Completeness is the most reassuring metric in product data and one of the least informative. It answers a question about your schema: of the fields you decided to have, how many are populated? It cannot answer the question you actually care about, which is whether those were the fields worth having.
A schema that never asked for screen size cannot report screen size as missing. The gap does not appear anywhere in the reporting, because from the schema's point of view there is no gap. That is how a catalogue reaches 93% complete and still fails the shopper who knows exactly what they want.
What shoppers say about it
Akeneo's 2025 PX Pulse survey found the share of consumers frustrated by incomplete or unclear product descriptions rose to 34%, up from 28% the year before. That is a vendor's survey and worth reading as such, but the direction is consistent with the usability research: the problem is getting worse as catalogues get deeper.
Two ways of scoring the same laptop shelf
our own measurement, reproducible from the repositoryCompleteness counts the middle bar. A shopper experiences the bottom one.
View as table
| Value | |
|---|---|
| Products on the shelf | 216 |
| Fields populated (93.1%) | 201 |
| Still indistinguishable | 94 |
Readiness asks a different question
Instead of asking how full the fields are, ask this: after every question your catalogue can currently answer, how many products is a shopper still choosing between?
That number has a property completeness does not. It gets worse when you add products without adding facts, which is what actually happens as a range grows. And it improves only when you record something a decision turns on, which is exactly the behaviour you want from a metric that is supposed to direct work.
On the laptop shelf we measured, the catalogue scored 93.1% complete and still left 94 of its 216 products indistinguishable once every question it supported had been answered. Both numbers describe the same shelf. Only one of them describes the shopper's experience of it.
Why this is not an argument against your PIM
A PIM governs, enriches and distributes product information, and it does that well. Completeness is the right metric for the job a PIM is doing: it tells you whether the process is running. It was never designed to tell you whether the schema is right, and holding it to that is unfair to a tool that is doing what it was built for.
Completeness is a process metric. Readiness is an outcome metric. Reporting the first and managing to the second is the mismatch.
The practical consequence is that the priority list has to come from somewhere else, and then go into the PIM as work. Not instead of it.