Cosmetics store SEO: brand, series, shade — and three different pages

Most often because it is not in the site map. This is a literal find from our point of view: a store with a working language switch and declared hreflangs did not have any address of the second language version in the map. The search engine can reach it by links, but it receives a completely different priority of bypass. First, we add these URLs and compile the hreflang so that the versions don't conflict with each other in the same release.

A familiar situation. In the wholesale store we measured, the canonical map gives 4,167 addresses, and the old one, according to the standard path, gives 7,782 entries for 2,755 unique addresses, of which about 2,706 are commercial. One is duplicated, the other is incomplete, and the crawler most often follows the standard one. We do this: we leave one card, remove the second, write the address in robots.txt and check the bypass. We immediately warn about a side effect: the number of pages in the index may drop in the first weeks. This is the expected course of events - duplicates come from the index, not your directory.

It depends on whether they are looking for a specific shade. If in your niche the requests look like "lipstick 214", a separate address is justified. If not, options in one card give fewer pages and fewer duplicates. There is no universal correct answer, there is a calculation: how many addresses will be added and how many of them have a chance to receive demand. Our default recommendation is options inside the card, and to open individual addresses point by point, for requests that are already visible in your statistics. Expanding the structure later is cheaper than gluing together thousands of nearly identical pages.

Address - yes, text - not immediately. First, all brands get their own addresses with the correct title and page name, this is closed with a template. Then a normal text is written for 15–30 of the most important ones. To write a description for each brand at once is to spend the budget on pages that no one comes to. And a fair boundary: the presence of a page in the site map does not mean either indexing or traffic - Google writes it directly.

For general queries like the name of the category, practically no, and it is more honest to say it right away. For narrow links "brand + series + type of means" and for wholesale requests - yes, because there is less competition and more sense in the detailed page. We build our plan precisely on the second. Separately, about the brand: the manufacturer itself competes with you for its own name, and no one can beat us.

He decides, but not in the way that it is customary to scare. We measured a natural cosmetics store with a first response time of 126.3 ms—the fastest in our sample—while having 271 sitemap entries for 116 unique addresses. A fast site and a clean index are two different things. Speed ​​removes technical loss: a person does not get tired of waiting. It does not make the offer more attractive and does not glue doubles.

Because "standing" and "collecting" are not synonymous. The cosmetics store in our sample had the Universal Analytics counter UA-148400680-1 disabled by Google back in July 2023 in the code, and there was no GA4 property at all. Formally, there is analytics on the website, but in fact, nothing was collected for three years. Also, no illusions: the counter itself does not improve anything, it only gives an opportunity to see, and the blockers and the consent mode make the numbers a trend, not accounting.

Technical changes are visible in Search Console in 3-6 weeks: the number of indexed pages and crawling frequency change. The demand for specific items depends on the season and assortment. Monthly report: what has been done, how many unique addresses are in the working card, how the "duplicate" and "not indexed" statuses are moving, what proportion of cards has the correct marking. Everything to your addresses - we do not bring other people's interest.