Creating effective ads for Yandex.Direct

Benchmarking can help you create effective ads in Yandex.Direct.

Benchmarking is the process of identifying, understanding, and adapting existing examples of a competitor’s effective performance in order to improve their own performance. It equally involves two processes: evaluation and comparison. Benchmarking can be considered as one of the areas of strategically oriented marketing research.

How does this work in Yandex.Direct?

If you have read the reference information, you know about such parameters as «productivity» and «quality factor«. If you simplify it, then productivity evaluates the theoretical efficiency of the phrase, and the quality factor — the practical one. Yandex.Direct also has a function for testing ads in groups.

Conclusions from this can be drawn as follows:

1. Each ad can be checked for compliance with the recommendations for improving productivity.
2. It can be assumed that the part of advertisers that uses the functionality of ad groups already shows winning options in the search.

In other words, if we collect competitor ads for a specific query, we will get both untested and obviously winning options. All we have to do is check them for compliance with Yandex’s recommendations and sort out the garbage. Based on what remains, you can already create your own ad variants.

Practical example

To begin with, let’s take the segment of the semantic core that interests us, and select from it all queries with a frequency higher than zero.

Here is a sample of 130 phrases containing occurrences of the words «ad» and «yandex direct»:

The «Frequency» column shows the number of impressions per month with a wide correspondence, and the «Competition» column shows the number of ads (advertisers) at the time of analysis.

You can view competitor ads in Yandex. Direct on this page — https://direct.yandex.ru/registered/main.pl?cmd=showCompetitors&phrase=ключевая%20фраза

For further work, we add up the received ads, for example, in Excel. Format — two columns, the first one contains the ad title, and the second one contains text. It turned out to be about 9,000 ads. We remove duplicates, and there are significantly fewer unique pairs left — about 1000. Now our task is to select the most «productive» ones.

Productivity is determined to a greater extent by the title and to a lesser extent by the text. The roots of this automatic assessment, apparently, lie in a study by Zakhar Plyushch published by Yandex in 2005 — https://blog.yandex.ru/post/10118/.

So, for soft filtering, we need ads that have:

• at least one of the query words is found in the text — +10% of the CTR
• at least one of the query words is found in the header — +15-17% of the CTR

If we want (and can) make the conditions more stringent (that is, if the amount of data allows it) then from the received ads, you should also select those with the title starting with the word contained in the request — this will give +25% to the CTR.

However, since we are conducting research and are not just going to «copy» our competitors ‘ ads — we can easily analyze the headlines and texts separately. Also, since we are benchmarking an entire «segment» rather than a single query, we will make the requirements more stringent and select ads that contain only both the source words «ad» and «direct».

As a result, we will get 51 texts, each of which should presumably increase the CTR by 10%:

And 58 headlines with various options for increasing clickability:

The results obtained should be studied and taken as a sample.

Based on the data obtained, write the desired number of your own headings and texts and test them in various combinations.