Aggressive ad formats in the Yandex network: when a high CTR is bad news

Anyone who learned display advertising in Google’s world carries a simple instinct: a placement with a high click-through rate is a placement that works. The audience is interested, the creative lands, the money is well spent. In the Yandex Advertising Network, known in Russian as РСЯ, that instinct will cost you money.

Some placements and formats there produce a CTR close to the theoretical ceiling and send no inquiries whatsoever. Nothing is broken in the reporting. The formats are designed to trigger a click, and interest has little to do with it. This article covers where that false success hides, why it looks normal in a dashboard, and how to separate working placements from the ones that burn budget. It also covers two things that are often mistaken for problems, and one change on the Google side that quietly widens reach for anyone running both systems.

Why a high CTR can be bad news

The main source of misleading numbers is aggressive formats. A full-screen block opens in the first second after a page loads, before the visitor has oriented themselves. It can be closed, but the close button does not appear straight away, and in that gap accidental taps accumulate. Many of them are entirely unintended: a thumb reaching for the article lands on the ad instead. On such placements the CTR jumps toward its limit while engagement and traffic quality stay at the bottom.

The formats keep developing. When a full-screen block cannot be shown because of a frequency limit, some placements compensate with several blocks in a row that cover the screen, and get the same forced contact in the first second. A smaller block may sit on top of that. This started as a mobile phenomenon. Similar interstitial formats were later extended to desktop, so abnormal CTR is no longer a purely mobile story, although mobile is where it is most visible.

The network’s usual response to questions about these formats is that test data shows no noticeable effect on user metrics. Observed practice disagrees: the CTR takes off, and no growth in relevant inquiries follows it.

The practical conclusion is blunt. A CTR well above the campaign average on a single placement or a single device slice is not evidence of quality. With high probability it signals that your ad is being shown in one of these formats. Such placements are found through the report slice and cut deliberately. Celebrating the overall rise in clicks is the wrong reaction.

Cutting them does not close the problem for good. New sites pass moderation and appear in place of the blocked ones, so the work is a standing routine of monitoring the placement and format slice, not a one-off clean-up. CTR has stopped being a reliable indicator of quality in the network, and it now has to be read with the placement format in mind.

What this means if you are used to Google

Teams arriving from Google Ads tend to bring over one of two habits when they meet the Yandex network: treating it as a secondary channel and ignoring it, or judging it by the same CTR logic they use for search. Both are expensive. The network can deliver inquiries, but only when it is read through placement-level reports and trimmed continuously. If your agency or in-house specialist sends you a monthly report with a single network CTR figure and no placement breakdown, that report is hiding exactly the thing you need to see.

Several of your ads on one page is not a bug

A second observation was treated as a violation for years: several ads from the same advertiser visible on one page of a placement. It looks like a glitch, or like something the network should penalise the site for.

In fact it follows from the architecture. A page can carry several independent ad blocks, and each one runs its own separate auction. The rule of one ad per advertiser applies to a block, not to the page as a whole. Your ad can therefore technically win the auction in several blocks on the same page. The network’s official explanation confirms this: the behaviour is expected.

Knowing this saves effort. There is no point filing complaints or hunting for a breach where none exists. The attention belongs on what you can actually manage, which is the quality of the placements themselves. For a foreign company that audits its Russian account from a head office, this is worth writing into the review checklist, so that nobody raises a ticket over a screenshot showing three of your ads in one feed.

When OR replaces AND and reach quietly inflates

Mobile and network traffic can also swell through targeting logic the advertiser may not be aware of. This part concerns Google, and it matters to anyone who runs display campaigns in both systems and assumes they behave alike.

In the Google Display Network the logic for matching conditions changed at some point. Previously, an ad set up with a topic and a placement was shown only when the site’s content matched all of the conditions at once. That was AND logic.

Now the same conditions work as OR: a match on any one of the parameters is enough, whether that is a topic, a placement, keywords or a video. For anyone who used to narrow delivery by intersecting several conditions, this is a reversal. A combination of conditions no longer narrows reach. It widens it. Settings built for the old logic start delivering more broadly than intended, and nothing in the interface warns you.

The habit worth keeping is simple: check how targeting conditions are actually combined in each system, and never assume that several criteria automatically mean a narrower, more precise audience. When a team copies a display setup from one platform to the other, this is one of the places where the two stop being comparable.

Filtering placements with ads.txt before they get in

Manual cleaning after the fact is one defence against junk placements. There is also a technical filter at the entrance. A network ad can only be shown where the site runs the corresponding monetisation system. That means a long list of domains can be checked in bulk in advance instead of one by one.

The mechanism relies on the file at /ads.txt, a public standard maintained by IAB Tech Lab. A successful server response to a request for this file tells you the site monetises through advertising at all. The content gives a more precise signal: a line that names the ad system together with the relationship type confirms the specific monetisation you need.

The check can be automated in two ways. A spreadsheet formula handles the quick version. A script handles parsing the file contents. If a match is found, the domain is marked as confirmed. If there is no response or no matching line, it is marked as unconfirmed. The point is to keep domains where your ad can never appear out of the whitelist, and to avoid spending manual review time on them. Confirmed domains go into the placement whitelist, where they are then sorted by topic.

The signal is indirect, and it helps to be honest about that. The file may be missing or incomplete even on suitable sites, and the line format changes over time, so the criterion itself needs periodic rechecking.

For a company running Russian campaigns from abroad, this approach has a practical advantage. Evaluating Russian-language sites by hand requires someone who reads the language and knows the local web. A technical pre-filter shrinks the list that person has to look at, which matters when that reviewer is a scarce resource.

Video: a hand-built whitelist

Video follows different logic, and automation hits a ceiling here. There is no mass tool that collects relevant videos and channels the way domain metadata can be collected. Selection is done by hand: browsing the platform and judging the content.

A browser extension speeds this up. While watching, you tick the videos and channels you want, then export them as two separate lists, one of channels and one of videos. Those links become a placement targeting criterion when the ads need to appear strictly on specific videos and channels.

The method is piecework. It suits a narrow, precise set of placements and does not scale to thousands of items. The extension is also a third-party tool that may stop being maintained, so a critical process should not depend on it.

What to do with your network campaigns this week

• Open the report sliced by placement and by device. Treat abnormally high CTR by cutting the placement, not by celebrating the growth.
• Stop spending effort on complaints about several of your ads on one page. That is block architecture, not a breach.
• Recheck how targeting conditions are combined. Today several criteria more often widen reach than narrow it.
• Run domains through their /ads.txt file before letting them into the network, and keep a whitelist at the entrance.
• For video, build a whitelist of videos and channels by hand, without counting on bulk automation.

If you sell into Russia from another country, add one more item: make sure whoever runs the account reports the network by placement and format, and that exclusions are reviewed on a schedule rather than once at launch. The detailed routine for that is in the article on cleaning placements in the Yandex ad network, and the fraud side of the same problem is covered in the piece on click fraud in the Yandex advertising network.

Aggressive formats and mobile placements survive because their metrics look excellent. A CTR near the ceiling, plenty of contacts, everything appears to work. That is exactly where the budget leaks: the number pleases the eye, and the inquiries never arrive. Whoever reads the network by the placement and format slice, rather than by the top-line metric, pays for attention instead of a thumb that missed.

If you want to find out which placements and formats are eating your network budget, book a free review. How placement control fits into ongoing account management is described on the Yandex Ads page.

Frequently asked questions

Sources

• Yandex Direct help, the Yandex Advertising Network and placement settings: yandex.com/support/direct/en/campaigns
• IAB Tech Lab, the ads.txt specification: iabtechlab.com/ads-txt
• Russian version of this article on belokrylovo.ru: Агрессивные форматы РСЯ и мобильный трафик