Ukrainian fields in a Prom feed: how many stores actually supply them

On 8 September 2026 we read 21 public YML feeds from Ukrainian online stores — 89,949 products — and counted exactly one thing in them: whether the Ukrainian-language fields name_ua and description_ua are present, and whether the keywords tag is filled in. This continues the measurement of 5 September, which did not count language fields at all.

The short answer: a Ukrainian version of the title and description is supplied by 5 feeds out of 21. That is 46,391 products, 51% of everything we read. The remaining 16 feeds — 43,558 products, 48% — do not have these fields at all. keywords is filled in by one store out of twenty-one: 5,819 products.

Broken down by what the feed is for

The feeds in the sample were built for different marketplaces, and the picture differs across them.

The first row is the point. Seven of the nine feeds built specifically for Prom do not supply a Ukrainian version of the title and description at all.

One more observation, small but telling: name_ua and description_ua always appear as a pair — in none of the 21 feeds is one present without the other. So these fields are not filled in one at a time: either the export module switches the whole block on, or it does not switch it on at all.

What these fields are

name_ua and description_ua are the Ukrainian-language variants of the title and the description inside the same product entry. name and description stay as they are, and the Ukrainian text sits next to them in the feed.

keywords is a separate tag holding the product's key phrases. Its role is auxiliary: relevance and the marketplace's internal search. It stopped being a "traffic magnet" of the kind people believed in around 2010, and stuffing a hundred queries into it does more harm than good.

Judging by the fact that seven of the nine Prom feeds live without the language fields and still work, these fields are optional. An empty name_ua breaks nothing: the product is exported, no error appears. That is precisely why they stay empty — nothing hurts.

Why they are empty

The reason is not laziness but arithmetic.

The description arrives from the supplier in one language. The export module puts into the feed what the store's database holds — the title and the description from the product card. For description_ua to appear, somebody has to write a second version of the text, and not once but for every product. For a 300-item catalogue that is a week of work. For a catalogue of 12,000 it is months that nobody has.

Then ordinary postponement logic kicks in: the field is optional, the products are on the marketplace anyway, so the task moves to "some day". In the measured sample, 16 stores out of 21 stopped exactly here.

How to fill them in across thousands of products

This is a job for a pipeline, not for a copywriter. We covered the pipeline itself — feed in, feed out — separately: bulk catalogue rewriting. Three of its properties matter here.

First. In Textory, our own product, each product goes through a single structured call to the AI that returns at once the rewritten description, a new title if needed, keywords and a meta description; the target name_ua and description_ua for Prom are generated in the same place and written into the export. When all the fields are born from one context they agree with each other: keywords match the description, meta does not contradict the title, the Ukrainian version does not drift away from the main one. Do it in four separate passes across four tools and that consistency has to be assembled by hand.

Second — choosing the export fields. Not every store wants everything rewritten: sometimes only the language fields are needed while titles and attributes must stay untouched. So the feed carries out what you ticked, with a Prom preset.

Third — a live preview of one product before the whole catalogue is launched. It is the cheapest way to see what you will actually get, before thousands of items are processed.

The order we recommend:

1. Open your own feed in a browser and search it for name_ua. Not there — question closed, the fields are empty.
2. Run the preview on one product and read the text with your own eyes.
3. Tick which fields go into the export and build the file.
4. Import one product, open its card on the marketplace, and only then upload the whole catalogue.

The fourth step saves hours of rollback if something is off in the mapping.

A separate limit of the rewriting itself: the description is built from your attributes, it is not invented. Numeric attributes must stay in the structured fields of the feed rather than dissolve into the text.

What this does not give you

Filled-in name_ua, description_ua and keywords do not lift positions on their own. They remove the technical under-filling of the product card, and that is all. Ranking on a marketplace, as in Google, depends on dozens of factors: assortment, prices, delivery terms, buyer behaviour, competition inside the category. The text is one of them, and far from the largest.

Nor does any of this relate to products being rejected. An empty description_ua is not the reason an item fails to pass: different rules operate there, and we went through them for another marketplace in the piece on why products get rejected when the validator says "OK".

And one more thing: the presence of Ukrainian text does not by itself make a description fit for citation in AI search. That is a different task and a different structure — we wrote about it in the article on GEO optimisation of descriptions.

The limits of this measurement

Precision matters here, because a number like this is easy to stretch over the whole market.

The sample is small and not random: 21 feeds, collected by searching for the typical paths of an export module. This is the state of public feeds on 8 September 2026, not a cross-section of Ukrainian e-commerce.

We counted the presence of a tag, not the quality of what sits inside it. We did not check whether name_ua differs from name at all, whether it is machine translation or human writing, or how long the descriptions are. A tag being present does not mean it is filled in meaningfully. It is entirely possible that in some of those five feeds description_ua holds a copy of the main text, and by our method it was counted as "present".

In short

Ukrainian language fields in a feed are not a marketplace requirement — they are simply what most stores do not fill in, because across thousands of products it cannot be done by hand. In the measured sample seven of the nine Prom feeds do without them, and keywords is supplied by one store out of twenty-one.

You can see what comes out on your own feed at textory.com.ua — the pricing model there is per volume of text, with current terms on the product site. Our other products are in the own products section. If the task is wider than text — exporting the catalogue to a marketplace, syncing stock, passing orders into a CRM — that is website integration with Prom. Reviewing the task is free, we answer within two hours, and we work under a contract.