A shopper narrows a search down with the filters on the left of the results page. Stainless steel, dishwasher safe, 4 litres and up. Your product is all three of those things, it says so in the second bullet, and it is not in the results. Nobody removed it. It was never eligible for that filter, because the field the filter reads was empty.
This is the quietest way to lose sales on Amazon, because there is no notification and no error. The listing is active, the ads still run, and a slice of the most qualified traffic never sees the page. A shopper filtering by material is further into the buying decision than one typing two generic words.
The part that will annoy you is that this is not a writing problem. The bullets can be excellent. The attribute behind them is a separate field in a separate part of the record, and free text has never populated it.
What attributes actually are
Every product type on Amazon has a defined set of fields. Not a suggestion, a schema. Amazon's own developer documentation describes it plainly: for a given product type in a given marketplace, Amazon publishes the attribute and data requirements as a JSON schema, including which fields are required, which are optional and which become required only under certain conditions. Source: Amazon, Amazon primary.
Two consequences follow, and they explain most of what sellers find confusing here.
The first is that the fields differ for every product type. A monitor has a display size. A shampoo does not. A dress has a neck style that means nothing to a kettle. So an account level rule like "fill in all the important attributes" does not survive contact with a real catalogue, because importance is defined per product type.
The second is that required and useful are not the same thing. A listing goes live once the required fields are present, and the fields that decide whether a shopper can find it by filtering are often the optional ones. That gap is where most catalogues sit: technically complete, commercially invisible.
- Persuades a shopper who is already on the page
- Gets indexed for phrases
- Can say stainless steel in any wording you like
- Does not make the product eligible for a filter
Decides whether they buy
- Uses a fixed value from Amazon's list for that product type
- Feeds search filters and refinements
- Gives an AI layer an unambiguous fact to match on
- Reduces returns by setting expectations early
Decides whether they arrive
Why blank attributes matter more than they used to
Filters have always run on these fields. What changed is the second reader of them.
Amazon's shopping assistant, renamed Alexa for Shopping in May 2026, answers shopper questions from listing details, reviews and community questions, and the intent layer underneath connects products to use cases, audiences and occasions rather than to phrases. Material stated in a structured field is a claim a system can rely on. The same fact buried mid-bullet has to be inferred from prose, and inference loses to certainty.
Amazon has been telling sellers this indirectly for a while. Its own generative listing tool is built to fill more than 70% of required product attributes, with examples like colour, style, theme, occasion type and whether batteries are required. Source: Amazon, Amazon primary. A company does not build automation for a field it considers decorative.
How to tell whether this is your problem
- Filter your own category the way a buyer would. Pick the three refinements a serious buyer would use, apply them, and look for your product. Absence from a filter your product genuinely qualifies for is the clearest evidence there is.
- Open the Listing Quality Dashboard and read the recommendations by benefit. The path is Inventory, then Manage All Inventory, then Listing Quality Dashboard, and Amazon groups missing attributes by what filling them is meant to improve, including search results and returns.
- Check the ASINs that came from a bulk upload or a migration. Files built to get products live carry the required fields and almost nothing else, which is exactly the pattern that produces an invisible listing.
- Look at your returns reasons. Wrong size, wrong material and not compatible are usually attribute problems wearing a customer service costume.
What to check before you start filling fields
There is a temptation to fill every box that accepts a value, and it is worth resisting, because a wrong attribute is worse than a blank one. A blank field makes you invisible to a filter. A wrong one puts you in front of a shopper who returns the product and says so publicly.
- Confirm the product type is correct first. Attributes are defined by product type, so a product filed under the wrong type is being measured against the wrong schema, and filling that schema carefully just makes the wrong answer more complete.
- Pull the current values rather than trusting the front end. The Category Listing Report shows what is stored against your ASINs, which is regularly not what the detail page implies.
- Decide who owns the answers. Material composition, capacity, compatibility and age range are product facts, so they come from whoever holds the spec sheet, not from whoever is writing copy that day.
Auditing attributes without turning it into a quarter long project
This is the sequence our team uses, and on 50 to 100 SKUs it is a day of work rather than a month.
- Export the catalogue and group it by product type. Everything downstream happens per product type, because that is the level the schema exists at.
- Pull the flat file template for each product type in play. The template carries the field list and the valid values for that type, which is the fastest way to see what exists that nobody filled.
- Mark the fields a shopper can filter by. Material, size, capacity, colour, item form, compatibility, age range, quantity. These are the ones with a direct route to eligibility rather than a theoretical one.
- Fill them from the spec sheet using Amazon's valid values. The valid value list matters more than the wording, because a value Amazon does not recognise reads as no value at all.
- Upload in small batches and read the processing report. A file that partially processes is a real outcome, and the summary at the top is where that gets discovered.
- Wait, then verify. Attribute changes commonly take up to about 48 hours to reflect, so an immediate refresh proves nothing.
- Re-run the filter test from the diagnostic section. The product either appears in the refinement now or it does not, and that is a cleaner test than any score.
Where the flat file itself fights back, our flat file errors page covers the failure patterns, and a misfiled product type is a different fix from a missing attribute. Our team re-categorised an educational products listing within 24 hours after a flat file miscategorisation, and it re-indexed with no further suppression on that ASIN.
Which fields carry the weight, by category
Field names differ per product type and per marketplace, so this is a starting map rather than a list to copy.
| Category | Fields that decide findability | Fields that reduce returns |
|---|---|---|
| Apparel | Size, colour, department, fit type, sleeve and neck style | Material composition, care instructions, size chart |
| Supplements and grocery | Item form, unit count, diet type, flavour | Ingredients, allergens, dosage, storage |
| Home and kitchen | Material, capacity, colour, item dimensions | Care instructions, heat tolerance, what is included |
| Electronics and accessories | Compatible devices, connector type, display size, power source | Included components, cable length, model compatibility |
| Beauty and personal care | Skin or hair type, item form, scent, size | Ingredients, usage instructions, quantity |
| Toys and baby | Age range, theme, character, material | Safety warnings, assembly required, battery requirement |
The pattern across all six rows is the same. The findability column is what a shopper narrows by before choosing anything, and the returns column is what they check once they are already interested.
Keeping the fields filled after the audit
- Make attribute completeness part of the launch checklist, not a clean-up job. Twenty minutes at creation costs less than a re-upload and a 48 hour wait later.
- Re-check the dashboard monthly. Amazon adds and reweights fields, so a listing that was complete in March can carry recommendations by August without anyone touching it.
- Keep one spec sheet per product as the source of truth. When the answers live in someone's memory, values drift and the UK and US records stop agreeing.
- Treat backend search terms and attributes as separate systems. They are filled differently and read differently, and our backend keywords page covers that field on its own terms.
What we would do first if this were our account
Take the ten ASINs carrying the most revenue and run the filter test on each, because that takes an afternoon and turns a vague worry into a specific list.
Then pull the flat file template for those product types and fill the findability fields from the spec sheet, since those have defined answers and no judgement in them.
We cannot promise a ranking change from this, because attributes are one input among several and Amazon does not publish the weighting. What is fair to say is that a product cannot be filtered into a result set on a field it has not filled, and that is a mechanical fact rather than a theory.
If the page also has to show those facts to a shopper once the fields are right, our A+ enhanced brand content page covers that side.
If you want to know which of your ASINs are sitting outside the filters your buyers actually use, that is part of the free, no-obligation audit, and if your attribute coverage is already strong we will tell you that instead of selling you a project.
Related guides
Common questions about Amazon product attributes
Do attributes affect ranking or only filters?
Filters are the provable effect: a blank field means no eligibility for that refinement. Beyond that, attributes are structured facts Amazon's systems can use with certainty, which matters more as intent based matching grows. Amazon publishes no ranking weight for them, so anyone quoting one is guessing.
Is it enough to put the information in the bullets?
For persuading a shopper who is already reading, yes. For eligibility in a filter, no, because the filter reads a structured field and not prose. The two live in different parts of the record and they do different jobs.
What happens if I fill a field with a value Amazon does not recognise?
Usually nothing visible, which is the frustrating part. The value sits in the record without matching a recognised option, so the listing behaves as though the field were still blank. The valid value list in the flat file template is the reference that avoids this.
How long do attribute changes take to show?
Commonly up to about 48 hours to reflect after processing, and filter eligibility can lag a little behind the detail page. A refresh five minutes after upload tells you nothing.
Should I use Amazon's AI to fill attributes?
As a drafting tool it saves real time, and Amazon states that sellers are responsible for the accuracy of content they accept from it. So generated first pass, human check against the spec sheet, because a confidently wrong material value costs more than an empty one.