Before exploring this study, we advise you to read the previous report, “Ecommerce Ranking Factors – 2019”. It provides detailed industry ranking information worth knowing. And now – about this year’s data and changes.
Already a year ago, in Yandex, the top three positions in the niche query “gloves for fitness” were occupied by giants. Now the first niche store does not appear until positions 6-10 (the issue is often rotated). Above are the largest stores, marketplaces and aggregators.
The dominance of large sites in Yandex has intensified: it is no longer only in the top 3 and top 10, but also in the top 30. At the same time, the sites in the search results continue to age at an accelerated rate. Links are still important (especially in Yandex), and the “link weight” of the site as a whole is more important than its relevance and the weight of a particular page. Commercial parameters (phone 8-800, reviews, selection by parameters, etc.) and technical ones (https protocol, adaptive layout, micro-markup) continue to grow, and the size of the SEO text is reduced.
Now – in more detail.
Search engines don’t like risk
Search engines don’t like to take risks. They prefer large sites, usually universal, over smaller sites, including specialized ones.
Parameters directly or indirectly related to the size of the site (ICS, traffic parameters, a huge assortment, presence in many regions, link parameters (site), the number of pages in the index, the number of pages found on request, etc.) are very strong, and the impact over the year this parameter in Yandex has grown significantly.
Another consequence of search engines’ dislike for risk is the aging of search results. In 2019, the average age of sites in the top 3 (in our “general commercial” sample) was 10.7 years for Yandex and 11.5 for Google. Moreover, over a year, the average age increased by more than a year – that is, sites were replaced on average with older ones! This trend continued: the age of sites continued to grow at an explosive pace. In the top 3 of Yandex, sites have aged from 12.3 to 13.9 years in a year, in the top 3 of Google – from 12.2 to 13.5 years.
Links to the site are more important than to its individual pages
Reference parameters can be divided into four groups under two categories:
to the site as a whole (domain) and to a separate page (URL);
aggregate link mass (purely link parameters) and links with texts relevant to the request (link-text).
Over the year, the parameters related to the total link weight of the site have grown, and therefore are directly related to its size. For example, the number of domains referring to the site. The median values for the top 3 Yandex have almost doubled, for the top 30 – four times, and now for e-commerce, site link parameters in Yandex are noticeably stronger than in Google. Parameters related to links to a separate page and link-text have dropped, especially in Yandex.
“Social” parameters are not directly related to the size of the site, but usually large sites and social networks are doing well. Therefore, over the year, the share of sites with groups in social networks has grown significantly (although it was already rather big); parameters in both search engines have increased. The highest correlation coefficients with position are VKontakte in Yandex, Facebook and Twitter in Google.
The value of commercial parameters is maximum
Commercial parameters and e-commerce are simply created for each other (which is already clear from their names).
Presence in several regions is very important – preferably in many at once. It is good if there are a lot of phones and / or addresses, or at least a few – this correlates with multi-regionality and site size. There is at least one address on 98–99% of sites in the top 30 of Yandex or Google; telephone – almost 100%. At the same time, almost a third – federal number 8-800, and only mobile numbers – only 1% (a year ago it was 3%).
The situation with ordering a call back is controversial. In Yandex, there is still a connection with getting to the top, but the correlation with the position is negative. On Google, the parameter doesn’t seem to affect rankings. The online consultant, according to our data, has never really helped the ranking, and now the parameters of this group have become openly negative.
The presence of prices on the site and on the page are important parameters, the values of which in our sample are close to one. The number of prices per page also matters. Large sites are characterized by large showcases with more than 50 product cards with prices. The prices themselves do not affect the ranking, but it is important that promotions and discounts are indicated.
photos, and several photos for the product;
selection by parameters and comparison;
reviews on product pages;
video on the site (search engines do not expect to find a video on every page, although there is a weak correlation with position here too);
certificates on product pages, if needed;
recommendations on the choice on the site (and on the page, but there – only for Google).
Payment and delivery are almost indispensable attributes of an online store. Their descriptions are not required on every page, but they should be easy to find – for example, in the site menu. The share of sites providing non-cash payment options has noticeably increased – by card (already 86% in the top 30 and 96% in the top 3 of Yandex), Yandex.Money (about half as often) and Webmoney (a little less often);
Of the delivery methods for both search engines, fast local ones (courier, self-pickup) are much more important than slow long-distance ones (mail or transport companies), which are also usually provided, but did not and does not affect the ranking.
Technical parameters must be up to the mark
Sites that make it to the top 30 for competitive searches tend to follow search engine recommendations – the higher the better. Over the year, the corresponding parameters have grown even more.
The protected protocol https: // is already used by 94–96% of sites in the top (and Yandex is ahead of Google here), and a year ago it was “only” 85%. The situation is about the same with responsive layout – only in Google there is no correlation with position. It is also important to have micro-markup – in particular, Schema.org markup. The parameter value has grown over the year, especially in Yandex.
The size of the HTML-code of the pages (including scripts, etc.) has noticeably grown in the top 30 of both search engines. At the same time, the share of text in the code is a strong “negative” parameter: as it approaches the first position, it drops, and for sites in the top 3 Yandex, text is now responsible for only 4% of the total page weight. The large size of pages in the top does not lead to the fact that they are loaded for a long time. The median time to load the code (albeit without loading external scripts and even more so their execution) is less than a second.
SEO texts need to be abandoned
The size of SEO text continues to decrease, as does its share in the total page text – not only in Yandex, but also in Google.
The number of occurrences of query words in the SEO text is a strong negative parameter (the less, the better). The same is true for other parameters for SEO copy.
The parameters for entry into “text snippets” – the part of the page without SEO text – are stronger than the corresponding parameters for the entire text of the page. Occurrences of individual words of a query are more important than occurrences of all words in a row, and even more so – the exact query. The exception is
) on Google. Here, just for the exact query, a strong connection with getting to the top is visible. There should be a lot of query words on the page. In Google, they are typed by almost 35 requests (it has grown over the year), in Yandex – by almost 39 (it has stopped growing).
During the year, the answers to questions have noticeably grown in Yandex, but “weakened” in Google. Publications (articles and reviews) have become more important in both search engines, especially in Yandex, but reference materials in both Yandex and Google have “weakened”.
The risk of Baden-Baden is still high: Yandex applies this filter in a metered fashion, and what you have always gotten away with may disrupt your business tomorrow. Use Turgenev and sleep well!
The full version of the report is on the Laboratories page.
The study was prepared by Mikhail Volovich.