Product data quality: seven mistakes that cost sales

Quality shows up in sales

A catalog with poor-quality data does not look bad on a dashboard: it shows in operations. Products a channel rejects, filters that find nothing, returns because the size was different, customer service questions the product page should have answered. Each mistake looks small; together they cost sales every day.

These are the seven mistakes that repeat most in catalogs of every industry.

1. Duplicate or changing SKUs

Two products with the same SKU overwrite each other; a product that changes SKU between uploads appears twice. Since the SKU is the key that links every system, a mistake here spreads to the store, the ERP and every marketplace.

2. Unnormalized values

“Navy blue”, “Navy”, “Marine” and “NAVY BLUE” as four different colors. The store filter shows four options and each one finds part of the products. List attributes with allowed values prevent it.

3. Measurements without units

“250” does not say whether it is milliliters or grams; “12” does not say whether it is inches or centimeters. Without the unit, the value cannot be compared, filtered or converted for another market.

4. Specifications lost in the text

“Made of 2 mm stainless steel with a non-slip handle” inside a paragraph is information the store cannot filter or send to a channel. Specifications go in attributes; the description mentions them.

5. Different records on each channel

When each channel is edited separately, information drifts: a different price, dimension or photo in each place. Buyers who compare lose trust, and buyers who purchase receive something they did not expect.

6. Unreviewed images

Photos that don't match the variant, images without alt text, the same photo for three different colors. The image is the first thing buyers look at and the last thing anyone reviews.

7. Data without an owner

When nobody is responsible for a field, nobody fixes it. An attribute a supplier filled in two years ago is still there even though the product changed. Without an owner and a change history, nobody knows who changed it or when.

How to spot them

You don't need a huge audit to find these mistakes. A few simple checks are enough to start:

• Repeated SKUs: sort the catalog by SKU and look for the ones that appear twice.
• Unnormalized values: review the list of distinct values for each list attribute; spelling variants jump out.
• Measurements without units: look for values that are just a number in measurement attributes.
• Differences between channels: compare the record of your ten best sellers on each channel where they are published.
• What customers ask: the questions that keep reaching customer service point to the field that is missing or wrong.
• What channels reject: the error reports from each marketplace upload are a free list of the fields that need attention.

How to keep them from coming back

Fixing by hand works once; rules work every time:

• Typed attributes with allowed values and units.
• Required fields by family and by channel.
• Review before publishing, with a clear approval step.
• A change history that says what changed, who changed it and when, and lets you go back.
• A single source for the product record, exported to each channel instead of edited in each one.

How Bruxel handles it

In Bruxel attributes are typed, with value lists and measurements with units, and families say what is required. Each product record keeps a history of what changed in each version and lets you go back to an earlier one, and the catalog quality reports show what is missing and where. Everything is exported from one place to each channel, as the export guide shows. To measure what is missing, continue with catalog completeness.