Expand my Community achievements bar.

☕[AT Community Q&A Coffee Break] TUESDAY 11/29/22, 8am PT: AI-Driven Personalization & Recommendations with Cristinel Anastasoaie, Senior Product Manager for Adobe Target☕

Avatar

Administrator

Join us for the next Adobe Target Community Q&A Coffee Break 

taking place Tuesday, November 29th, 2022 @ 8am PT - 9am PT 

REGISTER TODAY!

We'll be joined by Cristinel Anastasoaie aka @cristinel, Senior Product Manager for Adobe Target, who will both be signed in here to the Adobe Target Community to chat directly with you on this thread about your questions related to the recent 9/15/22 Adobe Target Skill Builder Expedition series Webinar on AI-Driven Personalization and Recommendations and to continue the conversation! 

Be sure to review the Webinar Recording and then bring any and ALL of your follow-up questions here to the thread on Tuesday, November 29th at 8am PT for comprehensive insights, deeper explanations, and strategic suggestions for your unique use cases ~ DIRECTLY from the Adobe Target experts 

• Optimize real-time recommendations, one-click personalization activities, and automated offers
• Create and launch AI-driven personalization activities using Visual Experience Composer (VEC)
• Integration with Adobe Experience Platform (AEP) for real-time segment and profile attribute sharing

Adobe Technical Support Engineer Simon Chen aka @simochen will also be joining to provide additional insights and answers to your questions.

REGISTER TODAY!

REQUIREMENTS TO PARTICIPATE 

  • Must be signed in to the Community during the 1-hour period 
  • Must post a Question about Adobe Target
  • THAT'S IT!  *(think of this as the Adobe Target Community equivalent of an AMA, (“Ask Me Anything”), and bring your best speed-typing game) 

INSTRUCTIONS 

  • Click the blue “Reply” button at the bottom right corner of this post
  • Begin your Question with Cristinel's Community Username: @cristinel

 

Amelia_Waliany_1-1667428041357.png

 

 

 

Cristinel is a Senior Product Manager for Adobe Target, focused on machine learning personalization in Adobe Target, Analytics for Target integration, and the product user interface. Before his current Adobe role, Cristinel held product management positions on various content management and digital marketing products focused on consumers.

 

 

 

 

 

Curious about what an Adobe Target Community Q&A Coffee Break looks like? Be sure to check out the thread from our latest 10/26/22 Adobe Target Coffee Break with Adobe Target Expert Drew Burns aka @Drburns27.

Topics

Topics help categorize Community content and increase your ability to discover relevant content.

17 Replies

Avatar

Level 1

@cristinel is there a document with out of the box profile scripts?

Avatar

Administrator

Hi @cristinel ! Thanks for joining to answer Community Qs today, and thanks @Studley2021 for kicking off the chat!! 

 

This question was posted by mharper5:

"I am looking for a log that can tell me the last time a feed was successfully loaded. I have one that has been stuck on Importing Items for a while."

Original Question: Recommendations Feed Log?

 

Avatar

Employee

Unfortunately Target does not have a log for showing when a feed import has failed. However, you can subscribe to a notification feed which can send UI, push or email messages when a feed import has failed. See this Experience League document for more details: https://experienceleague.adobe.com/docs/target/using/release-notes/system-status-updates.html?lang=e... 

Avatar

Administrator

@cristinel , this Question was posted by KMarchewa:

"In their documentation, Adobe suggests the following guideline for determining sufficient traffic when a a Conversion metric is used for Auto-Target

"1,000 visits and at least 50 conversions per day per experience, and in addition the activity must have at least 7,000 visits and 350 conversions.:

My question is, in general, if these requirements are met, how long should it take the models to build? More generally, there is a calculator for Auto-Personalization Activities - can one use this to estimate a sufficient duration for Auto-Target or is there an alternative calculator that will work? Thanks"

 

Original Question: Auto-Target - Model Building and Activity Duration

Avatar

Employee

If the traffic and conversion thresholds are met, first models are built within 24 hours and will keep learning and adapt to new traffic moving forward.

 

Automated Personalization and Auto Target are pretty similar, with the the note that AP usually requires more traffic given that it builds experiences based on the offers specified in the activity.

Avatar

Administrator

Hi @cristinel , this Question was posted by 

 

"I wonder if there is a minimum of traffic or conversion needed for Recommendations to work well?

For Recommendations to work effectively/develop its algorithm is there any guideline or documentation specifically looking at the sales/conversion volumes required for the machine learning to work?

Thanks,

Nicolas "

Original Question: Is there a minimum of traffic/conversion needed for Recommendations?

 

Avatar

Employee

Recommendation minimum traffic is hard to defined as it can be influenced by the a number of factors such as: algorithm, filtering and exclusion rules, number of items in catalog. Some algorithms (like popularity based ones) require a low volume of traffic in order to start generating recommendations. Others, like user based and cart based require some traffic to start learning visitor behavior and be able to give recommendations. You can find more information on Adobe Target recommendation algorithms on this page: https://experienceleague.adobe.com/docs/target/using/recommendations/criteria/create-new-algorithm.h...

Avatar

Administrator

Hi @cristinel this Question was posted by mysticDeveloper: 

"I want to create a "most popular in your favourite brand" recommendation, which with the new update it should be through the "Popularity-based" recs. 

What does the field "Profile Attribute to Match" do on the new "Popularity based" recommendations? 

 

Amelia_Waliany_0-1669739268941.png

Not sure why Adobe released an update without also making a proper documentation at the same time. Currently the documentation says "(Info coming soon)" for Top Sellers by Item Attribute and Top Sellers by Item Attribute and Most Viewed by Item Attribute.  https://experienceleague.adobe.com/docs/target/using/recommendations/criteria/base-the-recommendatio..."

 

Original Question: Popularity based recommendations. What is "Profile Attribute to Match"?

Avatar

Employee

Thank you for raising this. Similar to the "Most viewed by item attribute" algorithm, the Top sellers by Item attribute let you select which item attribute you want the recommendation to be based on. For example, you can specify which profile attribute stored in the visitor profile to match.

 

We will update the documentation for Top seller items as well.

Avatar

Level 1

@cristinel How can I gain insights as to how Adobe AI-Driven Personalization & Recommendations features & capabilities stack-up to key competitors in the market? Is there any comparison documentation that can be referenced?

Avatar

Employee

Adobe Target is a testing and personalization tool recognized as a leader in by various industry reports. Here is some more information on how Target can contribute to your experimentation and personalization programs: 

Avatar

Level 2

@cristinel 

 

Thanks for taking the time to answer some questions today. If you don't mind, I have some questions unrelated to the AI-driven personalization aspect of Adobe Target. If not, no worries.

 

Our org has been looking at using on device decisioning. We're excited by the simplicity and performance that's suggested in its documentation. However, most (if not all) of our use of Target is for AB tests w/ mutually exclusive test groups, and on-device decisioning doesn't (yet?) support Visitor Profile audiences. Is there any possibility that it ever will? If not, are there any suggestions on how we could achieve conducting A/B tests across mutually exclusive groups with on-device decisioning?

 

Unrelated, we are also seeing odd behaviors with our analytics reports. KPIs and other metrics continue to shift for groups in a test we've *after* deactivated tests, sometimes days after. We're also seeing events attributed to users in mbox assignments where it should be impossible for them to see these events (e.g., a user clicked a button that wasn't shown for the experience that Analytics claims they were given). Any thoughts on how this could happen or we could troubleshoot these issues?

 

Again, I hope these are not too off-topic for this venue and thank you for your time in responding to these questions.

Avatar

Employee

@seeplusplus  - thank you for your questions.

 

Unfortunately we do not have any plans to add support for visitor profile audiences in On-Device decisioning, but I would like to learn more about how you are differentiate your user groups.

 

As for the unrelated question, there might be multiple reasons for that. Some of them could lie in the difference on how Target and Analytics attribution work, with Analytics having longer attribution times. It would also depend on the activity type as some activities are optimizing for visitor while other per visit.

Avatar

Level 2

Thanks for your reply. As for how we differentiate our user groups, we are using a setup similar to this guide:

 

https://experienceleague.adobe.com/docs/target/using/audiences/visitor-profiles/use-profile-scripts-...

 

Re: Activity types. All of our activities are A/B tests w/ form based config. (Not sure if that answers your question). Are there any resources where I could read more about Target/Analytics attribution? I'm fairly fresh in the process of looking into all of this, and any further reading would be great!

 

Thanks again!

Avatar

Community Advisor

@Amelia_Waliany @cristinel 

Is there a way to export all the audience segments from Adobe Target? Also, will it be possible to see the Target activities in which the audience segments are used?

Please guide me to the reference documentation or how-to videos, if there's a procedure to achieve it. In case it is possible through APIs, that would be helpful too!