Item Specifics and Structured Data

01

What structured data actually does

A marketplace listing has two kinds of content. There is unstructured content — your title, description, and photographs — which a person reads. And there is structured content: category, brand, size, colour, material, model, condition, and dozens of category-specific attributes. The structured part is machine-readable, and it drives four things that happen before a buyer ever sees your words.

1. Filters. A buyer who ticks a size, a brand, or a colour removes every listing that has not declared that attribute. Not ranks them lower — removes them.
2. Search relevance. Attributes contribute to matching, often more reliably than keywords stuffed into a title.
3. Catalogue matching. On platforms with a product catalogue, identifiers attach your listing to an existing product page along with its accumulated demand.
4. Buyer confidence. A complete specifics table answers questions that would otherwise become messages, or become a decision not to buy.

02

Category comes first, and constrains everything

Category is not a label. It determines which specifics fields exist, which are required, which filters your listing can appear under, and how the marketplace interprets everything else you enter. Choosing it loosely propagates errors through the whole listing.

• Pick the most specific accurate category, not a comfortable general one. Buyers browsing a broad parent category are browsing; buyers in a leaf category are shopping.
• Check what the fields are before committing. If a category is not offering the attributes that describe your item, it is usually the wrong category.
• Look at where comparable sold items sit. The market has already made this decision many times; copying its answer is faster and more reliable than reasoning it out.
• Do not category-hop to escape competition. Listing in a quieter but wrong category reduces competition by also reducing buyers, and it can attract enforcement.
• Expect the same item to be categorised differently across channels. Taxonomies are not shared, and a mapping that works on one marketplace is often meaningless on another.

That last point is the one that makes multi-channel selling harder than it looks. Each platform maintains its own tree with its own depth and its own naming, and the correct category for your item on one is frequently not even conceptually present on another.

03

Filling specifics properly

1. Complete every required field honestly Required fields are usually required because buyers filter on them. Fabricating a value to clear a validation error is worse than leaving the listing unpublished — it puts you in results where you do not belong and generates returns.
2. Then complete the recommended and optional ones This is where most of the available advantage sits, because most competing sellers stop at required. Every additional accurate attribute is another filtered search you can appear in.
3. Use the platform's own values, not your own wording A free-text "navy blue" may not match the "Blue" filter. Where a dropdown exists, the dropdown value is the one that counts.
4. Add manufacturer identifiers where they exist Model numbers, catalogue numbers, and standard product codes are what connect a listing to a catalogue entry and to buyers searching by exact identifier.
5. Match measurements to the stated convention Units and sizing standards vary by region and by category. A number entered against the wrong convention is a wrong number, and in clothing it is one of the surest routes to a return.
6. Leave a field empty rather than guessing An absent attribute costs you one filter. A wrong one costs you a return, a refund, and possibly a defect on your account.

The compounding effect is what makes this worth the minutes. Ten accurate attributes do not put you in ten searches — they put you in every combination of filters a buyer might apply, and buyers who filter narrowly are buyers who know what they want and are close to purchasing.

04

Identifiers and catalogue matching

Where a marketplace maintains a product catalogue, a correct identifier attaches your listing to an existing product record — inheriting its images, its specification, its reviews where applicable, and the search demand that record has already accumulated. It is the single highest-leverage field on a catalogued item.

It also has the sharpest failure mode. Matching to the wrong catalogue entry produces a listing that describes something you are not selling, with photographs of an item the buyer will not receive, and the resulting dispute is difficult to defend because the listing genuinely misrepresented the goods.

• Verify the match against the item in your hand, not against the title you searched with. Variants, editions, regional releases and years diverge in ways the name does not capture.
• Be careful with near-identical products. Successive model years, pressings, and revisions are the usual source of mismatches.
• Do not force a match that does not exist. Some items genuinely have no catalogue entry, and a listing without one beats a listing attached to the wrong one.
• Where the catalogue is wrong or incomplete, say so explicitly in your description — this is common in music and collectibles, where cataloguing is community-maintained.
• Never rely on catalogue images to describe your actual item. They show the product; your photographs show the goods. See product photography and condition grading.

05

Keeping structured data consistent across channels

Structured data is where cross-listing quietly degrades. The item is the same everywhere; the fields describing it are not, and the usual result is a good listing on the channel you built it for and a thin one everywhere else.

The structural fix is to stop treating each listing as the record. Capture attributes against the item once — at intake, while it is in your hands — and let each channel receive a translation of that record. This is the model Instica is built on: one item is one record, and each connected channel reads its listing from that record — so a correction is made once rather than three times and forgotten twice. It is also the practical argument for leaving spreadsheets behind, covered in migrating from spreadsheets.

06

Auditing what you have already listed

Existing listings are the cheapest inventory you will ever improve, because the item is already photographed, described, and live. A structured-data pass over stale listings frequently produces more sales than new sourcing does in the same time.

1. Start with aged listings that get impressions but no sales. Traffic without conversion is a listing problem; no traffic at all is usually a specifics or category problem.
2. Count blank fields in your worst-performing category. The result is usually startling and immediately actionable.
3. Check the category against comparable sold items, and move anything sitting in a parent node.
4. Search your own listings using a buyer's filters and confirm they appear. This is the fastest way to find an attribute that is missing or entered in a form the filter does not recognise.
5. Fix the intake process afterwards, so the same gaps are not recreated on everything you list next month.

When a listing has good specifics, a fair price, honest photographs and still does not sell, the problem is elsewhere — which is the subject of when listings do not sell.