Employee

[Mentor Sree Charan Tekumanda] Adobe Real-Time CDP Community Mentorship Program 2024

Forum|Forum|1 year ago
July 15, 2024
15 replies
4201 views

Hello Team,

Welcome to the Adobe Real-Time CDP Community Mentorship Program 2024! This is the featured Community Discussion/Contextual thread for your Adobe Real-Time CDP Community Mentor, Sree Charan Tekumanda!

Sree will be your dedicated mentor, providing valuable support and guidance on your Adobe Real-Time CDP queries as you upskill yourself and prepare for Real-Time CDP certification throughout the program.

Know your Mentor Sree Charan Tekumanda (aka @sreecharan73)

Sree has extensive Data engineering and Modeling experience. He is a certified AWS Solution Architect, Databricks, and Adobe RTCDP Architect. He is an expert in strategizing digital applications and experiences.

He is looking forward to revolutionizing Customer Experience together with you all.

Aspirants mapped

1) Aditya Chabuku ( @aditya_chabuku )
2) Ryan Ku ( @ryanku )
3) Chris boehmer ( @cboehmer83 )
4) Hitesh Kumar ( @hiteshkumar98 )
5) Arya Rajendra Kumar ( @arya_rajendra_techm )
6) Nitin Indrajith Vishwakarma ( @nitinkarma )
7) Grandhi Lakshmi Pravallika ( @lakshmipravallika )
😎😎😎 Niharika Chhabra ( @niharikach-1 )
9) Chetan Gaikwad ( @chetanga )
10) Praveen Prabhakaran ( @wolflytics )
11) Rajendrasing Bhadoriya ( @rajendra168 )
12) Brandon Olvera ( @brandonolvera )
13) Austin Vaught ( @austinva1 )

How to participate in the program

Post your Questions in this thread to connect with your Mentor, Sree, and fellow Aspirant peers.
Stand a chance to win the ‘Most Engaging Aspirant’ recognition from your mentor by participating in a weekly quiz.
Test your knowledge by replying to the unresolved questions in the Real-Time CDP and AEP community and tag your Mentor to get recognized as an ‘Exceptional Contributor’ by your mentor.
Stick to the schedule to cover one module/week and clear Adobe Real-Time CDP Certification during the program: July 15 – Aug 30

Suggested Next Steps for Aspirants:

Update your Community Profile photo with your latest headshot to stand out to your Mentor and Peer Aspirants.
"Like" this thread to confirm your participation in the program.
Introduce yourself to Sree and your Aspirant peers by Replying to this Thread! Break the ice by introducing yourself (location, org/ company, etc.) and your experience with/ interest in Adobe DX stack.
Post your Questions to this thread as you begin learning more about the Adobe Real-Time CDP Business Practitioner Professional Certification (Exam ID: AD0-E602)
Stick to schedule and ensure you track your progress in the exam prep guide.
Test your learning by replying to weekly quiz by your mentor
Practice the modules by replying to unresolved queries in the AEP community & RTCDP Community and tag your mentor.

Remember that every post / like / comment you make in your contextual thread and the Real-time CDP Community throughout the program helps increase your chance to be recognized by your Mentor and win exclusive Adobe swag, so bring your best efforts!

We wish you all the best as you embark on this learning experience!

Real-Time Customer Data Platform

This post is no longer active and is closed to new replies. Need help? Start a new post to ask your question.

C

ChetanGa

New Participant

Hello @sreecharan73

Hope you are doing well. Want to get your thoughts on this & everybody else's as well

Is there a way to resolve an identity at a household level in AEP ?

Example:
An OTT subscription service being used by 3 members (Mom, Dad & a Teenage kid) in a household.
Each member would access the service using their own channels/device - Web, app & TV.

However, as a marketer, I want to be able to tie them to a household to understand patterns & reach out to them for specific communications.
E.g.
1) Promotional Messages: Reach out to all three for promotional messages messages thereby as per age & viewing patterns.
2) Servicing/Payment reminders: To mom & dad. If one doesn't respond, I would like the ability to remind the other one.

Can we use AEP to tie multiple people together. Currently, we only seem to be tying profile fragments together to build a profile for a single person.

Thanks,

sreeCharan73

New Participant

@chetanga Good question to follow up. Providing my views below:
Ideally, AEP is meant to stitch only fragments belonging to a single profile, hence it is always advised not to try marking any shared identifiers as people identifiers.
The recommended way to handle this is to have a Lookup(multi Dimensional entity) for the loyalty/HouseHold account info and mark this as a non-person identifier.
On the related Exp Event schema, setup a relationship to this lookup schema(based on the account ID) and then enable both the lookup and the Exp Event schema to enable advanced segmentation, thereby able to segment audience using the additional attributes present in the lookup schema.

In your usecase,
1.Promotional Messages: The promotional messages are anyway sent based on the individual behaviour and preferences, hence we can treat them seperately and personalise the communication. if only certain family accounts are considered, then use the above mentioned method to check, if the profiles belong to shortlisted account types/categories.

2. Servicing/Payment reminders: This may seem like special case, but can be handled using the same method. segment profiles based on the "expired/due to expire" dates of the accounts(lookup) and then you have all the related profiles in the Segment.
Now, based on the communication channel and application, we need to add the additional logic. Assuming that account info, also has some info on the primary contact and secondary contact stored in the lookup, but not marked as identity for profile stitching.
for ex:
using AJO, we have all the profiles entering the journey and then we could send the reminder to the primary contacts first in one journey. if there is no response, there can be a wait and condition check , so then to trigger to the secondary contact.

if using other FTP based destinations, like Adobe Campaign or MailChimp, export the previously built segmentation along with the Account ID, expiry date, primary contact and secondary contact flags and relevant profile info ( do not deduplicate based on the account ID, but deduplicate on profile email or mobile no). The orchestration should then be built on the relevant app, as similar to mentioned above.

Hope this provides some idea on to achieve your usecase.

C

ChetanGa

New Participant

Great, Thanks a lot @sreecharan73 !!

NimashaJainAuthor

Employee

Hope you all had a great learning experience. Here is your 50% waive off Voucher form!!

You have to submit your request here and you will have the chance of winning! Please fill out this form by August 19th, 2024

Best of luck

sreeCharan73

New Participant

Hello@AustinVa1 , @BrandonOlvera, @rajendra168, @wolflytics, @chetan01, @NiharikaCh, @LakshmiPravallika, @nitinkarma, @Arya_Rajendra_TechM, @Hiteshkumar98, @cboehmer83, @RyanKu, @Aditya_Chabuku

Hope you had a good learning over the week. Here is the weekly quiz 2.

I get answering the open ended might be time-consuming, so please respond with the choice, in a short reply.

W2 - Quiz

1. For Adobe Analytics/webSDK event data, what would be an ideal primary identity?

a. ECID

b. Cross Device Identifier(CRM ID/Customer ID)

c. Identity Map

d. Need more info

2. A data engineer tries to ingest data into a profile fragment with 25MB of data, what happens?

a. Since 50MB is the limit, Data is ingested, with no errors.

b. Data ingestion fails, as size exceeded.

c. Data can be ingested only up to 25MB, so gives a warning.

d. only 10MB data is stored and rest is truncated.

3. While setting up identity symbols, the Data architect needed an identifier for the Product schema. what type would you suggest?

a. Cross Device Identifier

b. Primary Identifier

c. Non Person Identifier

d. Device Identifier

Looking forward for your responses. Please provide feedback, on the questions, if you feel any.

R

RyanKu

New Participant

1. For Adobe Analytics/webSDK event data, what would be an ideal primary identity?

a. ECID

b. Cross Device Identifier(CRM ID/Customer ID)

c. Identity Map

d. Need more info

in general the ideal primary identifier is an identitier that represent a person, such as CRM ID/Customer.For WebSDK / AA an identity map would be useful and the PK would change depending on scenario: the person identifier becomes the primary identity (authenicated scenario) if available, otherwise ECID becomes the primary identity (unauthenicated scenario)

2. A data engineer tries to ingest data into a profile fragment with 25MB of data, what happens?

Based on Default Guardrails for Real-Time Customer Profile Data and segmentation | Adobe Experience Platform

a. Since 50MB is the limit, Data is ingested, with no errors.

3. While setting up identity symbols, the Data architect needed an identifier for the Product schema. what type would you suggest?

Based on Identity Namespace Overview | Adobe Experience Platform

c. Non Person Identifier

sreeCharan73

New Participant

Hello

@AustinVa1 , @BrandonOlvera, @rajendra168, @wolflytics, @chetan01, @NiharikaCh, @LakshmiPravallika, @nitinkarma, @Arya_Rajendra_TechM, @Hiteshkumar98, @cboehmer83, @RyanKu, @Aditya_Chabuku

hope that you all went through the basics of Data Architecture and RTCDP overview & Profile Ingestion, in the last weeks. Since we are in the 2nd half of the program, lets switch to below topics(these cover 52% of the exam questions, so you now know the importance) refer https://experienceleague.adobe.com/en/docs/certification/program/technical-certifications/rtcdp/rtcdp-professional/rtcdp-p-business#:~:text=Exam%20objectives%20and%20scope

1. Segmentation & Activation

2. Data Governance

sreeCharan73

New Participant

Hello@AustinVa1 , @BrandonOlvera, @rajendra168, @wolflytics, @chetan01, @NiharikaCh, @LakshmiPravallika, @nitinkarma, @Arya_Rajendra_TechM, @Hiteshkumar98, @cboehmer83, @RyanKu, @Aditya_Chabuku

Hope you had a good learning over the week. Apologies, I could not post the weekly quiz earlier.

Here is the weekly quiz, based on the last week modules.

W1-Quiz:

1. What is a profile fragment? How do Identities would help stitch fragments?

2. How can one choose between Batch Ingestion vs Streaming Ingestion? What are the advantages?

3. If using a customer segment(Audience) based on the hashed Email IDs, what type of destination(export) is the ideal? a. file based b. streaming segment c. streaming Profile export.

Although the answers are open ended, try to answer in 1-2 lines for each, basically like a log line to summarise the understanding of the topic. I request everyone to respond. Since Data modelling in AEP is trivial, but gets better with collaboration. Pls do not hesitate, we are all in this together.

Regards

Sree

C

ChetanGa

New Participant

Here are my responses, please do correct as necessary:

1. What is a profile fragment? How do Identities would help stitch fragments?
Profile fragment is a set of data ingested in a dataset. This data should have at least one identifier (either primary or non primary) - for it to play a role in stitching identities across namespaces.

2. How can one choose between Batch Ingestion vs Streaming Ingestion? What are the advantages?

Batch ingestion is a scheduled ingestion & happens at a certain frequency using files; Streaming ingestion is more real time & leverages HTTP APIs/streaming services.

3. If using a customer segment(Audience) based on the hashed Email IDs, what type of destination(export) is the ideal? a. file based b. streaming segment c. streaming Profile export.
I am guessing, this may be b) streaming segment since we have hashed email IDs. Since, this is a segment/audience & not a profile- it should not be streaming profile export. Not sure, of the exact differences though. Would love to get a different view on looking at batch/api based destinations.

Thanks

sreeCharan73

New Participant

Thanks for taking time to answer, so quickly.

All answers are good and correct.

1. Profile fragment must have a primary identity and it is corresponding to a dataset in RTCDP. Identity Service and Real-Time Customer Profile | Adobe Experience Platform

2. Correct. Although Batch can also be used for Database sources and CRM, while streaming ingestion can also be extended to AWS Kinesis, Azure Event Hub & Kafka based pub-sub models.

3. Correct, streaming segment export would be an ideal candidate, Meta and other social destinations are good examples for this.

C

ChetanGa

New Participant

Hello @sreecharan73
Hope you are doing well !

While going through the XDM schemas & Datasets, some of the aspects are not 100% clear to me.

I am looking to unify a profile for a simple use case for a beauty products retailer as given below.
I am looking at creating schemas & datasets, creating basic relationships (similar to PK, FK in RDBMS). The said retailer has web & physical stores.

To keep things simple:
I am considering a Web schema & an associated dataset. This dataset would hold data when consumers register & update their profile.
Schema 1: For Web
Below is the schema:
Cust_ID | Phone_number | Email_Address | First_Name | Last_Name | Gender | Dob

Cust_ID -> primary identifier generating from a backend CRM system.
Phone_number -> Non primary identifier
Email_Address -> Non primary identifier
This schema is in a way the main schema.

Schema 2: For Store
Another schema & dataset is for a store dataset to ingest feed coming from the store.
Two key aspects here:
1) No Cust_ID since store systems are not tied to CRM
2) Customers may chose to give either phone number or email address depending on their preference.

Hence, the data would come in the following manner.
Case A) Feed using Phone number as a primary key.
Phone_number | First_Name | Last_Name | Gender | Skin Type | Products Purchased

Case B) Feed using Email Address as a primary key.
Email_Address | First_Name | Last_Name | Gender | Skin Type | Products Purchased

My Ask:
Q1) Unify web & Store: Connect the web schema & store schemas together using either email address or phone number for a customer as and when the data comes in. How would one approach this ? I want to ultimately use skin type, products purchased from store in the segmentation.
Q2) Schema Relationships: Does the answer lie in building relationships between the schemas ? What use would the relationships serve given that (Email, Phone number are identifiers) and are present in Web Schema.
Q3) Priority of updates: I want to give preference to web, app data over store data. Since, store data may have typos & common errors (e.g. Name is misspelt in store data). I believe, this would be related to merge policies?
Q4) Enabling Primary Identifier: Do I need to enable the store dataset to profile & give a primary identifier ? Would email & phone work as identifiers without use of Customer ID coming in from store.

Thank you so much !! Just lot of questions currently.

sreeCharan73

New Participant

Hello @chetanga

Thank you for asking this one. This focuses on the backbone of AEP, the Identity Service.

1. Unify web & store:
This is the primary advantage of AEP as RTCDP, we should have the right fields marked as identity. Since the PK of the records are not consistent, we should approach this using identityMap as the Primary Identity and have the email and phone as secondary identities on this schema.

2. Schema relationships:

this is basically for non people Entities like product(here as example) and other lookup (multi dimensional) entities. This is a very good feature which enables advanced segmentation, but should be dealt with care.

3. Priority of updates:

Yes, you are correct. Merge policy is indeed the answer. Create a merge policy of type "Dataset precedence" and use this merge policy for all segmentation and activation needs.

4. Enabling Primary Identifier:

Yes, primary identifiers are tied to the relevant schema, so we could have phone or email as primary identifier on the store schema, however since the primary identifier cannot be null, if the values are only available inconsistently (either email or phone) then as I mentioned above, try having the IdentityMap as the primary identifier and utilise Data prep to map them, as it comes.

Please check my post below for the links to learn more about this.

Regards

Sree

R

RyanKu

New Participant

Thanks @nimashajain for the introduction. Nice to meet everyone and thank you for being our mentor @sreecharan73 !

Ryan Ku, Sr Manager Digital MarTech at Rogers in Toronto, Canada. I have limited experience (less than a year) with the Adobe Experience Cloud products but am jumping in to learn the new Adobe Experience Platform suite. Looking to learn as much as I can to help our organization succeed in personalization across web and mobile app and optimize performance for campaigns / paid media.

sreeCharan73

New Participant

Hello @ryanku

As you might've understood, WebSDK, CJA, Target and AJO are very sophisticated tools to help create omni-channel campaigns, delivering consistent and personalised experiences to the end customers. Since, AEP is the sole data platform that boosts these apps, makes it a must-know.

Hope this activity helps to get you into the game of AEP.

Regards

Sree

sreeCharan73

New Participant

Hello

@AustinVa1 , @BrandonOlvera, @rajendra168, @wolflytics, @chetan01, @NiharikaCh, @LakshmiPravallika, @nitinkarma, @Arya_Rajendra_TechM, @Hiteshkumar98, @cboehmer83, @RyanKu, @Aditya_Chabuku

Below is the section-wise details about the AEP RTCDP Business Practitioner exam:
Professional certification | Adobe Real-Time Customer Data Platform

Since most of the team are already having the foundational understanding of AEP RTCDP.

The Experience League documentation is a huge ocean of knowledge on the topics, however lets focus on the important topics required for a good start.

Let's dive into the topics, with related links to the Adobe documentation.

AEP Data Architecture:

Stay tuned for some trivia, starting this weekend.

C

ChetanGa

New Participant

I was going through the key concepts - schemas, datasets & so on.
As I understand, schemas are the foundation and then one has to create datasets to ingest data.
Now, there are 2 approaches to create a dataset:
1. Using XDM Schema
2. Using CSV File.
In case of 2 above, it seems we don't need a schema in place to create a dataset & can directly create a dataset.

My questions:
Question1: Wanted to understand, what would be the use case/situation when one could not create a schema & directly create a dataset from CSV (i.e. go as per approach 2).

Question2: Is it a best practice to create a schema first & then create dataset ? Is there a preference or recommended approach.

Ref: https://experienceleague.adobe.com/en/docs/experience-platform/ingestion/tutorials/ingest-batch-data

NimashaJainAuthor

Employee

@chetanga Awesome Question

@sreecharan73 Can you look into the query?

C

ChetanGa

New Participant

Hi All
I am Chetan, I work at Epsilon & based out of Bangalore. I have been working on different products of Adobe Marketing Cloud since quite some time. Had a long standing wish to understand AEP & RT CDP in detail.

Nice to meet you all.
Cheers,

sreeCharan73

New Participant

Hello Chetan

Good to have the "Adobe Experience", makes it a great start for learning "the Platform". Since AEP is built around and to be the real-time data warehouse for all Exp. Cloud applications & much more. Let's talk more.

Show more replies

Sign up

Login with SSO

Login to the community

Login with SSO

Scanning file for viruses.

This file cannot be downloaded