If a typical PPC specialist with no experience in Yandex.Display decides to test display campaigns in Yandex.Direct, then most likely he will try to use them as a performance tool. I decided to take the place of such an advertiser (especially since I personally have no experience with display advertising either) and try to get conversions from the “media”.
Let me remind you that display advertising has slightly different tasks and they belong to the highest level of the sales funnel. For example, with display ads you can:
- increase brand awareness;
- convey brand positioning to the target audience;
- reach a wide and new audience.
The latter option seems to fit our performance goals: for our webinars, we need a new audience that is not yet familiar with the eLama brand.
My cunning plan consisted of three points:
- Launch a CPM display campaign.
- Collect audience by pixel.
- Launch CPC retargeting on it and thereby reduce costs.
I was promoting a webinar for beginners about the basics of PPC advertising. According to Crypta, most of the users who have already registered for our webinars are interested in IT, real estate and finance, travel the world and the country, drive cars, play sports and buy online, some of them are students.
Advertising launch and results
I created an ad campaign with geo-targeting to Moscow (since Yandex should have more data for Muscovites) and several targetings – each in a separate ad group. For the purity of the experiment, users who have visited the eLama site in the last 90 days and all users under the age of 18 have been excluded from the campaign. Since Pixel only needs one impression per user to collect audience, the strategy was chosen – reducing the cost of repeated impressions.
Ad group 1: targeting persistent ad interest
This group generated 67 clicks and no conversions.
Here’s what Krypta said about the interests of users selected for this targeting:
It turned out that the most targeted interests for eLama – finance and IT – are at the very bottom of the list. User categories also don’t match what we expected, with housewives and gamers on the list.
Ad group 2 – targeting persistent interest in creating and promoting websites
This group of users includes webmasters, SEO specialists – people who know what contextual advertising is, but, perhaps, have never launched it and would like to learn how to work with it.
The results were almost the same as in the first group: 79 clicks and only 1 conversion.
Let’s see what other interests this audience has, in addition to creating and promoting sites:
In the first positions among the interests are again pets, cooking, and finance and IT – at the bottom of the list. The categories overlap with the characteristics of eLama’s target audience, but only partially.
Ad Group 3 – Targeting the Marketing Profession
Only this ad group was able to show at least some result in terms of performance – five conversions from 29 clicks.
Segment analysis shows that marketers are the audience most similar to the audience of listeners to eLama webinars. Interests in IT, finance and real estate came to the fore.
Ad group 4 – impressions by semantic core
As an experiment, I launched another ad group – on a keyword list. But he soon turned it off, since the keys can be used to launch regular contextual campaigns. In a few days, she gave only six clicks.
Remarketing in action
In the last step, I launched retargeting to the audience that saw the display ads. The plan as a whole worked – we managed to get an average cost per click of 8.31 rubles (which is several times cheaper than the cost of a click when paying by CPM). True, during testing there were only four clicks, so it’s too early to draw conclusions.
How to evaluate the media effect
Okay, getting conversions from display ads wasn’t easy. There is nothing surprising in this – when launching display ads, you should not count on clicks and traffic. It is assumed that the audience will constantly meet display ads, over time they will remember and begin to recognize the advertiser. Users who have seen ads more than once will begin to trust the seller and are more likely to become customers of the company in the future. There are several ways to track post-launch display ad performance:
- To assess the media effect, it would be logical to select two segments in Metric and compare the conversion rate by them:
- have seen display ads;
- have not seen display ads.
This is called post-view analysis. But neither Yandex.Metrica nor Yandex.Direct has such a report. Display ad professionals recommend using dedicated platforms like Adfox, which have post-view and post-click analytics tools.
- Another option is Brand Lift. To understand how effective the advertising was, two surveys of the target audience are conducted. One before the launch of the advertisement, the second after. The results are then compared to determine how brand awareness and interest have changed. Such an analysis can be carried out using the Yandex.Vzglyad service.
- The standard tools from the arsenal of a contextual advertising specialist can only approximate the effectiveness of display advertising. You can connect a Metrica counter to a segment collected using Pixel and set conversion goals. So you can find out the percentage of users from this segment who have achieved the goal. But it will not work to compare it directly with those who have not seen display ads.
In addition, display advertising is a game of long duration, ads can be shown to the user for a long time before they convert. And the data on visits is analyzed only for the last 90 days.
If in this screenshot you, like me, were confused by the low similarity of users, then there is an explanation for this. Interest in the creation and promotion of sites is only one of the many interests of this audience. In other interests, they may be completely different from each other. Therefore, you should not try to launch a look-alike on the audience gathered by Pixel.
The experiment lasted only a month. This time is not enough to judge the media effect, but certain conclusions can be drawn.
First, and purely practical: it is difficult to achieve uniformity of impressions within one campaign. In the first week of testing, almost all impressions in the campaign were received by the most reach group targeting ad interest. The rest of the groups were not shown, and for the second ad group to work, I had to disable the first, and so on.
Second and most important: from the transfer to Yandex.Direct, display advertising has not ceased to be media advertising and has retained its tasks and features. And turning it into a performance tool will not work outright. If you consider this type of campaigns as an additional opportunity to drive traffic and get conversions, but pay by CPM, then I have to disappoint you – this is not the case.