Struggling with a Text Filter
1. We started with hiding from indexing all the texts that existed at that time. Next, a text analysis was carried out across the entire list of requests. All clusters (all of them or all high-frequency queries from the cluster) were investigated according to the following procedure:
- first, a group of queries was loaded into a text analyzer,
- then, based on the results obtained, the corresponding page was edited and styled (meta tags, headings, text field, text fragments, etc. changed / added / removed / closed from indexing).
2. In addition to individual changes on each page, the following alterations have been implemented on the website as a whole:
- template texts from catalog pages and product cards were removed;
- the number of the word "price" on the listing has been optimized;
- A "Buy in 1 Click" option was added onto the listing and in the product card;
- the number of H2 tags on the listing was reduced;
- in order to combat overspam, the number of keywords in the text block for linking tagged pages was minimized.
It should be noted that consistent work on correcting the site's shortcomings based on the results of text analysis began to take effect after a couple of updates, which largely contributed to the optimism of the project team.
In total, it took about 2 months to get the site out of the text filter.