalog completeness: how to measure what each product is missing

Why the percentage misleads

“The catalog is 72% complete” sounds like information, but it does not say what to do. Is everything missing a little, or are a few products missing everything? Is the missing data required by a channel or data nobody reads? Are these products that sell or products nobody looks for?

Completeness is useful when it answers a concrete question: what is this product missing to sell on this channel?

Complete against what

Two references matter:

• The product's family. The attributes that belong to that type of product, with the required ones marked. A shirt is not missing “power”; a drill is.
• The channel. The columns each destination requires. A marketplace may ask for a global identifier, a product type or photos with a certain background that your store does not ask for.

A product can be complete for its family and not ready for a channel, or the other way around. That is why both should be measured.

From measure to work list

The measure only helps if it turns into concrete work. What works is a list with three columns: the product, the channel and what is missing. With that list:

1. Group by missing field. If two hundred products are missing the same attribute, it gets solved once, not two hundred times.
2. Sort by what sells the most. Completing the biggest earners first pays off more than going in alphabetical order.
3. Separate what can be filled automatically from what needs a person. A lot of data is already in your spec sheets or descriptions; some only the supplier knows.

See what the channel would receive

A number on a dashboard does not feel the same as seeing the product record the way the buyer would. It helps a lot to open a product and see what the store would receive: the category, images, description, details, and where there is a gap. That gap is the difference between “attribute 14 is missing” and “this page goes out without dimensions”.

What to do with incomplete products on export

When it is time to send the catalog to a channel, there are two options for incomplete products:

• Send them as they are, if the channel accepts it and the missing data is not critical.
• Leave them out until they are complete, if the channel would reject them or, worse, if an empty cell would erase data the channel already had.

What matters is deciding it on purpose, not finding out when the channel rejects the upload.

Common measurement mistakes

• Measuring against every attribute that exists, instead of the product family's. Everything looks incomplete and nobody knows where to start.
• Counting a junk value as filled. An “N/A”, a dash or “see description” is not data; treat them as empty.
• Measuring only your own catalog and forgetting the channel's requirements, which are what decide whether a product gets published.
• Celebrating the average. A catalog that is 90% complete can have its ten best sellers without photos. Always look at the list, not just the number.

Measure it every week

Completeness is not fixed once: new products arrive, channel requirements change and attributes are added. Reviewing it every week, with the same measure, shows whether the catalog is improving or falling behind, and on which channel.

How Bruxel handles it

In Bruxel completeness is measured product by product against its family, and each channel states which columns it requires and what each product is missing for it. “In the store” shows the product record built from what your channel would receive, with the gaps marked and a link to where they are filled; the In the store guide explains it. When you export you can leave incomplete rows out, and the catalog quality report puts everything in one file. For the basics, read product attributes.