So, when trying the judge whether SKAN 4.0 is an improvement over SKAN 3.0 for app-to-app campaign performance measurement, it comes down to the question how new privacy thresholds relate to the old privacy threshold.
In 4.0 there are at least two thresholds: (a) low-end to medium-end crowd anonymity threshold, and (b) medium-end to high-end crowd anonymity threshold.
Questions:
(1) The first question is: When the 1st postback (the one after 2 days) contains the fine-grained ConversionValue, does it also contain the 4-digit SourceID, and vice versa? Or do ConversionValue and SourceID have separate privacy thresholds? Or is it even a combination of both? (And the same question goes for coarse ConversionValue and 3-digit SourceID.) I would imagine the SourceID has a higher privacy threshold than the ConversionValue.
(2) The second question is: Can we imagine the medium-end to high-end crowd anonymity threshold to be in the area where the SKAN 3.0 ConversionValue privacy threshold is? In other words would I get the fine-grained ConversionValue when about as many installs are reached as were necessary to get the ConversionValue in SKAN 3.0 (which was essentially a fine-grained ConversionValue)? Otherwise it would be quite a step back, right?