Contextual advertising trends in 2021


In 2020, Yandex and Google had some pretty big changes in terms of advertising tools. Each of the changes made quite a stir, especially the changes in Yandex. As a result, many contextualists were faced with a problem: previously effective techniques are no longer working, the cost per click has increased, the conversion fell. What to do?

To solve this problem let’s make a retrospective review of changes in Yandex and Google in 2020, carry out their analysis, identify trends and how to adapt to them. At the end of the article we will give specific recommendations for businesses on how to work with contextual advertising in 2021.

Contextual advertising in Yandex: what has changed in 2020?


  • Payment for conversions


It is a logical continuation of the development of machine learning, optimization technologies and automatic bids, which promotes Yandex in recent years.

  • Disabling customization for additional relevant phrases


This change was positively received, until complaints emerged in professional communities on social networks. Complaints pointed to the fact that with the innovation there was a noticeable increase in traffic by synonyms, and some keys were not even synonymous to the phrases specified in the settings. However, in most cases, the synonyms are the same as those specified in the settings. It should be noted that communities are a good indicator of negativity, indicating that there are problems.


  • Autotargeting


Recently we are seeing more and more examples of ad groups on autotargeting bringing conversions similar to the results of ads for which manual settings are used.


  • Managing bids for different devices


The ability to separately manage bids on different types of devices: desktops and mobile devices (tablets have been left out for now).


  • Correcting coefficients for different OS


It is possible to add correction coefficients for different operating systems of mobile devices: iOS and Android. However, the setting for separate management by types of operating systems is not done very well. Now the specialists cannot, for example, raise the rates on iOS, and lower them on Android: you can select only one thing. There is hope that in the near future this will be finalized and the tool will become full-fledged.


  • Auto-segments for retargeting


The appearance of auto-segments for retargeting in Yandex.Metrika was not a notable event. In fact, this is quite an important change. We will talk about it in more detail below, in the recommendations for businesses on setting up autotargeting.


  • Look-alike audiences


Now created in fact automatically directly in the settings of Yandex.

White list of advertising platforms


The fact that caused the greatest amount of discussion: Yandex has decided that he is now sufficiently well aware of where the traffic, which sites can be called more quality, and which – no. This he took away from the experts on contextual advertising possibility of minusatsii traffic from these sites. There have been a lot of complaints in the core communities that these sites bring the “wrong” traffic, it converts “not so”, the applications from this chart are not as desired. Yes, at this point this change seems unpleasant. But it fits in with the overall strategy of developing the Yandex.Direct tool. Yandex is also actively experimenting with the placement of online ad campaigns.  For example, ads appear below the search bar in Yandex.Browser when the user has not yet entered any query. Although the contextual advertising system is based on a semantic kernel of queries, predictive analysis is spreading in search engines.

Opinion of an expert

Alexey Avdeev, Performance Group Head at Ashmanov & Partners:


Both Yandex and Google have a huge amount of data on each user: gender, age, history of visiting various sites. Search engines know what the user is looking for, what the user is buying, what the user is interested in. So they can predict the likelihood of purchase on your site even before the user entered the query.

This also applies, by the way, to large marketplaces. Here comes to mind Amazon with its claimed system that allows you to be proactive, “to send a parcel to the buyer before he even orders it”. All of this would be impossible without the history of orders, analysis and study of user behavior and their data. What conclusions can we draw? Some changes were received positively by the market, while others, on the contrary, were assessed negatively. The main thing is to understand where contextual advertising in Yandex is moving, what you need to change in your approach, or if there is no way to change it, how to adapt to them.

Read also:   How to come up with a company name: best examples + 11 ways

Changes in Google Ads in 2020

The changes that have taken place here echo those that have taken place in Yandex.Direct to some extent.

On the one hand, Google has expanded the types and formats of ad campaigns and developed “smart technologies. On the other hand, it reduced the manageability of the ad campaign by abandoning the usual ad formatting. The expansion of coverage was similar to changes in Yandex due to traffic by synonyms.

The amount of data available for analysis in Google Ads for search queries has decreased. The search engine thus says that it does not consider it necessary to process more data, because the main conversions are carried out for high-frequency queries. In fact, he considers unnecessary statistics for low-frequency queries longer than 70 characters. There is logic in this: they have a dramatically decreasing share of traffic. Such queries a lot, respectively, labor costs for the collection of semantics and analysis of statistics are high, and the share in the profit is very small.

An important change – the optimization indicator in the advertising campaign settings and the introduction of recommendations on settings for Google Ads. It should be noted that recommendations on settings also appeared in the new Yandex interface.

In addition, Google in 2021 will change its approach to assessing the quality of the site, read about the new requirements here.

What should contextual advertising professionals prepare for in 2021?

Analyzing the changes, we see that both Yandex and Google are tending to reduce the manual nature of the advertising campaign. At the same time, there is an upward trend towards automated management of ad campaigns through machine learning by the ad systems themselves.

In fact, ad systems are now telling experts, “We know better where to place what bids, give us the main direction, and we’ll do everything ourselves.”

With dynamic ads, it is enough to make a list of headlines, ad texts, pictures, and Google itself will make ads based on this data. And it will show them to users who, in Google’s opinion, are more likely to click on the site or take a targeted action.

Opinion of the expert

Alexey Avdeev, Performance Group Head at “Ashmanov and Partners”:


Contextual advertising professionals need to retrain themselves. On the one hand, they need to immerse themselves in machine learning technology to understand how it works and be competitive in the future. On the other hand, they need to move from the technical work of assembling a complete semantic core and writing a separate ad for each key to working on ad campaigns from a marketing perspective. You need to learn how to form a UTP and create a seamless logic between the target audience, the creative and the landing page.

How should businesses run their advertising campaigns in 2021? 7 ground rules

Based on changes that have already occurred, we determine the trends for further development and can make recommendations for contextual advertising in the coming year.

Business should enlarge advertising campaigns

Optimization and machine learning search engines are at the stages of keywords, ads, ad groups, campaigns and do not go to the level of the advertising cabinet. Accordingly, in order for ad systems to have more data to optimize an ad campaign, we need to give more data to machine learning. The bigger the campaign we make, the more data we collect and give to search engines, the better the results of automated strategies will be. The development of machine learning in search engines is not going backwards, so we need to learn how to work with this in mind.

Use auto-strategies

Often there is a contradiction: business is interested in the cheaper clicks, and Yandex, on the contrary, more profitable expensive clicks. Does not become a contextual advertising under autostrategy too unprofitable? No, it will not. Like any company, Yandex is interested in maximizing its profits, but ration it per unit of time. Yandex does not benefit from taking all the “cream” from the market at once and thereby killing it. The company is trying to grow its market to get a steady income from it.

Read also:   TikTok launched a new feature - automatic subtitles

Quite a few experiments were conducted on auto-strategies in 2020. It turned out that the advertisers, for which there is a fairly large number of statistics, autostrategies work successfully. The following trends are worth noting:

  • The development of voice search leads to an increase in the number of queries that contain many words and have lower frequency. The number of such queries and their share in the semantic core are increasing. At the same time, in most cases, these queries will get few hits. This is precisely the reason why Yandex and Google are actively promoting the idea of synonyms.
  • Machine learning technologies can now quite clearly recognize speech, translate it into text, and determine the meaning of text. They can do the same with pictures – by the visual row of a picture, both Yandex and Google can understand what is shown on it (for example, a car of a particular model). Therefore the search by picture also works effectively.
  • Text classification and tone analysis is the next stage of machine learning. Thanks to him, search engines determine by the content of a test query whether it is a commercial or an informational query. Even if the request looks the same, search engines determine what the user wants now based on previous requests.
  • Video is a trend, and the share of spending on video ads is growing because they have more engagement than pictures. We can assume that the next change in ad systems in 2021 and beyond will be an increase in the number and types of ad creatives with more engagement than the average picture. These are, of course, video and animated banners and other elements.
  • Pictures, text and music are what machine learning tools can create. By and large, headlines in our ads can already be made automatically. Soon it will be possible to generate pictures directly for ads. Video editors already know how to create videos that can be used for contextual advertising systems.

Send more data to analytics systems

For autostrategies to work better, you need to pass more data to analytics systems. Be sure to transmit call-tracking data, and preferably offline conversion data. It’s also helpful to send cost and click data from other advertising systems.

If there are not enough macro-conversions to make autostrategies work, you can use micro-conversions as an intermediate link between clicks and macro-conversions. Not all of them, of course. You need to calculate the Pearson correlation coefficient between these micro and macro conversions. If the correlation coefficient is high enough and there are more microconversions than macroconversions, you can use microconversions as the target on which to optimize your ad campaign.


All changes in advertising systems are made based on the analysis of statistics. For example, if you sell emergency services (tow truck for a broken down car, repair a burst crane, etc.), your customers simply do not have time to choose long, they buy from the first run. Then it makes sense to block all retargeting and generally separate primary and secondary traffic.


Now there is some confusion between the contextual advertising tools of Yandex.Direct and Google Ads and targeting advertising systems. Yandex and Google are working more with search queries and are actively moving towards audience and behavioral targeting. It should be noted that Mail.Ru now has a fairly large audience – VKontakte, Odnoklassniki, MoiMir. This is largely targeted, audience and behavioral traffic. This year launched targeting by key requests in VKontakte, the most popular social network, in fact, entering into competition with search engines.

Properly work with the creatives and PPM

Creativity – a way to communicate with the visitor, which should make an immediate impression and tell clearly what we sell and why you should buy from us. Now less time is spent on gathering semantics and setting up rates, but more attention is paid to the creation of your own PR. It is necessary to create a creative that will be relevant to the USP and which would be aimed at the selected segment of the target audience.

The creative should contain concreteness. Clichés like “fast delivery”, “100% guarantee” people do not react, you need to broadcast an elaborate SUE. Of course, you need different formats of creatives for each audience segment. The extent to which creativity creates engagement determines the quality of traffic.

Read also:   Twitter will begin testing the tweet-editing feature in the coming months.

Do splitting: for each segment of your target audience, find out why that segment buys from you. Perhaps because of price, delivery time, brand, trust in the store, etc. Then for each segment, you can create essentially a separate UTP. From them you need to build a seamless logic between the segment, the creative, the message and the landing page to which the traffic goes.

Pay attention to quality when creating landing pages

In 2020, because of the pandemic, there has been a fairly large increase in competition and paid advertising budgets. At the same time, the audience has not grown significantly, its increase is much less than the increase in competition. Particularly increased competition at the lower stages of the sales funnel, where there is a user with an already established demand. Plus, competition from social networks has increased at these stages. Therefore, attracting visitors to the site is becoming more and more expensive every year.

In 2021 you need to optimize not only contextual advertising, but also landing pages, so that the conversion on them would be higher. Working with paid sources of traffic, you need to look at what can be done with this traffic, how to process it further, so that the visitor is more likely to take the targeted action on the site.


Working with the entire search engine space is a factor in getting additional traffic and conversions

Businesses need to understand that they are not the only ones selling a given product. Type in any search term for which you sell products and see what and who surrounds you in the search space. These are competitors, marketplaces, reviewers, aggregators, geoportals, search engine services, video review sites and more. You also have to work with all of these sites, not just your own site.
There is a certain limit of the market that can be covered by the number of ad impressions as much as possible. For example, a user has asked a search query, but hasn’t clicked through to the site because the budget doesn’t match what the marketplaces offer in that cluster. Nevertheless, we can use the surrounding search space for additional touch points with potential visitors.

If a business can’t compete with marketplaces that have million-dollar promotion budgets, they can sell their product there. There are times when a product is not bought on the first click, but there are a fair number of informational requests or requests for product reviews. Then efforts should be made to have the customer read the reviews not only on the site, but also on third-party review sites.

Another trend: the growing share of Yandex services in organic output. We can not affect this, but we can and should use it. If a business provides services, it is worth placing an ad in Yandex. Services and thereby increase the number of contacts of the service with the target audience.

Opinion of the expert


In 2021, it is effective for businesses to look not at how services and marketplaces prevent them from rising to the top of search results, but how they can use these services and marketplaces. Clicks are expensive, and these places of presence provide additional opportunities, including free ones, to connect with potential buyers.

You can read tips on how to work with search space in different niches here.


The changes taking place in Yandex.Direct and Google Ads have a common foundation: the development of machine learning in search engine technology.


  • Machine learning in contextual advertising systems reduces the manual controllability of the advertising campaign. This is the main trend of 2021. Autostrategies, which require more data, more statistics from businesses, are the mainstream.
  • A marketer needs to know how to work with this. If some actions do not lead to the planned result, you need to understand that it is not in the settings, but in the approach. It is important to work with the entire search engine space, promoting your product/service on sites that are at the top of search results.
  • Reducing the manageability of the advertising campaign brings to the forefront the work with the audience, creativity and UTP, with landing pages and search space. There is always something for a specialist to do.


Leave a Reply

Your email address will not be published. Required fields are marked *