Synthetic User – The future of online marketing and analytics

Yes, absolutely, consent optimisation remains necessary. Users with consent are necessary to identify predictors for users who do not give consent. We cannot identify predictors and perform sampling with a 0% consent rate.

We don’t know precisely at the moment. Our lowest example was 27%, which has already worked great. We are currently trying to find the bottom limit.

The entire legal basis can be seen in the memorandum. The memorandum refers to “strictly necessary”. I.e., it is necessary for you to generate synthetic users and for this we collect these few parameters in the non-consent area also with a user recognition in the session area, however, the data is processed in the name of the customer, i.e. the data remains first-party data.

No, because collecting the data is allowed by the memorandum to generate synthetic users.

We expect a significant impact on sales, as users are activated who were not activated before and the data quality is massively improved. On the other hand, all marketing today is based on third-party cookies, with only 15-20% of activatable users. With JENTIS, we see a data improvement of about 60% in the A/B tests during onboarding. You are also welcome to view the JENTIS and Pixum case study here.

We can provide the use of Synthetic Users only if JENTIS is installed. Otherwise, there are four classes of slots for the pilots. One is exclusively for data quality testing; there are no additional tools here. The other three slots are performance marketing, social media marketing, and affiliate marketing – here, of course, the A/B test must run in an appropriate tool.

This is an advantage of the core technology of JENTIS. We now have over 125 connectors, and we can create new connectors quickly because we use twin-server technology, i.e. when passing data to third-party tools, we do not use APIs but the same technology that is used on the browser for tracking – just on the server.

We do not capture the data but the predictors. If the predictors include “number of page views in a session”, for example, then the session and the page views would also be recorded. When the non-consent user is assigned to a segment, the data is eliminated again. After the assignment to the segment, the user object is empty. So nothing is overwritten because nothing was measured but refilled and calculated.