Adobe Target Experts on Unlocking AI-Powered Personalization

@gokul_agiwalGreat questions!  Here are my responses:

1) Adobe Target's AI-powered capabilities, especially Auto-Target and Automated Personalization that evaluates in real-time each individual's profile (or a subset of profile data), utilizes all data made available to Adobe Target, so it automatically evaluates your customer data if brought in through API, Profile Core Services, or as a Destination for Real-Time CDP.

2) Best Practices are well-documented in our Adobe Target documentation.  Nurturing of these campaigns depends on if you are looking for quick lift/performance improvement and feel confident in the experience variation choices or if you want to take time to evaluate the performance of the variations built relative to the conversion goal that the model is training on.  This will dictate traffic allocation as you ramp up the algorithm to full performance.  Also, these activities are very dependent on the location, experience variations and goal of the activity in terms of their performance as well - strategy is key based on your business goals!

3) Adobe.com's homepage is an Auto-Target activity and a good example of the ability to create a fully dynamic lobby experience while reporting on metrics beyond just conversions, including page load performance and many other important KPIs.  These activities are automatically tested against a default experience, either random selection or a selected default experience; it is best practice to always test any strategy, manual or automated personalization, to ensure it is improving the experience over a static or default original experience and we provide this OOTB.

Indeed it is and thank you @niocolasmeriel for joining! What burning questions do you have for today's hosts? CC @rajneesh_gautam_ @3068898 and @shelbygoff 🙂

@amelia_waliany @coops I agree that the human can determine the audiences evaluated within an Auto-Target activity, as it evaluates audiences from your audience library, but the random forest ensemble also assembles its own audiences that it discovers as predictive during the course of its evaluation (up to 4 attribute combinations) which it provides in our Target Segment Discovery report.

You can use blockList API (https://experienceleague.adobe.com/en/docs/target-dev/developer/api/models-api/models-api#_blank) however its not available within the Adobe Target UI. I believe its a really good idea to raise under the Ideas section of Experience League

Adobe Target uses multiple data points collected by Adobe Target, Adobe Experience Cloud tools (AAM, RTCDP) and external data uploaded  via Customer Attributes and RTCDP. The page below provides a comprehensive list of all data-points used: https://experienceleague.adobe.com/en/docs/target/using/activities/automated-personalization/ap-data

You can ask Adobe Target to exclude some of the attributes/features from personalisation-models using the blockList API: https://experienceleague.adobe.com/en/docs/target-dev/developer/api/models-api/models-api#_blank  

You can block specific attributes/feature or entire feature-sources e.g. AAM, GEO but not the specific segments.

@niocolasmerielI will need to dig into this!  We do have the ability to ingest a custom model with Auto-Target, but it is not an out-of-the-box feature so I will need to dig into how this is done.  Stay tuned!