Jul 31, 2022
In Welcome to the Forum
The end of conversion lift studies? Essential for doing conversion lift testing is being able to measure conversions of the two groups: Your test group (users who have seen your ad), and your control group (users who haven't seen your ad). But with the developments of the cookie-free world , it is becoming increasingly difficult for marketing platforms to track users including data. For example, iOS 14.5 has made it impossible to do conversion lift studies for advertisers on Facebook. We're not sure what the future holds for the rest of the marketing channels. However, it is certain that the disappearance of third-party cookies jeopardizes the correct measurement of conversion lift studies. Also read: The 6 hidden gems of Google Ads Testing incremental value yourself with Geo-Experiments You can also do a conversion lift without the tooling of marketing channels, even without cookies. An alternative is testing by regions. This is called a fax number list geo-experiment. A geo-experiment is basically simple. You pick 2 very similar regions and assign them both a group. In your test region, you show users an ad, but not in your control region. You make a prediction in advance how many conversions both groups should get, and how much you should get extra from the marketing activity. The difference in the data of the regions is your incremental value. Although the results of this method are not as accurate as those of a conversion lift of the marketing channels themselves, geo-experiments are future-proof as they do not use data at the user level. A nice alternative for the future. How do you use the results? Just because a marketing channel has incremental value doesn't mean you should advertise on it. To get a better idea of the value, you can look at the cost per incremental conversion or at the incremental value / cost (ROAS or POAS). When you have determined the value, you can trade on it. If some of your campaigns have less value, for example retargeting compared to prospecting, or branded compared to non-branded search, you can decide to bid differently.