CDP curriculum

Step 2: Data modeling Steps to promote data source objects to data model:

1. From data stream record home, click on mapping 2. Identify which object(s) the data source object should map to. Create your own If necessary 3. Complete mapping and save 4. Define relationships between the data model objects to which you just mapped

Dynamic segmentation vs Calculated Insights

Dynamic segmentation Self-service, simple use cases, specific targeted campaigns w/out need for reusability CI -Don't need to calculate everything every single time the query runs -Gives a level of abstraction & standardization -Nontrivial calculations -Support for longer date ranges -More complex use cases

Courses 1 and 2 Known data

Known data is information definitively tied to a specific person. It's like an email address or a phone number used when a customer sets up an account.

Known identity Recognized state Anonymous identity

Known identity- like name, email, phone number and recognize the person Recognized state- between known and anonymous where someone is logged into website and can give info to turn into known Anonymous identity- not logged in (to website for ex)

Course 7 - Identity resolution

Learn how to unify customer data by using identity resolution match and reconciliation rules

Changes in CDP

"Unified profile" term replaces MC term "contacts" Match rules & identify resolution can create unified profiles for each customer instead of having multiple contacts for each customer Segment publishes measured every time customer publishes activates a segment. Each publish operation for a segment counts as (1). Engagement Events- sum of all data that represents engagement between company and customer (like email engagement or refunds or purchases or website visits). Have use cases for configuring event data streams into CDP

When is an opportunity qualified?

1, Good fit accounts for C360 will have these attributes: Salesforce sales/service cloud, marketing cloud, audience builder, Ad Studio, MC Connector Salesforce sales/service cloud, marketing cloud Heavy on MC, often over on contacts, multi-channel journeys MC and Audience Builder (but unhappy) 2. Bad fit accounts for C360 will have these attributes:' No salesforce footprint, and request a CDP with complex use cases Need marketing automation, small footprint General tone: follow a best of breed strategy (only the best point solutions) Large inhouse development focus (resources and skills internally) 3. Further clarification accounts: Identified as big bet account (would need in-depth discovery for long term success) No footprint, but interested in Salesforce Platform (cross-cloud positioning). Know their goals though Real time activation (what does real time mean and to where) May need just IS or both

Components that make up the filter in segments: Aggregation Operator Value

1. Aggregation= select aggregation for an entity based on count, sum, average, or max/min, along with an operator and value for a new container attribute Count: ex. At least 5 purchases Sum: lifetime purchase of $500 Average: individual average lifetime value of $500 Max: max purchase amt is $1000 Min: min purchase amt at least 5$ 2. Operator= specify how filter criteria relate to the value entered (like the verb in a sentence) -Date= last year or this year or last number of days etc -Ex. To set up a batch email that sends to profiles on their birthday you may use attribute= birthdate, operator= is anniversary of, value= today's date -Numeric= no value, is equal to, is greater than, is between, etc -Ex. Send a special email offer to customers who spend more than $100. Attribute= grand total amount, Operator= greater than, value= 100 -Alphabetic =is equal to, is in, is not in, contains, begins with, etc -Ex. You want to send an email to customers who live in a certain state. Attribute= state, Operator, =is in, Value= IN, Indiana 3. Value= thing you want your filter to find. They are not case sensitive

Knowledge Check 1. Which CDP feature helps calculate reusable attributes like lifetime value and customer satisfaction score? 2. What step in a segment configuration determines the attributes that show up in the attribute library? 3. Which attributes in the attribute library have a one to many relationship with the segment target?

1. Calculated Insights 2. Segment On 3. Related Attributes

.Feb '21 release

1. Can import data sources from many places like Loyalty Cloud besides MC and data extensions 2. Calculated insight can do spend value by product category

Trailhead Salesforce CDP Basics Prepare for Salesforce CDP

1. Clean your data 2. Ingest your raw data (including 2nd and 3rd party data) First party data- your customer data that you own and manage Second party data- someone else's data that you purchased from another company/source Third party data- collection of data from many sources (ex. Behavioral/demographic) 3. Map and model your data Select model matching your use case or create a custom model Map data to identify relationships 4. Identify your customers You have known and unknown customers in your data Customer identity resolution rules helps you manage duplicate contacts in your imported data 5. Select your channels 6. Segment and activate your audiences Segment with drag and drop. Activate that segment across channels you selected 7. Do your marketing magic Ex. Can use your CDP audience segments as entry source in journey builder 8. Measure your success Ex. Higher customer lifetime values, increased growth

Knowledge check 1. What does CIM stand for? 2. What 3 connectors are available out of the box for data ingestion? 3. When mapping your Contact Point Email object, which 3 fields required to successfully use identity resolution, segmentation, and activation? 4. When setting up the data source object or schema corresponding to the data set that you're importing which category would you select when bringing sales order data?

1. Cloud information model 2. MC, Sales/Service Cloud, and AWS S3 3. ContactPointId PartyId EmailAddress 4. Engagement data

Salesforce CDP for Admins Trailhead Additional admin tasks (part 1) 1) Create a package in Salesforce CDP. 2.Identify your account's functional domain.

1. Create a package in Salesforce CDP. As an admin, you can package and install your Salesforce CDP S3 data streams for distribution. A package is a container for Salesforce Metadata Component that can be either individual configurations or an entire custom Salesforce platform app. After the package is created, you can distribute it to other Salesforce users and orgs via an install link or AppExchange. 2. Identify your account's functional domain. The functional domain refers to the Salesforce and public cloud infrastructure where your instance is located. This information is helpful to provide when troubleshooting any potential issues with support. (Similar to knowing your MID in Marketing Cloud). From CDP Setup, click Setup Home. Your functional domain is listed at the bottom of the page.

Course 3 3 solutions for CDP: (depends on the use case)

1. Customer 360 audiences- long term CDP, system of insight, a single view of your customers 2. Interaction studio- real time personalization engine, some CDP functionality like unifying customer data with web and mobile capture, system of engagement 3. Audience builder- legacy exact target segmentation tool, more for technical marketers. Long term trying to migrate people from audience builder to CDP

Knowledge check: 1. Which 3 match rules available to link multiple records into a unified customer profile? 2. Which match rule allows you to unify records based on an external loyalty ID? 3. What reconciliation rule is defined by default when selecting profile attributes in the Unified Individual?

1. Exact phone number Exact Name and Email Exact Email 2. Exact Party ID 3. Last Updated

Example: external identity graph

1. Include identity linkages from external sources -Import identity links -Match rules can use imported identity links that are mapped to Party Identification object -Ex. use subscriber key as way to connect records to MC -Multiple types and names -Use type (ex. For customer identifier and Name (for ex. CRM identifier) per record to control which IDs should be matched -Unknown user matching -Configure Device IDs as Party Identification objects to enable matching of records for the same device 2. Mapping for Party Identification object -> Identity resolution based on IDs -The ID used for Exact Match -Records with the same Identification Number will be unified by mapping to Identification Number -Populated with Individual ID -Determines which Individual objects get unified -Map to Party -A unique key for this object -Only one record per Party Identification ID will survive 3. Bringing in ID Name-> CRM customer ID and ID Type-> person identifier 4. Also bringing in External ID and Customer ID -Map into contact point and individual objects -Then the External ID is mapped to Party Identfication ID, Customer ID mapped to Party ID, Party Identification Name is mapped to CRM Customer ID, and Party Identification Type is mapped to Person Identifier -The external ID to Party Identification ID is what allow each record to be unique -The unified individual then has GUIDs for their party IDs in unified contact point object and Individual IDs in unified individual object

Prospects looking for system of insights or system of engagement or both

1. Insights (C360)- data unification with batch ingestion, holistic customer identity resolution, segmentation & insights, outbound messaging activation. Development platform at large data scale 2. Engagement (interaction studio)- real time data flow & profile store, experience cut across online & offline, real time interaction and management for personalization. Event triggers & AB experience testing. AI + Next Best Action 3. They will get better integration going towards a Unified Marketing Platform Vision ^

Knowledge check: 1. Which CDP permission set manages overall segmentation strategy and identifies the target campaigns? 2. Which tab in navigation manages the data coming into CDP? 3. What must the CDP admin do first when setting up users in CDP?

1. Marketing manager 2. Data streams 3. Create profiles for each CDP user role

Knowledge check: 1.What are 2 root causes to why companies face fragmented data issues? 2. Which 3 topics are covered during a standard discovery process?

1. Misaligned business goals and department silos 2. Use case definition, identity resolution requirements, and identity data sources and mapping

Salesforce CDP for Admins Trailhead Additional admin tasks (part 2) 3. Auditing and ongoing maintenance:

1. Monitor usage entitlements. As an admin, you need to monitor your account for activities that impact your contract. These include: Unified profiles (similar to Marketing Cloud contacts), segment publishes, and engagement events or records . 2. View identity resolution processing history. Once resolution rules have been enabled by your data-aware specialist, it is a good idea to regularly visit the Identity Resolution page to check the processing history and consolidation rate over time. 3. View record modification fields. All objects include fields to store the name of the user who created the record and who last modified the record. This provides some basic auditing information to help troubleshoot. 5. View and monitor setup changes with setup audit trail. Admins can view a setup audit trail, which logs when modifications are made to your organization's configuration. 6. View and monitor login history. You can review a list of successful and failed login attempts to your organization for the past 6 months. 7. Monitor account status on Salesforce Trust. Monitor service availability and performance for both your Marketing Cloud connection and your Salesforce CDP account. By doing so you can quickly identify issues

Some summer '21 release improvements

1. Navigate more efficiently. Use drop-down under data stream's name to navigate between streams in mapping 2. Trash Can Icon -If you want to remove all field mapping between data stream and selected data model object or Delete data model object entirely from data model - "Remove mappings" or "Delete Object" -Note: unmapping from manually is supported for non-primary key/event date time fields and shouldn't be used to remove ENTIRE data stream's mapping to a DMO (because can't unmap with primary key fields) 3. Dependencies -When attempting to delete a DMO with dependencies, a confirmation box lists out object elements throughout CDP that must be remediated before object can be deleted. -Clicking on each object element will redirect user to details page (they list URLs for you that will open a new tab) -Data streams that are mapped to DMO are included in dialog box just for info purposes. Does not impede deletion 4. Enable Value suggestion- > Select Edit Properties -Configure an attribute of data type text with 'Enable Value Suggestion' which allows you to search data values as you type in segment definitions -Click 'edit properties' in mapping screen-> check enable value suggestion inline next to the field and save -You can have up to 500 attributes for your entire org for value suggestions -for text attributes, you can use up to 100 values but it's 500 for value suggestions 5. Create DMOs before you map (3 ways to do this below) -Choose an existing DMO to use as starting template for customization -Choose a CSV file to upload with 4 ordered columns: Label, DeveloperName, Type, IsPrimaryKey -Manually build a new object by adding one field at a time 6. Data Model Graph -Navigate to Graph View -Visualize how data is connected -Color of each node: Profile, Engagement, Other -Click node to view details and make edits, just like list view -Only mapped DMOs show in this view

Party identifier

1. Party identifier = any customer or vendor supplied id for a person Examples: Subscriber key Contact Id SSN Loyalty Id 2. Party Id match -You're ingesting data from more than 1 source which shares a common party identifier and you want to unify those records together using identity resolution -Ex. You have system which tracks customers including their Customer ID and Loyalty Id and separate loyalty system which uses Loyalty Id -You can do party id match on Loyalty ID to unify data from loyalty system 3. Party identifiers must be 'named' in source streams using 2 formula fields on DSO -Party identification Type -Party identification name -Ex. Field label: PartyIdType, use transformation formula to create text of 'Person Identifier' -Ex. Field label: PartyIdName, use transformational formula to create text of 'MC Subscriber Key' -Then, you map those to the DMO fields like Identification Name and Party identification type. -Map the customer/vendor supplied identifier that's the primary key (ex. Subscriber key) to Identification Number and Party -Then, make sure the relationship is right since it's auto added -Then, you will see exact party identification match rule in setup (chosen party identification numbers must be a match) -Party identification type-> person identifier -Party Identification name-> MC subscriber key

Summer 21' release Segmentation Improvements (part 1)

1. Suggested filters for filter definitions (like autocomplete) -For values as you type in filters or view values in order -Can multi-select values for 'is in' operator -Load/ingest data -Pick which attributes to set up data ingestion and basically choosing attributes to enable on data stream field list (up to 500 attributes and only for text data type) -Create segmentation 2. Easily copy a segment to allow for common segment across teams -can copy to populate the same filters and customizations -can't copy activation targets nor publish schedule 3. Enhancement to operators -'No value' filters and addition of day-based operators -Ex. Today, yesterday, this month, last month, etc -Null as a value for Dates, Numbers, and Text. (previously was only for numbers) 4. Processing guardrails to ensure no service degradation and handling of restriction of processing

Salesforce CDP for Admins Trailhead Set up Salesforce CDP

1. Update your admin user (login for first time and assign admin permission set) 2. Provision Salesforce CDP (complete setup in CDP setup) 3. Create profiles, users, and add permission sets 4. Connect to marketing cloud account using admin credentials 5. Select appropriate data bundles and Bus in MC (data bundles are based on email and mobile channels in MC) 6. Connect to sales/service cloud (if needed) 7. Prepare for ongoing tasks/maintenance (ie. Auditing) Created profiles for each type of user ^ (marketing manager etc) New users will automatically get an email when you create them

Salesforce CDP for Admins Trailhead Your role as admin and the other roles

=user management. User permissions—determined by permission sets—specify the tasks users can perform and the features they can access After creating each profile for each user and then create users, permission sets assigned to each user: Admin-> setup, user provisioning, day-to-day config, support, maintenance, performs regular internal system audits Marketing manager-> manages overall segmentation strategy and identifies the target campaigns Marketing specialist-> creates, manages, and publishes segments of the messaging campaigns the marketing manager identifies Data-aware specialist-> manages logical, marketer-friendly data model defined by the marketing manager/specialist. They create and manage data streams, map data, and build calculated insights to be used in segmentation

Course 4 Trailhead Salesforce CDP Basics Single Source of Truth

A CDP is a place where company collects and stores data about its customers Single source of truth= a suite of products that give you a single, unified view of every customer (with a single customer ID)

A Data Model Object (DMO)

A Data Model Object (DMO) is an object in the data model created by customers. If a new object is created, it can use a reference object from the Cloud Information Model. If the DMO uses a reference object, it inherits the name, shape, and semantics of the reference object. This is called a Standard DMO. Customers can also choose to define an entirely custom DMO, or a Custom DMO. Data cannot be used for segmentation, activation, or analytics until it is mapped from a Data Source Object (DSO) to a DMO

Container paths

A container path needs to be selected when container has multiple access paths back to segmentation entity (this is called Segment On) because there's multiple data sources -Ex. If want to send an email to individuals who bought a specific product, they'd choose attribute connected to retail purchases in a data stream You CAN'T change container path after you select it. You'd have to delete the container and create another one to change the path

3 marketer use cases solved with CDP

Audience suppression, LTV modeling, and Customer resolution

Pricing of CDP

Corporate edition ($150,000) Enterprise edition ($600,000) Enterprise plus ($780,000) Premiere support included

Course 9- Activation Activation Target Activation Activate on

Activation target= create connection to messaging and journeys or Amazon S3, for example, to publish audiences -In setup, you can choose which BU to activate to Activation- publish audiences to Activation Targets for marketing campaigns -Improvements include -> can choose client contact point to message Activate on- select an entity to activate on. Activate on can be segmented on the entity or on the entities with 1:N relationship with the segmented on entity -ex. individuals or households

Create a marketing cloud activation target (have to identify one before you can publish a segment)

Activation targets tab->new activation target->MC->name-> add or remove columns of Bus to receive published segments Activation target is used to store authentication and authorization information for a given activation platform -You can publish your segments, include contact points, and additional attributes to the activation target -Before you can create a Cloud File Storage Activation Target, determine your S3 access key and secret key. -The S3 credentials provided must have these permissions: s3: PutObject, s3: GetObject, s3: List Bucket, s3: Delete Object, s3: GetBucketLocation -You can create only 1 CDP activation target for your org

Additional Attributes Publish history

Additional attributes- push additional attributes to Activation Targets for Journey decisions, message, content personalization Publish history- history of activations to track the statuses

More Summer '21 releases

Advanced functions like ranking, bucketing, lead, lag, exponential, StdDev, Approx Count Nested logic (ex. Nested queries and wrap up with top-level query with dimensions and stuff) View CI results in data explorer Can enable & disable, edit friendly names, and function library in CI management

Step 2: Data modeling Identity resolution

Audiences setup in gear icon -can set the Match Rules to match on Exact Email since we would like the email addresses from the Marketing Cloud email bundle to merge with the email addresses from the runner profiles where appropriate -Under reconciliation rules-> can set rules to work on Most Frequent or provide a Source Sequence (which is most trusted source/hierarchical ranking) A data model object inherits its Category from the first data source object that maps to it. Once the data model object has its category set, all subsequent data source objects mapping to it must have the same category. Overall, can create a new custom object and related it back to the standard Individual object. After activating identity resolution rules and running them, summary statistics are available for your ruleset. -View these in the Processing History tab of the ruleset record homepage True or false: you can rank hierarchy of rules used for identity resolution -True Reconciliation rules allow rules to work on which criteria? -Most Frequent -Source Sequence

Known and Pseudonymous data can fall into one of 4 categories.

Basic data- personal information, and it can be unique to an individual, like phone number or email address. Or it can be nonunique, like zip code or last name. Interaction data- actions people take and how they interact with each other and brands Behavioral data- Marketers begin attributing characteristics to a customer, getting to know them better. This happens over time, as they gather more and more basic and interaction data. With the right behavioral data, marketers determine the best ways to communicate Attitudinal/preferential data- gets to the heart of the customer, identifying how they feel and what they believe about certain things, such as how often they want to receive communications.

Ingest->map-> model Step 1: Data Ingestion (part 1)

Bring in all fields from a data set exactly as they are so you can always revert back to original shape of data if you make a mistake or change business requirements (initiate) Authentication details page -Refresh only new files- once files have been retrieved, we track which files have been picked up and which haven't and pull only new files accordingly -Treat case of missing file as failure- when data stream runs on a schedule, if no files are located in directory, this setting will cause data stream to fail for that run as way to provide an alert -My production file does not have headers- use this if file used during setup had headers, but future files will not

Calculated insights (part 1)

Build multidimensional metrics like LTC Can use in segment builder to understand customers better Can activate for personalization Calculated insights available via API for use at every touchpoint Using SQL: -Select attributes, from <data model object>, join, where, group by -Inner join- returns records that have matching values in both DMOs -Left join- returns all records from left table and matched records from the right DMO -Ex. Left join UnifiedIndividual on SalesOrder.PartyId=UnifiedIndividual.Id -Right join- returns all records from the right table and matched records from the left DMO -Full join- returns all records when there's a match in either left or right DMO

The ideal C360 Customer part 1

Business goals: single source of truth across marketing with unified customer profiles. Improve data governance from consent management to deduplication. Make segmentation processes automates. Activate data in real time Marketing and operations: large-scale advertisers, multi-channel marketers, with integrated marketing goals, automated lifecycle marketing initiatives, challenges with accessing and leveraging customer data sources, limited tech resources within marketing org, and personalization

Trailhead Salesforce CDP Basics Cloud Information Model

CIM= Is a schema/model used to communicate between connected data sources with different data structures and formats Uses APIs (developer tool that allows systems to talk to each other) and other mappings to connect applications and data Open source data model that standardizes data formatting across apps Structured or unstructured data (with or without a formal data model) Creates standardized data models that can be used in common scenarios based on subject areas (like sales orders). Creates a standard way for platforms using the CIM model to map to each other CIM is used in CDP to make data modeling marketer-friendly using pre-built models based on common marketing use cases. These data models are called data bundles (ex. Tracking customer loyalty). Data bundles help you standardize your data Implementing CDP won't impact your current contact model in MC

CRM Connector Data Source (sales & service cloud connector)

CRM connector will provide access to all standard & custom objects that are supported by Salesforce Bulk API You can disconnect a salesforce org from CDP anytime but have to delete data streams, segmentations, etc associated with it first Create data streams: 1. Select salesforce org to create streams from -The salesforce objects you have access to will be listed for selection 2. Can create a data stream with one object at a time 3. Field selection: -Object details category defaults to profile but can change it -Primary key is auto populated and read-only -Deselect any fields not required for data stream because they are already all selected -Use formula fields to define formula expressions for your data stream 4. Deploy data stream: -Name of stream defaults to include name of object + Salesforce org ID -Can change if needed -Refresh frequency for incremental is only allowed hourly but full refresh occurs weekly -Upon deploying, can create mappings to use objects and fields for segmentation and activation 5. TWO ways to create data streams from Salesforce -Data bundles contain prepackaged objects and mappings -All Objects where you can view & select all of loyalty objects you have access to Loyalty management integration: Helps execute personalized messaging Pre-built data bundles Unlock new segmentation powers Segment individuals who joined loyalty programs in last 6 months for example

Course 5 Salesforce CDP for Admins Trailhead Some tabs in the CDP app

Calculated Insights are predefined and calculated metrics that can help marketers build segments. Data Explorer and Profile Explorer are data-viewing tools, allowing a view into ingested data and unified profiles Activation Targets and Activations to manage where the segments are exported—like Marketing Cloud Data Streams are the connected data sources added into your CDP platform

Salesforce CDP for Admins Trailhead Additional admin tasks (part 3) 4. Enhancements

Can package data streams and models as well as Standard data model and calculated insights (package for migration purposes or Appexchange solutions) -Can create a package: setup->package manager -Can download in another org to use and you'd create a new data stream and choose the package you want Control and status events -Can get latest status and info of your CDP status without viewing actual records (objects supported: data stream, segment, activation) -Use cases: reports and dashboards with latest info from CDP -Flow- flows will reliably execute on status change or other field updates -Ex. When data stream status = error or when data stream records >5000 -> Actions send an email, post to chatter, other channels

More types of data in CDP

Contact data Engagement data- required to enhance and give depth/substance Purchase data- Preferential data- can buy from 3rd party. Have to mis with contact data Service interaction data- customer service department. Fix things while cooking Behavioral data- enhance and make more specific to end customer

What should I ask in a discovery? (know their capabilities and what products we have to help them)

Data capture- What data sources do the customer have? In what systems are they stored? Data unification- How do you manage deduplication? How do keep data quality high? What is your leading system for Customer Identity management? Segmentation- What are your customer personas/segments? What rules and attributes define them? Which team manages your segmentation? Activation- Which are your activation channels? Insights/AI- How do you optimize your segmentation strategy & your ROI? How do you find new potential segments?

Ingest->map-> model Step 1: Data Ingestion (part 2)

Data is first ingested from source in raw form and stored in Data Lake object Data is retrieved from source by way of a connector Connector= establishes communication between servers so data can be continually accessed Data streams assist the connectors by dictating how often and when connections should be established and also help with actually populating data into data lake object once connector gains access Ex. Cloud Storage Connector, MC Data Sources and Connector, ,etc For MC, CDP provides starter data bundles that give you predefined data sets for email and mobile (ex. Einstein engagement data) -Data sets retrieved by connectors refreshed hourly and profile data sets refreshed daily -With Data extensions you can do Full Refresh (data retrieved daily) or New/Updated Data Only (hourly) -Refresh mode is setting that determines how incoming data affects existing data in table -Full refresh- clears out table entirely and replaces with new data each time it refreshes -Upsert- select if your data set contains 1) new records only or 2) new and/or updated records only Refresh History tab is good way to validate data being retrieved at expected cadence and without errors Note: If using custom data set from like a DE, have to do modeling step yourself

Data ingestion-> data extensions SFMC data extensions to CDP

Deploy settings- -"Data Extension Extract Mode"- controls how data is exported from marketing cloud by DFU to S3 - User options: - Full Refresh (what we did for client. taking data from today and throwing it in) -Delta Extract by Date (see date fields in data) -Delta Extract by Number -Refresh Mode & Schedule- controls how and when data is imported from S3 into data lake -User options: -Upsert -Full Refresh Process 1: MC to S3 - DFU drivers extract data from MC SQL Server into Einstein AWS S3 bucket. Drivers are wrapped in automation studio activities Process 2: S3 to CDP -Process historical data (last 90 days). After setup, data streams will run on an hourly schedule Underlying automations: ex. PSMain moves system data on a daily basis and a PSMain_1 hr moves denormalized engagement data on an hourly basi -PS extended- moves data extensions in full extract mode. Runs daily -Data extension name with MID- moves data extension in incremental mode. Runs hourly Note: If you select wrong Data Extract option, you can't go back and change it without salesforce support

Trailhead Salesforce CDP Basics CDP and MC connection

Each CDP instance is allowed to be connected to one enterprise Marketing Cloud instance or a single EID, with all associated business units (each with separate MIDs)

Identity resolution summary

Entity rulesets= let you configure match rules and reconciliation rules about an entity -you can create up to 2 entity rulesets for the individual entity to test different combos of match & reconciliation rules -consolidation rate- number of unified individuals/number of source individuals A unified individual object contains info, like first name, of customer record and is created by identity resolution process. Unification is determines based on ruleset configurations CDP also creates unified contact point objects automatically, which aggregate contact points from diff data sources. CDP also automatically creates unified link objects which serve as link between Source and Unified projects. Unified link objects allow you to view source data of each unified profile and required for creating calculated insights

Ingest->map-> model Step 1: Data Ingestion (part 4) Header vs. Field Label 3 Types of Data Steps to bring in Data Source Object "As-Is"

Header label- what columns are called in raw data Field label- what columns from raw data will show up as in CDP -------------- 1. Profile- segmentable entity. When defining a population to segment on. An individual or account 2. Engagement data- time series data. Associated with a snapshot in time, behavioral events, transactions, purchases, etc -Event time field- pops up if choose engagement and means when did the event occur. Doesn't change (immutable). Ex. Purchase date. Events for that date are put into a folder which proves ingestion and query performance 3. Other data- if not profile or engagement. Product catalog data, store info, etc ---- 1. Authenticate to data source -MC or sales/service cloud or amazon S3 2. Select Data Source Object (Dataset) -Object name or file name -The same "source" can be utilized for multiple data source objects -'Directory' is an optional field in the authentication details page 3. Confirm Data Source Object details -Fields and data types, primary key, category, transforms 4. Configure updates to Data Source Object -Refresh mode, authentication, schedule

Salesforce CDP for Admins Trailhead Additional admin tasks (part 4) 5. Enhancement-> Core extensibility- sharing rules

Helps control visibility Use sharing rules to extend sharing access to users in public groups, roles, or territories for your CDP objects Objects supported: data stream, calculated insights, segment, activation, activation targets Use cases: control access of CDP config like data streams with sharing rules Sharing rules consist of 3 main components: 1) Records- whether owned by certain user or meets certain criteria 2) Users- an individual or certain group of user by roles, territories or groups 3) Access= user can be granted read-only or read & write access Admins can use sharing rules to extend access to users and public groups but NOT remove access

Understand use cases to identity solutions (listen, discuss, suggest)

If client wants to account for unknown and known data in real time. If client wants segments for delivering real time next best offer message? = Interaction studio If client wants to leverage segments for lifecycle experiences= CDP Is client after a way to activate audiences in an integrated way, across all channels? Yes =Customer 360 & IS If client comfortable to integrate today, buy both 360 and IS. If not yet and need deep integration, buy IS first Example Use Case: Listen- We're trying to resolve customer identity, across all our diff databases and systems + Discuss- Is this specific to customers in your salesforce systems and external sources? Yes = Customer 360 Audiences + Are you trying to account for both unknown and unknown data in real time? Yes =Interaction Studio + Do you want your marketing org to own and manage this? Yes =Customer 360 Audiences

Trailhead Salesforce CDP Basics Audience suppression example

Isabelle is collaborating with the Customer Support team to suppress customers with open support tickets from receiving marketing communications. Jason immediately notices this improves customer satisfaction scores.

Salesforce CDP for Admins Trailhead Steps to connecting MC and Sales/Service Cloud

MC 1, CDP Setup-> click MC under configuration 2. Enter your admin user credentials by clicking 'manage' 3. Data source setup-> click 'manage' to select data bundles to import into salesforce CDP (options include, email, mobileconnect, and/or mobile push) -> click 'start' 4. Click 'manage' to select Bus to ingest data from->click save (after configuring your data bundles, CDP creates set of automations and automation activities in MC automation studio. These automations transfer data from between the 2 products) 5. BU Activation Setup-> click 'manage' to select which of your available MC Bus to activate. Connect Sales and Service cloud: 1. CDP setup-> select Salesforce CRM under config-> click connect (option to click connect for salesforce org where CDP is provisioned or connect another org) 2. Enter user credentials

Mandatory Party Identification Fields for mapping And the mandatory relationship....

Mandatory Party Identification Fields for mapping= Party identification Id Identification name Identification number Party identification type Party Mandatory relationship = Partyidentification.party-> Individual.IndividualID

Step 2: Data modeling

Map the data streams to the data model to create harmonized view across sources -Choose which CIM objects to map or create your own -Configure identity= set reconciliation rules -Relate= define relationships between data model objects The CIM model consists of several objects covering subject areas (ex. Party, product, sales order, and engagement) Adding a new data layer that confirms to a single taxonomy so everyone understands that, for example, an offline order and online order references to "Order ID" Can add custom objects and can add custom attributes to standard objects -For example, you can use the CIM's notion of an Individual ID to tag the source field corresponding to the individual who purchased a device (one data stream), called about a service issue (another data stream), received a replacement (yet another data stream)—and then review every event in the customer journey (yep, one more data stream).

In-app unified profile: Match rules Attribute value merge rules Joined behavioral data First party identity graph

Match rules- define what individuals have the same identity. A unified profile created. - Ex. All records with same email address and match them -Attribute value merge rules- rules that reconcile conflicting attribute values and define which the unified profile should have -Ex. Unifying 2 records of known person and names are gabby and gabriella. Do you want to store gabby or gabriella? Joined behavioral data- unified profile links together all engagement data for all different individuals with same identity -happens automatically First party identity graph- unified profile can be powered using 1p IDs

Identity resolution Match rules and reconciliation rules

Match rules= link together multiple records in a unified customer profile and let you specify conditions to match and unify a record Reconciliation rules= determine how to select profile attributes in unified customer profile or basically allows you to select criteria, like last updated or most occurring, to select what data to use in the profile -which data source has most up to date info about your customers essentially

Calculated insights (part 4)

Max number of measures per calculated insights= 5 Max number of dimensions per calculated insights= 10 Create calculated insights with SQL -They will show up in segmentation canvas, under attributes and objects referenced in the SQL query or just search - For example, Drag LTV calculated insight to canvas as attribute and can add dimensions like product category. So can build segments on insights you create and use them in activation To update/delete calculated insights First, delete segment referenced in calculated insight Best practices -consistent with what you see in segmentation tab -They suggest to have different calculated insight views and group metrics together on the right object -Max of 5 measures per computed view

Course 6 Salesforce CDP Data Ingestion and Modeling trailhead

May be helpful to map out a matrix of your data sets in columns and create rows for special considerations for each Want to take inventory of all data sources you want Identify all data sets required for each data source Special considerations: -Primary key for data sets= value that uniquely identifies a row of data - Identify any foreign keys in data set (may be linked to diff data set) -If data not subject to change once a record is sent or if data set needs to incorporate updates to existing records -Any transformations like simple formulas to clean up names/row-based calculations -Review attributes (fields) coming from each data source. If same field and multiple sources, decide on which source is most trusted and set up ordered preference -Authentication details needed to access data sets -How often does data get updated

Calculated insights (part 2)

Measure and dimension columns (have to end in _c) Can have up to 5 measures and 10 dimensions -Only numeric measures allowed Supports typical aggregates like sum and max and functions like less than or and or greater than DateTime Functions will return a time stamp like the hour, day, month, year instead of just like a number. -TimeStamps stored in UTC -Timezone across CDP is picked up from org timezone Automatic refresh (every 24 hours unless specified by aggregation time granularity)

Calculated insights (part 3)

Metric retention periods affected my metrics time granularity -Ex. Hourly metric time stored for 48 hours -Daily stored for 365 days -Monthly and quarterly stored for 5 years -Yearly stored for 20 years Metrics roll=ups -Metrics data is compacted as becomes older in this process -Only max, min, avg, count, and sum can be rolled up -Rolled up metrics can be queried -Hourly rolls up to daily, daily up to monthly, monthly to quarterly, etc

Step 2: Data modeling In Data ingestion, there are starter data bundles that automatically create data streams and associated mapping relationships in data model There's email studio, mobileconnect, and mobilepush bundles

Stream name: (below streams have Daily Full Refresh and one stream only) 1, SFMC Subscriber -can't filter by BU 2 SFMC Campaign 3. SFMC Journey 5. SFMC Journey Activity Email studio bundle data streams examples: 1. SFMC Email Engagement Send 2. SFMC Email Engagement Click 3. SFMC Contact Point Email 4. SFMC Ent Profile Attribute

Step 2: Data modeling Data Models (Party, Product, Sales Order, Engagement, etc)

Party data model provides information about trading relationships, such as customer and supplier information. -Used when creating a data model -Let's you set rules for how everything relates about a person aka party and not just their contact record -If trying to build segment based on who has a driver's license in CA, but not showing up when creating the segment, it's probs bc it's not related to the Party ID in the data model Product data model provides information on a product available for sale or service. Sales Order data model provides information on future revenue or quantity for an opportunity, including product family, territory, and other information. Engagement data model provides information on interactions with a specific party, such as an email message or telephone call.

Personalization and Optimizing for Success

Personalization: Personalize each email send Providing metadata in audience batches/segment and can be used to tailor message for a specific person in a segment but in a scalable way Optimize for success: Listen to customer feedback, what's working/selling, test things, etc to help optimize Customer preferences always changing so CDP success is always ongoing

Population Containers Example

Population= number of records within your current segment. -Refreshed after publishing or doing an on-demand count Segment containers can only include 1 metric Containers= attributes can be dragged into individual containers to create AND/OR relationship logic between the attributes -Provide way to create relationships between related attributes and are used to build filter logic -When attributes placed in 1 container, your query engine looks for attributes that relate to 1 another in this way (attributes in separate containers aren't connected) Ex. If want to send an email to customers who purchased as yellow scarf, they would use 1 container with: -Product_Category is equal to scarf -Product_Description is equal to yellow -Logic: AND

Pseudonymous data

Pseudonymous data isn't quite so straightforward. It can't be tied to a known person, so there's some guessing involved. ex. hashed IP address

Step 2: Data modeling After mapping, do 2 quick checks:

Review relationships -Data model tab->click contact point email for ex.->tab over to relationships and make sure it's right Check identity resolution rules

Ingest->map-> model Step 1: Data Ingestion (part 3) Schedule and Refresh

Schedule- Data streams can be retrieved hourly, daily, weekly, or monthly -Refresh initial file immediately- field retrieved immediately upon saving data stream as opposed to waiting for first scheduled run You can also extend data set by creating additional formula fields to clean names or row-based calculations (transform). At least 36 functions so far Text manipulation-> ex. EXTRACT (), FIND (), SUBSTITUTE () Type conversions-> ex. (ABS(), NUMBER (), PARSEDATE() Date calculations-> ex. DATE (), DATEDIFF(), Logical expressions-> ex. (IF(), AND (), OR(), NOT() The required syntax to reference a field from the raw data is: sourceField['Header Label'] where Header Label corresponds to the column name in your raw data. Each data set is going to be represented by a data stream in CDP You can write in the name of the data source, specify data set you're bringing in from a source which is the Object Label and Object API Name. The data set is known as the Data Source Object -When defining the DSO, designate a field as the primary key -Primary key needs to basically uniquely define a row in data -Can combine fields to make a primary key -New formula field->CONCAT function to combine, for example, "Order ID" and "SKU" as the primary key

Data explorer tab

See your data in the CDP UI. Generally available now Allows viewing of all Data Source Objects and Data Model Objects Filtering support by Object fields By default shows first 100 rows alphabetically Limit to max 100 rows to preview data Permission to control access

Segment, Publish, and Activation

Segment= filter your data to create useful segments to understand, target, and analyze customers Publish= process of searching and building segment based on filter criteria. You can publish segments on chosen schedule or as needed Activation= moving audience segments to activation target -ex. During activation, an audience segment created in a shared data extension that can be used in journey builder

Course 8 Segmentation Segment, Activation Targets and Activations tabs

Segments tab is where you create your filtered audience segments Activation Targets and Activations are used to manage where segments get exported, for example Marketing Cloud

Setup-> Identity resolution rules

Source individuals- # of records Configure match rules here Ex. Reconciliation rule: Source sequence lets you determine priority of sources for what field should be Have to activate for rules to be applied

Trailhead Salesforce CDP Basics Cloud Information Model terminology

Subject area- business concept identified by CIM (ex. Sales order). Contains 1 or more entity groups. Ex. Party Entity group- logical grouping of related entities within a subject area (ex. Retail sales). Each entity group contains one or more entities Entity- unique object that an org collects info about (retail customers). It's like a MC data extension Attribute- unique characteristic of an entity (ex. Customer's First name). It's similar to a data extension field in MC

Which aggregation would you use if you are looking for a lifetime purchase value over $100?

Sum

Step 2: Data modeling When mapping, you can create custom data model

Tab over to custom data model-> click new custom object-> review if want to keep field names inherited from data lake object To indicate how new custom object relates to rest of data model -Data model tab->search for data model custom object you created and click it->tab over to relationships-> click new->click new relationship -To complete the relationship, you must set the cardinality. -Example given: Recall that in the RUNNER_STATS data set, the same runner indicated by MAID could run multiple times. This means that the RUNNER_STATS data set may have multiple instances of the same MAID. The multiple MAID instances tie to a single MAID, or runner, in the RUNNER_PROFILES data set though, suggesting that the relationship between the objects is N:1. You would selecs N:1 in the dropdown =Click save

The ideal C360 Customer part 2

Target existing MC customers 3-5 months out as easiest to sell as part of renewal Messaging and journeys, audience builder, advertising studio Pain points to solve: disparate data & marketing systems, always over on contacts, limited tech resources in-house Early renewal quote or swap quote (swap out existing m&J editions, contacts, etc)

The ideal C360 Customer part 3

Target existing SF, no MC customers immediately Sales, service, & commerce customers Pain points to solve:: disparate data & marketing systems, limited tech resources in-house, and lack of consolidated tech Customer 360 comes with M&J (messaging and journeys)

Individual object has to be mapped to...

The Individual object and either a Contact Point or the Party Identification object must be mapped in Data Streams to successfully use Identity Resolution, Segmentation, and Activation. When mapping, enable value suggestion for the Outdoor Interest attribute, for example, to aid with usability during segmentation. This feature will enable users to choose values from suggested set, along with type-ahead functionality instead of relying on knowing the exact values stored in this attribute.

Create a 360 view of customer

They have a Global ID and agreed upon customer name You know their preferences and consent, holistic order history, holistic case history, holistic marketing response Single source of truth for every single customer ^ 1. Bring in tons of data and unify it and identify customer Bring in all data from sales, commerce, marketing in a user friendly UI Ex. Lifetime value history, propensity to open an email, order history. Resolve known identity- use exact match on combo of email, phone number, name, etc to resolve identify. Unify it and define rules on how to merge data together into unified profile/a single ID 2. Segment customers in groups and enrich with 2nd and 3rd party data. You bring people who are similar together into segments. Drag and drop marketer-friendly experience Bridge known and unknown data. Einstein predictive intelligence 3. Activate data- making personalized marketing, messaging, social media advertising Ex. Push through advertising studio or journey builder or marketing cloud, etc. Can analyze and realize what their preferred channel is and message them there to target more accurately 4. Manage consent- opted in marketing is best practice 5. Visualize and understand data 6. Build apps and packages on existing lightning platform (ex. AppExchange)

Logic of AND/OR

To determine to use AND vs. OR, ask yourself: Am I looking for any or all? Any of these can be true= OR All of these things need to be true= AND

Segmentation steps

To help with personalization, you can also add additional attributes to your published segment in your activation channel. Let's add first name and loyalty points balance so they can use these data points in their emails. Another great feature is that they can use this newly created segment in Journey Builder. When creating a journey, the team can simply select the Salesforce CDP segment in the Data Extension Entry Source in Journey Builder. 1. Create segment- properties - -Create your segment with basic properties like name and description -Segment target= defines entity on which segment will be built ex. Segment of accounts -Can choose any entity marked as 'profile' -Publish schedule- how often should segment re-filter for individuals that meet the criteria and notify activation targets that refreshed segment is available (default is 'don't publish') 2. Attribute library -Select data mapped in data modeling/ingestion to use in segmentation filters -Direct (1:1) and related (1:many) data 3. Rule builder -Build filters to define target segment using 1:1 and 1: many data from library. Features include on the fly aggregates, filter frequency, and relative date expressions -1:1 is so straightforward like gender is equal to female where 1:many can have different expressions -Text, number, and date data types attributes can be pulled onto canvas (count is at least 1) and containers 4. Count Segment -Request a count of the segment targets that will be in your segment based on your current data ingested and defined segment filters -Upon creation, a segment will display the entire CDP segment targets in this tenant as the Count -Can request the count again as filters added, removed, updated =Count always excludes based on lack of level 1 consent (right to be forgotten or opted out) 5. Publish -Publish segment either ad-hoc (unscheduled immediate campaigns, one-time campaign, or manually decide when) or on schedule (like for ongoing campaigns), to be used in activation targets like journey builder -Publish reruns the count and creates a materialized segment -Less than 5 min for segment count -Less than 30 minutes for segment publishing to activation target -Find best time to publish time from list view (helpful fields like last publish status date time or next publish date time) -Pause a scheduled publish before editing any segment to avoid confusion

Reconciliation rules

To standardize customer data that varies among connected data sources 1. Most Frequent 2. Source Sequence 3. Last Updated (default when selecting profile attributes in Unified Individual) The processing history tab shows source count, matched count, unified count, consolidation rate

True or False: Once an attribute path has been selected in container, it can't be changed

True

True or false: Attribute options are based on the profile entities mapped in your data mode Which type of resolution rules help determine which data to use?

True Reconciliation rules

Example: match on email

Unify records that have same email address 1. Match emails if individual emails -If all emails are personal or business, use this option for highest match rate 2. Match name and email if group aliases -If records include group alias emails, such as with household or business accounts, then use Name to distinguish between people that use the same email 3. Transform to optimize -Use formulas to clean up manipulated email addresses -Ex. [email protected] When individuals come in externally, they have to be mapped to both an individual ID on individual object and contact point object and then will go into unified individual and unified contact point with GUIDs

Modern enterprise data management requirements

Velocity- (ingest) ability to ingest data at various speeds, in batch mode or streaming format from separate systems to consolidate entire first-party data set (Enterprises are selecting customer data platforms based on where the weight of their first party data resides) Variety- (transform) ability to take data in multiple models with different field names and map them in one information model to unify data Veracity- (identity) capability to use direct and advanced matching to de-duplicate various IDs and contacts to provision a rich profile Volume- (data lake) need for a highly scaled, cloud-based infrastructure that can store massive amounts of fast-moving data Value-( ex. M&J, real time interactions) ability to connect that data to systems including an open ecosystem of developers to power enterprise-wide digital transformation

Adding party identification entity...

We want to add party identification entity, but have to do it separately from the first ones We added entities to avoid raising an error due to a conflict with required attribute mapping being removed. If you add that entity along with Contact Point Email and Individual the platform won't allow you to change the Primary Key mapping after you saved the mappings -Contact Point Email entity will maintain email addresses for an individual. - Individual entity represents people in the database along with their profile and demographic attributes. -Party Identification will enable unification of the individuals using various identifiers.

Segment On individual vs. unified individual

When creating a new segment, need to decide what to segment on= defines target entity on which the segment builds (so it matters what entity has been marked as 'profile' when creating data streams/model) Segment on 1. Individuals- a specific person or customer from specific data source like MC or 2. Unified individual- a customer profile whose data has been merged based on multiple sources using Identity Resolution rules -Recommended so there's no more duplicate contacts

Step 2: Data modeling Contact Point objects

When mapping, keep in mind that fields related to address, phone number, email, or other contact points don't belong on the Individual object. There are specific reserved objects for these, namely Contact Point Address, Contact Point Phone, and Contact Point Email On the Contact Point Email object, three fields must map to get this right: -Contact Point Email Id (primary key of Contact Point Email object) -Party -Email Address To set connection between Contact Point Email and Individual objects: -Set MAID, for example, as the primary key on the Individual Object and set the Party field to MAID for contact point

Knowledge Check

Which 2 CDP use cases can be met by unifying Service Cloud case data with Marketing Cloud subscriber data? Universal suppression Customer service follow up What are three pain points an existing Marketing Cloud customer might be facing which would make them an ideal candidate for CDP? Disparate data & marketing systems Always over on contacts Limited tech resources What are the three utilization pricing levers for a CDP license? Unified profiles Segment publishes Engagement events

Data Explorer w/ calculated insights

While validation segments provide opportunity to validate relational results of data mapping, Data Explorer enables user to actually explore the records in the data source and data model objects, and calculated insights -It is a quick way to do the spot/random checks at the record level to validate the expected outcomes for calculated insights or trace down relational data that is linked to a given unified individual ID.

Trailhead Salesforce CDP Basics Discovery and Design recommended approach

Why and What= identify initial Segments and Audiences to be built, what customer experiences to enable, what business impact Taxonomy= collate necessary attributes required for segmentation, align with CIM and identify gaps, design extended data model Data audit= explore available data sources, identify required feeds, existing data relationship Profile strategy= clarify all data points to be consolidated into Unified Individual Profile, hierarchical order of resolution and reconciliation rules Computations= derive necessary calculated insights that address segmentation and activation requirements for computed measure Value= clarify KPIs for use cases, establish operational reporting, provide recommendations for ROI measures and gaining audience insights

Mapping and unifying data and Segmenting and targeting at a high level

With CDP, Core ID/ingredient to hold everything like your batches (segments) together. Via data streams or flat-file imports. Ingredients in them need to be mapped and unified and connected to a Core ID (unique identifier to execute variations of things) Keep adding data to batches to keep it "fresh" and evolve over time. Can automate process to add batches of data From master mix of customer profiles, can segment them out to use for targeting What do customers like, their tastes, past purchases, etc to know what message you send them (basic or more enhanced or personalized and specific message for ex.) Creating a system of engagement: How display and deliver drives consumption and increases desirability. Deliver message or offer through channel to be consumed by audience. Channels connected to CDP in system of engagement. Powerful only if displayed where customers are

Back to segmentation Attribute library Direct vs. related attributes

attributes associated with a selected segment target that have been mapped in data model and marked for use in segmentation (basically use the library to create segments) Attributes may be from standard data sources (engagement data from MC), custom data object, or system data. -Attributes are used to narrow down a segment to your target audiences 1. Direct attributes- have a one-to-one relationship with segment target. Each segmented entity has only 1 data point for a profile attribute -(ex. Customer data->only have 1 entry for postal code or first name) 2. Related attributes- have multiple data points. -(ex. Customer data-> related attributes would be data points that 1 person could have multiple of like purchases or email opens)

Step 2: Data modeling Engagement category

data is action-oriented and has the time tracked in association with each action. (when map a data object with this category, MUST also map the Event Time Field)

Customer data platform (CDP)

is an integrated system that simplifies and connects all the data that businesses gather and uncover about their customers. It's the marketer's brain. Acquire data, process data, store data, analytics and decisions: predict and personalize

System of engagement

the channel you market to needs to be connected to CDP. Way your consumers will consume your end product

Example of identity resolution and definitions of the 5 fields to be mapped to match on party identifier

to match on Party Identifier, data streams must be mapped to the PartyIdentification object and a relationship must be set up between the PartyIdentification object. Party ID and Individual.Individual ID. Party Identifier refers to the exact match of Identification Number, Identification Name, and Party Identification Type. The following fields must be mapped first: 1, Identification Number = This ID is compared for Identity Resolution. 2. Identification Name= This required field is used to specify the name of the ID space, for example: Customer ID, LinkedIn URL. 3. Party Identification Type: This required field is used to specify an extra level of organization. This field is required for mapping, but optional for use in Match. -For example, a Customer ID could have further classification such as CRM ID, MDM ID. 4. Party: This ID is the same as the one used in the Individual object. 5. Party Identification ID: This ID is any unique ID.

Profile Explorer tab

where you can review all data points related to this individual. Once opened, and as an example, use a previously located Customer ID value to locate profile record by Individual Id attribute