Yandex Direct introduced a new approach to A/B-testing

Yandex Direct introduced a new opportunity to not only test different versions of advertising or landing pages, but now it is also possible to compare among themselves any settings, types of campaigns and even media plans on the basis of statistics. For this purpose the team of Yandex Direct released Experiments tool.

Advertisers with experience in working with analytics and with enough data saved up now have an opportunity to test different scenarios of working with advertising with Direct's new tool of A/B testing, which is available on demand.

"Think out the hypotheses, check them, try options which you doubted or with which you are just not familiar yet. Experiments will allow you to estimate on the basis of data what contribution to the general results of placement is received from branded or video advertising, announcements in Yandex Advertising Network and on search and many other things. Opportunities for different scenarios are almost not limited" — Direct's experts report.

Before starting the experiment you will need:

• To formulate a hypothesis that you want to check. The hypothesis has to affect change of one parameter. In one experiment it is recommended to test only one hypothesis.
• To check if you have access to Experiments in Audiences, Direct and Metrics.
• In Metrics you need to have goals set up.

To start an experiment, you will need:

• To create an experiment in Audiences — to get segments of users for an audience share.
• In Direct you need to link the segments to the campaigns under testing.
• To compare statistics on test segments and all other users using Master of reports by Direct and data of the Metrics.

Experiments are made to help you check which ad campaign settings are the most effective for your business. You will be able to learn how different types of campaigns affect each other, analyze and compare the effectiveness of automated strategies and manual ads bid strategies, and so on.