A good product at a fair price can sit unsold because of the listing itself. Two readers look at it at once: the shopper, who reads text and photos, and the marketplace's algorithm, which reads only structured fields — attributes, category, text format. A mistake the shopper never notices can keep the product out of a filtered search; a mistake the algorithm never flags can send an interested shopper away within seconds. Here are the five mistakes that hit both readers, and what to do about each today.
1. Incomplete specifications
A shopper rarely notices a missing attribute directly — they just don't see the product, because they filtered by "size: M" or "material: cotton," and an empty field never enters that filter. To moderation, an empty required field is usually grounds to hold the listing back, not just rank it lower.
Weak: "Blue dress," size, material and season fields empty — reachable only by someone browsing with no filter applied. Fixed: the same fields filled from the category's value list — size S/M/L, fabric composition, season — so the item surfaces in every filter a shopper would use.
Which fields are required for which category is its own topic, covered elsewhere. Today's action: open the listing with the lowest views in your best-selling category and check whether every field the add-product form offers is filled in, not only the ones visible on the page.
2. A description that says nothing
A shopper reads "quality product at an affordable price" and closes the tab — it answers none of their actual questions: what it's made of, what's included, how to care for it. Search gets nothing useful from it either: title and attributes carry more weight than description, but description is where longer, less predictable phrases get matched — the ones no attribute list covers.
Weak: "Quality backpack at an affordable price, comfortable and stylish." Fixed: "City backpack, padded 15-inch laptop sleeve, water-resistant fabric, reinforced straps, detachable waist strap."
Action for today: rewrite one description as made of → used for → included → what sets it apart, then check how much of the result carries information versus adjectives.
3. Poor or borrowed photos
A shopper looking at a dark or blurry photo can't judge colour or texture and moves on; one carrying someone else's watermark reads as resale from someone else's stock, true or not. Moderation reads that watermark differently — a legal risk, another seller's brand inside your listing — and rejects on that basis regardless of photo quality.
Weak: a photo shot on a supplier's phone — dim lighting, another shop's price tag in frame, a supplier logo in the corner. Fixed: the same product on a plain, light background, two or three angles plus a close-up on one distinguishing detail.
Action for today: check photos on your five best-selling listings first, where a weak photo costs most, and replace any carrying someone else's watermark or price tag.
4. One text for every marketplace
A shopper compares listings within one marketplace, and text built for a different one stands out as foreign: a block copied from elsewhere overflows its field or keeps formatting the platform won't render. Each marketplace reads the same data differently and moderates it differently — see how the generation process works shows how differently one source ends up formatted per marketplace. Unadapted copy usually fails review over structure, not content.
Weak: an emoji bullet list copied from an Instagram caption, dropped into a marketplace that doesn't render it — collapses into one unbroken block of text. Fixed: the same content split across that marketplace's own fields — title, attributes, description — each in its own format.
Action for today: pick the marketplace you copy text into unchanged most often, and check its field limits and format there.
5. Inconsistency across the catalog
A shopper scrolling a category compares listings without thinking about it, and when one item's size reads "42-44" and the next reads "M," the catalog looks assembled from several sources in a hurry. Search and sort degrade the same way: sorting by size or unit price slips whenever one field holds values in different formats, because the algorithm can't treat "42-44" and "M" as comparable.
Weak: three listings for the same model record size as "M," "Size 42-44," and just "42." Fixed: all three read "42-44 (M)," the template used across the rest of the catalog.
Action for today: pick the field where the mismatch is most visible — usually size or unit of measurement — and bring it to one format across the whole catalog.
Auditing an existing catalog without opening every listing
Working through a catalog of several hundred items row by row is the slowest way to find these five mistakes. Faster: export the catalog to a spreadsheet — your marketplace panel or CRM will do this — and look at columns, not products.
Empty attributes show up at a glance: sort the attribute column — size, material, colour — and empty cells collect at one end. For descriptions, sort by text length; the shortest rows are usually the same generic phrases. For consistency, count distinct values in a column that should have a fixed list: sixty ways of writing a size where six or seven should exist is visible without comparing listings by eye.
A spreadsheet won't show a photo, so start that check with your best-selling listings instead, where a weak photo costs the most.
This pass doesn't replace checking each listing individually, but it finds the bulk of the problem far faster than reading the catalog one item at a time.
How to speed this up
Selba generates complete listings to each category's rules and keeps one consistent style across the whole catalog, including the attribute format that would otherwise need manual alignment. You export the finished file and upload it yourself. Browse examples for your kind of business. Try it free.