Adobe Experience Platform [AEP] Training - Tutorial 1 - Introduction Online Demo Session - myTectra

Summary

Highlights

• Customer Data Platforms (CDPs) solve fragmented data issues by centralizing information hubs. 🗂️
• Adobe Experience Platform (AEP) uses deterministic matching for precise customer data profiling. 🔍
• Probabilistic matching in CDPs offers a statistical approach to merge data based on likelihood. ⚙️
• The importance of segmentation in targeting specific customer groups with personalized offers. 🎁
• Data governance in AEP ensures compliance with privacy regulations and protects customer information. 🔒

Key Takeaways

• Learn about Customer Data Platforms (CDPs) and their role in data centralization to improve customer targeting. 🎯
• Understand Adobe Experience Platform's (AEP) infrastructure and how it supports data collection and customer profiling. 🏗️
• Explore how CDPs use deterministic and probabilistic matching for comprehensive data integration. 🔄
• Get insights into creating detailed customer profiles for personalized marketing campaigns. 📊
• Discover the various data sources and integration methods supported by AEP, including batch and streaming data. 🌐

Overview

Chapters

• 00:00 - 00:30: Introduction to Customer Data Platforms (CDPs) This chapter introduces the concept of Customer Data Platforms (CDPs) and highlights their importance in the current digital landscape. It outlines a structured approach to understanding CDPs by starting with the basics and then exploring the different types available in the market. The chapter plans to provide an overview of the Adobe Experience Platform, including its architecture and the process of configuring sources and understanding schemas.
• 00:30 - 05:30: Understanding CDPs This chapter introduces the concept of Customer Data Platforms (CDPs). It begins by explaining the challenges businesses face in targeting customers accurately. These challenges arise because customer data is often dispersed across multiple systems such as email, social media, sales databases, and other sources. A CDP consolidates this scattered information, providing a unified view of the customer, which can significantly aid businesses in better targeting and understanding their customers.
• 05:30 - 11:00: Types of Customer Data Platforms The chapter discusses the issue of fragmented customer data being stored in various places, making it difficult for businesses to obtain a comprehensive view of their customers. This fragmentation complicates sending appropriate messages and avoiding redundant efforts. The example of a textile business, such as Max or Lifestyle, is used to illustrate the problem of targeting customers when their data is inconsistent and incomplete across different platforms.
• 11:00 - 19:00: Adobe Experience Platform Architecture In this chapter titled 'Adobe Experience Platform Architecture,' the discussion revolves around the challenges posed by multiple and inconsistent data sources in the context of using an application. For example, users might enter different email addresses across platforms like surveys and apps, leading to discrepancies in the data collected. The chapter highlights the difficulty in identifying the correct data source when there are conflicting entries, such as different emails being saved across various platforms, emphasizing the need for a unified and accurate data management strategy.
• 19:00 - 25:00: Data Sources and Integration The chapter discusses the need for businesses to understand customer preferences and behavior patterns to target them with the right offers at the right time. It emphasizes that a clear picture of customer data is essential for businesses to act effectively. Customer Data Platforms (CDPs) are introduced as a solution to handle fragmented customer data, enabling businesses to create targeted marketing strategies.
• 25:00 - 44:00: Schema and Field Groups in CDPs The chapter discusses how Customer Data Platforms (CDPs) solve data dilemmas by aggregating all customer data into a centralized hub. Data from multiple sources is cleaned and organized into unified customer profiles, integrating both online and offline interactions to enhance business operations.

Adobe Experience Platform [AEP] Training - Tutorial 1 - Introduction Online Demo Session - myTectra Transcription

• 00:00 - 00:30 sir today we'll start delving into concept of what are this customer data platforms or we'll understand what are they and how they're important in today's digital land landscape so once we grasp the basics about it we'll explore the various types of ctps available in the market uh next uh we'll get to know the overview of adobe experience platform then we dive into its architecture and then uh we'll explore the process of configuring sources and understanding the schemas in AP so let's get started
• 00:30 - 01:00 so first I would like to uh make you understand what exactly the CDP is uh so CDP is an acronym of this customer data platforms so the paino in general businesses are for example uh lots of businesses have a big problem targeting their customers correctly because they don't have all the information about the customers in one place so this happens because customer data is spread out across different systems like email social media sales databases a point of
• 01:00 - 01:30 sale many right because the customer data is not really stored in one place the data is present in multiple places so businesses struggle to get a complete picture of each customer so this makes it hard for them to send the right messages and avoid repeating their efforts uh for example I'm running a business maybe I'm running a textile business something like like a Max or a lifestyle so I want to Target you uh you are one of my customer but your data was very very partial and it's very very different in each medium for example if
• 01:30 - 02:00 you're using App you'll give some information in the point of sale you give your misspell your name some way or maybe something like that in surveys maybe you'll feel different way so uh the information of we know they have multiple sources but I don't know which source is correct which email is correct for example you have two different emails in one in survey one in app somewhere else so the data sources is are very much vast so what exactly the paino here is I want to exactly ident Y
• 02:00 - 02:30 what is your preference I want to identify what is your behavior pattern I want to Target you exactly like uh with the right offer at the right time because if you're interested in cosmetics I cannot give something regarding running shoes right because you're never interested on that stuff so uh businesses uh need some clear picture of this customer data only then they can able to do something so to resolve this issue customer data platforms are introduced so what the platforms do they serve as the solution to this fragmented
• 02:30 - 03:00 data dilemma by providing businesses with centralized Hub where all the customers data is aggregated and in the centralized Hub like they're aggregated for example our data present in multiple sources they get data into centralized Hub they clean it they organize into a unified customer profile so by integrating data from different sources and systems uh like it includes both online data and offline interactions these cdps enable
• 03:00 - 03:30 businesses to create comprehensive and up to dat view of customer so it give you some unified profile view unified customer view which contains demographic information of a customers purchase history browsing Behavior email interaction social media engagements and many more uh so it's painting something like a holistic picture of the customer journey and preferences it give you every information about the customers so after that we'll Target them uh So based on our requirement separately so using the cdps doesn't
• 03:30 - 04:00 just fix the problem of the scatter data it also opens up lot of opportunities for business so with the customer in all the customers ino in one place businesses can understand their customers better like what they like how they behave and what exactly trending what they're liking on it so this helps them group their customers more effectively and make the marketing campaigns more personalized uh like everyone needs personaliz experiences right this will work better so by using by using emails we can give the
• 04:00 - 04:30 product suggestion something like for example if you SE something on YouTube or something you'll be seeing the ads on Facebook or Gmail but something like that so whatever the customer uh required so we'll give you some offers or suggestions recommendations based on it so to to give that we need to understand customers better so that's basically cdps are not only solve the problem of scatter data but also help businesses to grow uh succeeding to this's very tough Market uh in the marketing digital marke is very important so uh the CDP will capture
• 04:30 - 05:00 everything on that so in simple terms that's that's what it does like it contains data in multiple places It capture the data and bring it to one centralized Hub then for example the customer may have partial data some customers doesn't have the full information everything for example we have 200 uh customers but in uh but in reality we have only 50 customers have rent data of it everywhere so what exactly this does we'll get the all the data together we'll Club it we'll use some a and ml
• 05:00 - 05:30 mechanisms there and we'll combine it and by getting the output we'll get a customers with 360° information those 50 customers for example we sending some 200 customers into it but only 50 are real customers out of it the AP will generate 50 customers out of it with all the information whichever related to Dos so after that uh we'll segment out what Exel requirement we have then we'll Target them separately so this is exactly what uh happens inside CDP so any questions you have in this
• 05:30 - 06:00 no but the uh the thing is that um CDP it's main like a uh customer data segregation in L okay means a lot of customers are uh uh they have different customers and different usage different choices it will just segregated and give one perfect solution to business to Target exactly yes so like it gives the UN ified profile view unified profile
• 06:00 - 06:30 view in the sense for example your data is very much scattered for example your preference data stored in somewhere your loyalty data is stored in somewhere so your transaction data is showed in somewhere and also the information is partial the same we know this repeating multiple times we don't know which we know this who right uh so when once we get all these data together in one single Hub then we'll give some holistic unified view of this customer like for example vinod Kumar B is one customer vinod Kumar a is another customer they doesn't have anything common so there will be two different customers so we
• 06:30 - 07:00 divide the customers accordingly and give the complete information about customers then we'll Target them separately what what is that is right but this what happens in cdps yeah got it thank you so uh let's understand the types of cdps so there are two types of cdps based on approaches we use in cdps so and other data management system to identify and match customer records across various data sources uh like for example I'm just telling you right we we're getting data from multiple sources we placing it one Central Hub right so
• 07:00 - 07:30 after that we want to check how these customers are combining uh like we are combining multiple sources of data how how this com based on some identity this should be combined right like how in databases we combine Things based on the primary key or a foreign key so there should be some mechanism involved in cdps how these customers are com combining based on what identity uh so how how your next information or other information of yours can be matched together so there should be some sort of of matching some
• 07:30 - 08:00 sort of things happening inside a CDP right and I told you about this ml mechanism right using that it will be combined So based on how they combining CDs are divided into two types one is deterministic matching another one is probabilistic matching so uh this deterministic matching involves matching customer records based on unique identifiers or keys that are shared across the data sources okay like uh exact it should be have exact matches of identities like for example you have information like
• 08:00 - 08:30 vinodkumar gmail.com if some other person have the same email then only you people will match together if there should be exact match with the identities we're giving it could be email address it could be CRM IDs it could be customer IDs it could be phone numbers or any personally identified information when there's an exact matches formed their identities are generated they they'll be mixed together and they'll form unified customer profile suppose CDP Rees a customer data from company's website or a mobile app or in store P Sy in each of
• 08:30 - 09:00 these sources it provides a unique customer ID or email address right in general so the CDP use this deterministic matching to link all the interactions associated with the same customer ID or email address into single unified customer profile so that that's what exactly deterministic matching is and there are many cdps in Market using it deterministic matching technology uh one is AP uh which we are discussing right now and one is sap CDP and Salesforce CDP and there are many things because there are around 40 50 cdps in
• 09:00 - 09:30 the market but majorly people use AP maybe Salesforce and things but yeah these are the examples of deterministic matching so uh there's another called probabilistic matching this involves using algorithms and statistical techniques to match customer records based on non-unique identifi or attributes that are common across other resources uh like these are not unique uh they match on non-unique things like they'll see the probability for example
• 09:30 - 10:00 these attributes might include demographic information purchase Behavior device characteristics or browsing patterns whatever so instead of relying on exact matches probabilistic matching assigns probabilities to potential matches based on the similarities between records so for example if there is vinod Kumar B and vinode it see it sees both see the probability matching there are 60% match or 70% match on your names it thought okay both persons are same we can match it instead of taking the exact matches exact identities it take some probability percentage if there's any
• 10:00 - 10:30 like how we do the permutation combinations it will check the probabilities if there's any match in the probabilities like May 60 or 70 based on the requirement we provided then it will match like consider the scenario where there are two customer records have similar but not identical attributes like similar names addresses and purchase histories the probabilistic matching algorithm might calculate likelihood score like uh what exactly that indicates the probability of these records represent the same individual So based on that these will be combined
• 10:30 - 11:00 together like simple difference theistic matching the common ID or a common something it will match accordingly if there's exact match probalistic matching think uh this could be common so they'll check for the percentage match maybe like email they'll check the percentage match 80% something it will give the numbers to it it will calculate based on it like there are few uh things that that in the market treasure data amperor examples of it and I told you about this Oracle Unity right they recently launched a CP it combines both
• 11:00 - 11:30 deterministic matching and probabilistic matching so it mix of both so recently they're also launching the combination of both cdps which is Oracle but it was not evolving just started but these are the basic types of cdps and examples so any questions here h no got it thank you I'll go to next I'll start with ap directly so as I told you AP is a deterministic matching tool it's a cloud-based dat mytic matching customer data platform it was developed by the Adobe Systems so as every CDP it is engineered to
• 11:30 - 12:00 empower businesses by centralizing and integrating customer data from diverse sources enabling to craft personalized and cohesive experiences across various channels and touch points uh so AP architecture is very crafted to support the collection processing and activation of customer data so it comprises different layers uh first layer is data inje layer you can able to see this one data inje layer so we can see data is
• 12:00 - 12:30 coming from uh different uh things like we have numerous sources of channels for example if you have S3 data the data is stored in S3 bucket or SFTP bucket or maybe some other Adobe platform because in Adobe Cloud there are many tools like there is CJ There Is AO there's a there's Adobe analytics all those things if the data is sh or stored somewhere in the Adobe platforms or there is any third party sources are there any streaming sources uh are there
• 12:30 - 13:00 any bat sources there are many types of sources right so it collects data from different sources that's what we are saying could be client side information could be server side information it collects the data from different sources called Data injection layer where we have two types one is batch data one is streaming data so batch data data is come in a batch way streaming is continuous uh data so this how different sources we just have in data injection layer then we have something called data storage layer so
• 13:00 - 13:30 once data is ingested it finds it a place in this layer AP employs is scalable and distributed infrastructure to accommodate this vast volumes of customer data so this is called Data Lake it kind it stores both raw data and structure data so whenever we have lot of tables we have lot of bulk tables we inest into AP then this thing will be stored in somewhere called Data Lake like it contains all the raw information whatever we ingested into this particular AP then we have something
• 13:30 - 14:00 called customer profile database okay so what happens as I Told You So once data got injested into this thing data layer based on the algorithm or based on the identities we provided into that particular data the things will match uh so it will do some stitching it will do some identity resolution it create ID graphs then it will combine them and making them as a different profile fragments maybe we have uh we have four Veno Kumar or three
• 14:00 - 14:30 soit they combine together and are making it as one or maybe we have we have so and soit B they have two variation two different customers uh we have six seven different data about them so they'll divide it into two two just two profiles so this this this type of thing this type of stitching this type of things are happens inside this realtime customer profiles real time customer profiles in the sense this raw data will automatically convert it into profile fragments with a unified View
• 14:30 - 15:00 so that's what the data L and next it's real customer View and after that we have something called segmentation service so I'll explain you later but I'll give you just a b description about it so once data got ingested profiles are formed so uh I am a company I want to Target for customers I don't want to give same offer for all the customers right so I have to I want to give different offers for the different customers based on my requirement for example uh I recently created an application no no one purchasing in the
• 15:00 - 15:30 app everyone is purchasing offline so I want to Target customers who never purchase anything online only on offline so I want to I want to separate the audience separate the profiles I want to Target them with a different offer uh maybe I want to Target a female customers whose birthday is coming this month so I'll Target them I'll do some rules I'll create some segments I'll divide them and I'll create them as separate audience this separate creating separate audience based on common traits or based on some rules is called segmentation like segmenting the
• 15:30 - 16:00 profiles into different ways based on the common features they just had so it's called segmentation service it will be two types one is batch one is streaming I'll explain you what exactly there in the later classes but yeah batches the batch F streaming is more like a real time process but in general the the segmentation Services dividing the unified profiles based on the common trades they just have and uh and we just have this data governance layer uh the data governance
• 16:00 - 16:30 layer like obviously we data we data should have some privacy and governance right it should be compliance with the Privacy regulations and data protection standards we just have could be CCPA it could be Hippa or the health care centers it could be uh brexit they have something for UK and us they have something CCP and all that stuff so all the data should be regulated on this so this will be automatically uh there so we are not doing anything from our side but we just ensuring that AP is taking care of these
• 16:30 - 17:00 things like whatever data we just collecting whatever dat data we just targeting and everything it are ader to this data governance and privacy rules so after this CDP there are many other tools like we are settling these things to the destinations like uh we are segmenting the things we're making the profiles and then we are just targeting the profiles maybe we can send them an email we can send them SMS we can send them a Facebook we can show the ad on a Facebook or a Google ad or like that we
• 17:00 - 17:30 have some 70 80 connectors uh in the AP where we can able to Target the customers differently like wherever you want to Target it you can do it like Instagram Facebook YouTube wherever you want to do it you can do it similarly there are many other uh tools which is very buil inbuilt to AP we can also take the data to there there is something called Journey Optimizer uh using which we can able to create campaigns offers and all that stuff that's another tool completely but is is built on on top of AP where we can able to get this data
• 17:30 - 18:00 directly into AP and we can create some messages create some interaction Journeys there and similarly there's something called customer gen analytics where it where it help us to analyze the data like how we use powerb and Tableau like different graphs and dashboards like it will clearly shows the dashboards and graphs like how many how many people's open the mail how many people didn't open how many errors are came how many was okay for the offers how many not it give the clear picture so that we can able to read Target the customers when something was not
• 18:00 - 18:30 happening like this we have multiple uh destinations after from AP to this but this thing was a separate sessions AJ and CJ because these are completely different tools there are a lot of things inside it but there are many other tools where third party tools we can able to Target things so this is a simple process it collects the data then it send to the this particular AP then based on whatever uh what whatever policies we just have whatever the identities we just have will Connect into profile the profiles will be uh segmented into
• 18:30 - 19:00 different audience based on the rules we just provided after that we'll Target to different destinations got it this is overall architecture of this particular so any questions here no thank you so then we'll talk about sources as I told you uh the sources is where from where we're getting the data right so within AP sources have various places where data is showed or collected so these sources contribute valuable
• 19:00 - 19:30 information used to create detailed customer profiles so for offering person experiences and understanding customer behaviors and everything the sources are very important because we're getting the data from it so examples of the sources are as I told you previously Adobe applications we can uh because Adobe uh is in AP is in Adobe cloudspace there are many many Adobe tools are present in here but there is no need to learn everything here but because whatever the implementation happening in Adobe right
• 19:30 - 20:00 there are many tools that happening like there's something called ACS like campaign service a is happening there's something called Adobe launch Adobe Target there's lot of things already present on the top of it AP is built so whatever the data present there it automatically can uh get get the things to AP as a source then we can use the data and create profiles so we not connecting anything on that we directly have a connector we'll use it directly and we'll take the data so like that we have something called Adobe application and we'll have uh this Enterprise Data
• 20:00 - 20:30 Systems as well as we have uh web and mobile sdks so as I told you there are Enterprise sources like we have many many connectors uh like we have default Source connectors in AP I'll walk you through the UI how exactly AP looks but in general there are many many sources present in AP like connectors we can directly use them and for example if you want to connect to the S3 of your client location S3 bucket of your client location they give you all the things like like a bucket name uh the password
• 20:30 - 21:00 RSA token everything so you will give the credential it will directly connect it from AP no need to use Python scripts and all the things that generally use in databases and cdbs here we have connectors like we have around 240 connectors in total in sources L of different types we can use it directly and connect to them and there is something called web and mobile SDK it's called Web software development kit mobile software development kit for example there's lot of data that present in the third party
• 21:00 - 21:30 that was neither sorry that was neither belongs to adobi apps nor belongs to any connectors that already present by default so there's any third party applications which doesn't have any connectors in AP you want to get that data you want to collect the web for example there's a website information you want to get that website information into your particular uh AP but you don't have direct connectors to it maybe Amazon information we don't have maybe we have some something like mintra for mintra we have there's a website we need
• 21:30 - 22:00 to get the website information if mintra got an app we want to get the mobile uh data information about it how the interactions are happening on that but there is no direct connectors use for all the apps right because there are many many apps AP cannot provide direct connectors for all the apps that scenarios we use the concept called web and mobile SDK where by default we'll code the things and we'll get the data and we'll use them and along with that we have this batch and streaming apis so through apis we can able to connect with the sources and we can able to get the data into
• 22:00 - 22:30 this particular AP it could be batch or streaming like streaming is Real Time batch it come as a batch so these are exactly what exactly the sources are yeah got it uh so in in this experience platform we have around Enterprise Solutions as 70 plus uh so these integrate these what happens generally have basic anation we need to do we need to these connections are very effortless we need to use the
• 22:30 - 23:00 tication then we'll connect directly to it we'll get the list of files we'll select what are the files we required there and we'll connect it directly to that and there lot as I told you there are lot of supporting things here like file storages data warehouses like S3 cloud and everything there are streaming thing like Amazon kesis Google psub uh lafate and live ramp event Hub and all those things and there are it's also connected with ETL Partners like Informatica softw trims and there is Talent there are many many sources as I
• 23:00 - 23:30 told you there are around 200 plus sources here so lot of things are happening so we can able to get around 120 terabytes of data every month we have permissions to take that much of data from these sources so we can able to do some a lot of things here like let me open uh the UI it will work so can you see my screen right yeah so this is how AP looks in general so uh I'll
• 23:30 - 24:00 explain you these things what what are the sandboxes and all that stuff but first I'll explain you what exactly the source is where it located so you can see uh this inside there are many things uh inside connections there you can see something called sources so when you click on the sources so it will give you all the source connectors that present in AP the now okay for example this these are the many connectors present in
• 24:00 - 24:30 AP like uh if you want to connect to a source that is related to any adob application you can able to see here these are the things that already present any advertising things any analytics things cloud storage sf3 PS3 so there are different types of uh these things different types of connectors that we directly have so you can use any one of it uh or may two of it three of it based on the requirement you just have based on because uh it Chang client to client because I work I
• 24:30 - 25:00 worked around some 13 projects in cdps so many clients store the data in different places based on their requirement based on their present implementation happening in their site right maybe they store somewhere in Adobe canes some some someone store in S3 bucket someone store in SFTP someone store in big quy so what we need to do is we need to get the credentials from them like uh view credentials or something like that they'll provide us directly like if this S3 they'll give the root access so we need to provide so for each thing we need to provide the authentication they provides like what
• 25:00 - 25:30 is account name S3 key and the secret key and what ELO required then when we click it automatically it will be connected uh when we need to give the authentication then uh the particular bucket is automatically connected so no need to use the Python scripts and everything it's a simple process so whatever the source connector you want to connect just go to Source select the source set up the connection give the information that you have there and click on uh create it will create directly so these things I'll I'll this are inje part I'll I'll explain you
• 25:30 - 26:00 later but authentication part this is what we gener leers if you want to connect to any Source just give the credentials and whatever you just have it it will work by default we have many we can use it but majority of places right we don't use any third party thing because majority of the clients only use this either they use Oracle or maybe Amazon AWS things but only rare case in my 13 projects I completed only one project they have data and the third party things other than this Con so rest of the things will have information in
• 26:00 - 26:30 these connectors only so this a simple uh UI process you need to click on setup and do this so that's what sources exactly does but how to after setting with the sources how we can inest the data into it I'll explain you later but the this what the sources are so any questions in sources and stuff no got it understood uh schema is nothing but think cdbs in this database point of view okay in database the main important things are tables where we structure the things and everything so similarly here
• 26:30 - 27:00 we have schemas like schema is like a blueprint for data it sets rules for how data should be structured and formated uh so what like this ensures some consistency and validity as data moves between different systems uh so in this schemas access universal language for describing the data uh like we can reuse without any conflicts or something so these are very much useful to store large and complex data efficiently this uh schema contains many things
• 27:00 - 27:30 inside it like it had classes it got field groups it got Fields it got data types I'll tell them exactly what they are but schema was nothing but a table which contains only the structure and there something called XDM XDM is nothing but experience data model it called XDM sometimes referred as XDM schemas and all you can see XDM here right XDM schemas they'll call it as XDM schemas or X M XDM is nothing but experience
• 27:30 - 28:00 data model so the meaning of it is so this particular a and Adobe they provide some definition to the schema like it has to be in this structure uh like how the table should be there in this way the hierarchy should be this way they have some structure on it by default so those those considered as an experience data model like they have some sort of data model everything was stored in certain place so that's why it's called experience data model that's it so a schema can many things inside but simply we can call a schema was class plus a
• 28:00 - 28:30 field Group I'll explain you what is class and what is field Group let's it's just class and field Group it's called schema and schema is just a table structure it's not a table I could say structure it's just skeleton of a table like it contains attributes uh it contains column terms that's it doesn't doesn't store any data it had only the structure so any questions schemas because it's important topic here that's why if you have any questions you can ask me um this is okay means this schema
• 28:30 - 29:00 it's like a table okay in a database means uh um means these categories they use like Behavior class field groups data type and Fields like that to store the data yeah right no no yeah yeah kind of yes but uh these kind of tables only but it doesn't have any data it have some structure that's it for example think like that you have a customer table in database the customer table have all the information about customers like customer name age demographics all the stuff it had all the information
• 29:00 - 29:30 about so this this schema similarly it's something like the table but it doesn't have any details inside it it just have a structure okay like it had like it had only column names that's it understood so like we have we can have multiple tables we fit into this particular schema you know the concept of views right like uh in the python and something we'll create a views where we store the data separately we have structure separately uh
• 29:30 - 30:00 no yeah yes something similar like we have structure is this is structure and data stored in some other yeah exactly exactly this is just a structure this a skeleton that's it okay okay so it had some definitions that's how it should be that's why it's called XDM experience data model because it's Bel to Adobe experience platform that's why it's called experience data model and it contains uh class plus schema field class I'll explain you what exactly class first because everything
• 30:00 - 30:30 was inside in the schema only so whatever because CDP is very very simple tool that's why I'll just start with schema only so there's something called class so a schema's class defines the behavioral aspects of the data that schema contain record based data or a Time series data uh for example if you're creating a schema first you have to Define is what is the class of that schema by default we have two classes uh in our AP one is uh experience event one is individual profile or one is record based one is event based what's the
• 30:30 - 31:00 difference is if it's data is record based like it is a a record of a customer record of a profile like age name uh loyalty details these are records of a customer the record value attribut value of a customer this is considered as an individual profile or record based there is some other other type is called experience event or time series these are events for example you perform some task at particular point of time it was captured that is called event data
• 31:00 - 31:30 or a Time series data for example you made card abandonment or you made a click on something you purchase something you trans you did some transaction these are you're doing some tasks at one point of time like you doing some event data you're doing some event at a particular time frame a particular time stamp so these information are called event data these are stored as in an experience event class so whenever you create a schema there'll be two types better
• 31:30 - 32:00 so but you you clear about this particular thing right what is schemas and stuff got so when so in the data management you can see schemas okay so when you click on this create schema and right hand it will ask you whether it's an individual profile or experience event or others okay as I told you uh schemas are of two
• 32:00 - 32:30 types the these are classes you can select a base class for the schema so inside schema the first important thing we need to select is classes so classes defines what type of data is storing in this particular schema it is record based or it is time series like it is a record profile level information of a customer either a phone number or an ID or or maybe loyalty ID maybe some profile ID something there a lot of details about a profile right those detail when if you want to store it to
• 32:30 - 33:00 this part to this particular thing if you want to create schema based on this record based then you select this individual profile if the data are coming is part of event like you are adding something to the card you abandon the cart or you you reviewed something uh you purchased something you made some transaction these are events in the marketing space these are called as events those are not uh profile information those are event information on particular time stamp you're doing for example because these are not coming on profiles these things
• 33:00 - 33:30 will keep on changing one time frame you did something in another time frame you do some other thing like you'll add cart in this particular time frame you abandon the cart in some other time frame you purchase the same thing in another time frame so one single time frame particular event you provide you did right so those kind of schemas you have to create like transaction schema then you'll say like experience event as a class there is something called others so this individual profile this experience event whatever data going inside these two right will form as profiles we'll go into profiles we'll
• 33:30 - 34:00 take that will considered as profiles because these two are called XDM schema XDM classes these are defined uh with the language with experience data model I told you they have some Rules and Things these are St structured like there if you don't want to keep the data into profiles you want to keep that as a separate entity like you don't need to involve into this the profil of the customers you want to keep it a separate ENT you want to use it whenever it's required like like in database we use lookup table if you know yes so like
• 34:00 - 34:30 that uh we use this like this doesn't belong to this uh stitching and everything but we can use it whenever it's required for example take a scenario of product product doesn't uniquely Define a customer right uh so product because you can purchase the same product I can purchase the same product we cannot Target a customer based on some product information so there is no need to have the product details in the profiles and events so what we generally does we'll keep only product ID here and we'll
• 34:30 - 35:00 create a SE separate lookup table in other called Product table it contains all the product information like product name color category all that stuff so it doesn't it doesn't comes under profile so but we make a relationship so whenever it is required any product information we'll grab it from here but we doesn't involve it in the profile but you'll understand whenever we create a profile and everything but yeah the basic differen is whenever you want to participate in profile if the data is you select individual profile the data
• 35:00 - 35:30 is time you select experience event if your data doesn't need to participate in the profile switching you can keep it as a lookup table then you keep it as another any questions here no got it yeah so this CDP implementation starts like this implementation what exactly when s is signed and everything they'll start something like this they'll give you group of tables uh they'll ask you uh May they'll ask you which could be goes under individual profile which can be an experience event what exact the schema names like schemas are nothing but table names you have to require
• 35:30 - 36:00 there's there's need of combining them or there need to there no need to have multiple attributes you can able to divide it so first and foremost thing you have to do is which belongs to individual profile which belongs to experience event what are lookup tables that's it lookup tables we hardly use but these two are very important so whenever data is with experience we'll get that whenever data is there based on the data thing we can able to because I'll by going deep inside we can able to understand the exact difference difference how it will work in interactions and all that stuff but in
• 36:00 - 36:30 general this is the difference and this main important thing whenever we want to implement anything in CDP so we need to figure out which belongs to individual profile which belongs to experience so once we select the class we need to give the display name don't worry about UI the UI keep on changing uh so I'm giving test C like it's a table name like schema name you want to give description you can do it
• 36:30 - 37:00 so after that it create something like this so this is I I told you about the XDM thing right this is the XDM thing they have this particular structure this automatically created by System itself you're not doing anything there so this is already defined by their own model the old experience model that's why it's called XDM schema here we can see the schema name we created here we can see the CL we selected so next we need to add
• 37:00 - 37:30 attributes to it like we created a table uh we created a type of table it is like it's record based a Time series then we have to add attributes into it right like a customer ID customer name and all that stuff like how we add the column names in the database similarly we need to add attribute names here there it comes the concept of field groups so field groups is nothing but a reusable component that defines one or more fields that Implement certain functions like personal details hotal preferences or address like these are intended to be
• 37:30 - 38:00 part of your schema that implements a compatible they're very much compatible to classes like uh in simple terms how can we say this field groups is nothing but a group of fields together we can use it in all the schemas whenever it's required so instead of creating for example you create some identities like customer ID CRM ID like all the identities into one single field Group so you can use that field Group in any other schema no need to create all the attributes again and again manually so field groups is nothing but group of
• 38:00 - 38:30 fields together was called as field groups that's it like instead of those are reusable whenever uh you you required to use it you can use it so all you if you want to add field groups or if you want to add for example there are many default field groups okay like out of the box field groups for example if you click on ADD you can see based on the industry you're working so if your client is retail your if your client is Healthcare if your client is Trav and thing they'll give some default field groups for this there
• 38:30 - 39:00 are many default field groups that already present here you can use them if it's required to you but no need to use them if it's not required if all the attributes are different than these things you can use it but few things we generally use it from here itself like there is something called demographic information like it contains the demographic details of a customer so if you click on ADD field groups it is automatically added here so it contains the new thing the new object is added here inside this you
• 39:00 - 39:30 can see name birth date name in the sense full name cesy name suffix prefix and everything the gender all those things because these are very much required for the customer thing right so I'll keep it as a customer table maybe that will be better so for the customer table these these things are important right so that's why we just added but if you don't want to add this you want to add everything from scratch your own words then you can able to click on ADD there's something called create new
• 39:30 - 40:00 field you can able to add your own field Group like uh we working some uh maybe we worked for Amazon so we do we just write Amazon customers demographic or whatever it is then we'll click on ADD field groups so the new thing will be created here we can able to add multiple attributes inside it for example if you click on plus here the fields will be displayed here you can the field name field names is nothing but a column name like column
• 40:00 - 40:30 name in how you provide the databases you'll provide the column names here for example this is a customer table I want to PR the customer ID first so I I mention the customer ID then I'll select the type of this is a data type there are many data types is suppos string double long integer array object and all that stuff so whatever it's required you can keep it and I'm making it as a string so here it will ask for the field Group in which inside which field Group I want to keep this so so field Group is we cannot add attribute directly here
• 40:30 - 41:00 okay for example I want to add customer Ed directly here instead of adding field group that I cannot do I want to uh I can keep Fields only inside a field Group we can I cannot add Fields outside the field so whatever attribute I'm just adding I need to P it some field Group that's it so that's why it's last Q at what exactly field Group it should go on that's why it's in stat symbols mandatory here if you do it without it it will throw us an error so here I'm I'm going to this Amazon thing so here I'm just selecting that so if if you want to add any default value if if
• 41:00 - 41:30 there is nothing there if you want to add any default value you can add it here is there any pattern should be there in the data that coming inside it you can able to add it here any format of it like it it's a DAT format email format or Ur or something like that you can able to add it like here uh are there any restrictions you want to keep any constraints you want to keep like is there any minimum length of this data is only six digits or seven digits there's a requirement from the client you can able to add it here so so after that if you want to make it as a required field
• 41:30 - 42:00 uh so only data inest if data is inest then the in happens then it will fail then make it as required or uh you want to keep it as an array you want to store multiple values inside this particular thing you want to get the data as an array you can select this there you know about enum right so enum as a type in in database in database same like enum like for example in enum we have it's a series of suggest registed values so it accepts only that values for example gender we
• 42:00 - 42:30 can keep male female unknown as three options in enum like three suggested values if other than three something came inside it it will fail the inje so enum is nothing but the set of suggestions or set of suggest values for this particular field we just targeting you want to keep it you can keep it maybe if few customers few clients want only this data should come inside it you you can keep it you like we need only country codes and everything for those scenar you can do it this identity part I'll explain you tomorrow class because
• 42:30 - 43:00 it's I want to connect with some other thing that's why I doing it it's more like a primary key secondary key and foreign key we use in our databases right it's a similar concept but I'll explain you detail in next class sure so so like this we can add multiple attributes in one field Group like this we There's No Limit we can have have add multiple field groups in the same class and same schema and and another concept of this field groups is these field Group grou for example we created a new field group called Amazon for I can
• 43:00 - 43:30 apply this will be added here so similarly I can able to add multiple attributes inside it I can keep on getting it I can keep on adding multiple attributes in particular field Group and this particular field Group is very much related to this class okay uh like uh I told you there some other class called experience event so this field Group I cannot use in that particular class and some other class other than this so I told you field groups are reusable but reusable only to
• 43:30 - 44:00 the same class okay so create any other schema at the same class then we can use similarly with experience event if you create something in experience event I can only use in the experience event schemas understood [Music]