Lots of users do not click on YouTube ads right away and visit the site later and directly. We developed a technology that was able to incorporate post-view conversions in the results, and set up optimization for in-game events (First Login, Second Login, time spent in the game). This helped us double our YouTube conversion rate and highlight our top users.

In 2021, we carried out several campaigns to attract new users to EVE Online. Initially, we performed optimization of registration, but quickly realized that we were more interested in in-game events that better correlate with the future profitability of users.

One of the first solutions was to use in-game events such as First Login, Second Login, and time spent in the game. Later we added these factors to the prediction of future income.

In cases where we take events from a game (which is located on the user's computer or laptop and is not connected to the browser) we can only use the User ID to determine conversions.

We were able to link this data to clicks, including Google Click ID. This made it possible to optimize YouTube campaigns for in-game events, but only for those users who clicked on the ad right away.

However, lately we have seen a big trend: many users do not click on ads immediately, but first watch the ad, then watch the video after the ad. If they were interested in the advert they would then go to the site (either directly or through a search), and then check out and download the game.

Using different models, we have seen that the effect of YouTube ads is greater than what we capture by clicks. We wanted to get more accurate knowledge of in-game conversions, including those users who came after viewing ads without clicking them.

We wanted to see the entire in-game funnel and, more importantly, use this data to optimize our ads, because in the case of post-click only users, the ad system failed to consider the users who went to download the game after watching the ad (because it did not know that they had converted).

We have developed a technology that can take into account post-view conversions, including in-game conversions, and automatically optimize advertising campaigns based on such data.

What is the novelty? #

We launched campaigns to attract new users using the GMP platform (DV360) with advertising on YouTube and optimization, including post-view conversions based on funnels from in-game events.

Today, almost no one knows how to upload in-game offline conversions to GMP (Campaign Manager 360), and therefore they usually make decisions based on either raw data on a piece of paper or only through external game events.

We have done what no one has done before us. We have created a new user ID architecture that allows capturing all users who watched a commercial on YouTube. Typically, in such campaigns, the only users who are recorded are those who watch the advertisement, then immediately follow the link and download the game (post-click). We have learned to see those who watch the video, but do not follow the link, and then go and download the game themselves after a period of time (post-view). We use this data to optimize our advertising campaigns.

What is the result? #

The first thing we saw was post-view registrations accounted for about 30% of the total volume.

But, more importantly, users who came after viewing ads had a conversion rate 2 times higher after their first accessing the game, than those who came after clicking.

This is probably due to the fact that such users are more motivated, since they remembered the ad they watched and downloaded the game for themselves.

Thus, using data on post-view conversions, we were able to see via the advertising system who are the very best users, those that we have been hunting for all along.

Finally, we managed not only to increase the conversion by 2 times, but also to significantly improve the Predictive ROAS indicators, that is, to bring much more future income to the business.

For whom is this relevant? #

The described technology for tracking post view conversions can be relevant for all game creators who work through a separate desktop program, and not via the web or apps. Additionally, it can be useful for creators and distributors of offline programs, e-commerce companies (for transferring data on redeemed orders from CRM) and banks (who are willing to pay for those who opened an account or card but not for applications).

Results #

• Post-view registrations were about 30% of the total volume, and we were able to start counting them;
• But users who came after viewing ads had a conversion rate 2x times higher to the First Login into the game, than those who came after clicking;
• More importantly, post-view users had a 3x higher conversion rate to the Second Login
• We improved Predictive ROAS variables;
• We built tracking for post-view events from the game for display and video ads from Reddit by integrating them with CM360;
• 30% of the total regs were Post-View registrations.