We have a (hopefully not unique) use-case. We fire a unique event for some users using the Google Analytics 4 Measurement Protocol (server-side events), and want to target these users.
Our approach:
1. We create an audience on Google Analytics 4 using the (Alpha) Admin API.
2. We fire some events from our server using Google Analytics 4 Measurement Protocol. These events contain user data like Email ID's, as well as other auxiliary info like Country Code, Region, City, etc. These are sent correctly using the user_data parameter in GA4 Measurement Protocol.
3. Our Google Analytics property is linked with our Google Ads account, so any audiences created on Google Analytics automatically show up on Google Ads.
Expectations:
1. The Google Ads audience should contain the same number of users as we see on the Google Analytics audience.
Actual result: The Google Ads audience is significantly smaller than the Google Analytics audience.
2. The Google Ads audience has equal sizes in Gmail, Display, Video, and Search segments.
Actual result: The Display populated first for us, after about 3-4 days of firing events. The Search and Video segments populated much later (3-4 more days), and are of much smaller sizes.
To test interference, we created another audience targeting the same group. We found that this new audience, that should be identical to the first audience populated first for Search and Video, and much later for Display. The Display segment is also much smaller here, and Search and Video are much bigger.
In both cases, Gmail fails to populate.
Can we get some clarity around this?
1. What can we do to increase match rates?
2. What's the situation around the different segments? Is this something arbitrary or are there some patterns/guarantees here?
Thanks,
Rohit.