Currently there are two methods to integrate A4T, either via the sdid approach or the data insertion api.
- The sdid approach is very limited as it has quirks related to first-time visitors. There is also a complete absence of technical guidance or documentation on how to generate an SDID if you are using server-side target on a modern platform that does not support one of the Target SDKs.
- This is particularly important as the most substantial/performant types of experiments/activities are likely to be run server-side, but server-side analytics lacks feature parity with client-side.
- The data insertion api approach is very expensive, as it increments an additional server call each time a user has an activity response from Target. It also introduces significant/unacceptable sequencing issues; either the A4T event arrives before the analytics event, causing misstatements of values related to things like marketing channels or it arrives after the page level hit for which the user actually qualified for the experiment, which requires more complex segmentation.
A less expensive and problematic approach would be to enable users to pass the tnta value via context data or an eVar and then map that dimension to tnta within processing rules. This would introduce meaningful flexibility, avoid unnecessary server calls (which is mutually beneficial for clients and Adobe) and mitigate any sequencing issues.
Why is this feature important to you -
This feature is important as the current A4T implementation choices are limited AND analytics can only support 3 listVars. To be able to analyze users in multiple experiments we need access to the tnta dimensions which behave as listVars.
How would you like the feature to work -
The tnta dimension would be an option within processing rules. Users would have the ability to map either existing dimensions or context data values to the tnta dimension.
Current Behaviour -