Low-value or low-demand page in Yandex: how do you fix it?
The "low-value or low-demand" label means Yandex removed a page from search, but did not penalize the site. This happens when the page copies another one, the robot cannot see its text, other pages answer better, or nobody searches for it. First check the text and copies in Yandex Webmaster (your site's account in Yandex). If everything is fine, name the page with the words people actually type.
Roman Kaufman, founder of Kaufman Lab · Data as of: October 2026
On September 9, Yandex put this label on two articles on our site. Each had more than two thousand words. The site stayed in search, only these two pages disappeared. We edited the texts but did not rewrite them, and the main fix was the title. Why one line can matter more than a whole article becomes clear from how Yandex rates pages.
The label removes a page from search and Alice, but does not touch the site
While the label is on, the page is not in Yandex search results, and so it is not in the answers of Alice, Yandex's AI assistant: Alice AI builds its answer from search results. Our articles were out of Yandex search for 6 and 15 days, even though both stayed in Google's index the whole time.
The rest of the site is not affected by the label. The Webmaster help page "Low-value or low-demand pages" (September 2026) says such pages do not mean any violation, and their number does not affect the site's rankings. A penalty is shown separately, in the "Security and violations" section.
Thresholds from agency articles, such as the claim that up to 10% is fine and rankings drop from 40%, do not point to any measurement. You do not need to save the site. You fix the page.
Yandex keeps a page that can rank high for real queries
According to the help page, the algorithm rates every page almost every day: whether search will show it for queries high enough for a person to find it. Yandex does not give a threshold in numbers. Length did not save our articles. And if there are several similar pages for one query, only one stays in search.
There is one label but two causes, and Webmaster does not say which one your page has. Yandex calls a page low-value if it is a copy of another page, or if its text is hidden in an image or a script and the robot cannot see it. A low-demand page works fine, there are just no queries for it.
A copy and invisible text are easier to check, so demand is left for last.
A page nobody searches for is fixed starting with the title
For a working page that nobody searches for, check the name first. Webmaster's recommendations say that without a word from the query on the page, search is unlikely to show it as a good answer. A title that nobody types will not bring queries to the page.
On September 12, we checked our old titles in Yandex Wordstat (Yandex's free keyword tool): "AI-ассистент для ежедневных задач" ("AI assistant for daily tasks") and "Автоматическое производство статей" ("Automatic article production"). Both got 0–19 searches per month. Meanwhile "как сделать ии ассистента" ("how to make an AI assistant") gets 396 searches per month, and the version spelled with the Latin "ai" instead of the Russian "ии" only 53. The word "автоматический" ("automatic") almost wipes out demand. The addresses stayed the same, the articles got new names, and "Личный ИИ-помощник" ("Personal AI assistant") came back to search on September 15.
The second one did not come back. On September 17, it got the name "Нейросеть пишет посты для соцсетей" ("AI writes social media posts"), taken from Yandex search suggestions. At the same time we noticed that the title promised a how-to guide, but inside was a project review. On the 20th, we put the steps at the start. On the 22nd, we sent the article for recrawling, so the robot would visit it again. On the 24th, it came back.
We cannot prove it was the title that brought it back. During the same days we edited the questions, added steps and asked the robot to visit. But a new text takes days of work, while a title can be checked with one query. Your page may also be named the way you describe your work, not the way people search for it.
How to find out what is wrong with your page and get it back in search
- Text. In Webmaster, open Indexing → Pages in search → Excluded pages. Paste the page address into "Page check", and you will see what text the robot gets. Turn text from an image or a script into plain text.
- Copy. If there is a similar page, set canonical in the copy's code, that is, the address of the page you are keeping.
- Demand. Type the title words into Wordstat without quotes and for all regions. Then it also counts longer queries with these words (Yandex Direct help; Direct is Yandex's ad service). If there are only dozens of queries, take a phrasing from the search suggestions on ya.ru (Yandex's search page) and change the title. Then send the page for recrawling and add it to "Important pages monitoring", and Yandex will tell you when the status changes. We aim for 300–3,000 queries per month, but this is not a Yandex rule.
If there is demand but the page is still not in search, check whether it gives what the title promises. And when you do not need the page in search, block it from showing with a noindex tag in the code, as Yandex advises.
Frequently asked questions
How long does it take for a page to come back after a fix?
A page can come back within two weeks after it is crawled again, but this is not guaranteed. Our first article came back three days after we changed the title, and the second two days after recrawling.
My page keeps appearing and disappearing in search, even though I did not touch it. Is this normal?
Yes, Yandex reviews its decision almost every day. A page can drop out and come back without any edits, because the search results themselves change. The robot also visits excluded pages, and before each index update Yandex checks again whether they have a chance to be shown.
I deleted the page or blocked it from search, but it is still listed as low-demand. Why?
Yandex uses the copy from its index until the robot visits again and sees the deletion or the block. If you block these addresses in robots.txt, the file with rules for the robot, they will drop out of the index within two weeks.