Before starting our Proof of Concept (POC) to understand Salesforce Data Cloud (Data 360), it is important to become familiar with some of the key terms and concepts that we will encounter repeatedly throughout this learning journey.
Rather than explaining these concepts independently, I will introduce them in a logical sequence so that you can understand why each component comes into the picture and how they are connected to one another.
Don't worry if you don't get the complete picture at this stage. The purpose of this post is to build a high-level understanding. I will explain each of these concepts in much more detail with practical examples in upcoming blog posts.
Let's start our journey.
1. Connectors
The primary goal of Data Cloud (Data 360) is to bring data together from multiple systems and make it available in a unified environment.
For example, an organization may have data stored in:
- Salesforce
- AWS
- Databricks
- External databases
- Marketing platforms
- CSV or other external files
- Other Salesforce orgs or platforms
To bring data from these external systems into Data Cloud, we first need to establish a connection with those systems.
This is where Connectors come into the picture.
In simple terms:
Connectors are used to establish a connection between Data Cloud and external data sources.
2. Data Streams
Once the connection with an external system is established, the next step is to bring the data into Data Cloud.
The process of bringing data into Data Cloud is known as data ingestion.
The configuration that defines how data is ingested from a connected source is called a Data Stream.
For example, if we have customer data stored in an external system, we can create a Data Stream to ingest that customer data into Data Cloud.
In simple terms:
Connector = Establish the connection
Data Stream = Configure and ingest data from that connection.
3. Data Lake Objects (DLOs)
Now that we have ingested data into Data Cloud through Data Streams, where does that data actually reside?
This is where Data Lake Objects (DLOs) come into the picture.
A Data Stream is primarily an ingestion configuration, while a Data Lake Object represents the ingested data within Data Cloud's data lake environment.
When we create a Data Stream, Data Cloud typically creates the corresponding Data Lake Object automatically. The DLO generally uses the Data Stream's name by default.
For example:
External Customer System → Connector → Data Stream → Data Lake Object
The ingested records are stored in the corresponding DLO, where they can be explored and processed further.
In simple terms:
Data Stream brings the data into Data Cloud, while the Data Lake Object represents the ingested data.
4. Data Explorer
Once data has been ingested into a Data Lake Object, the next obvious question is:
"Did my data actually arrive?"
This is where Data Explorer becomes useful.
Data Explorer allows us to inspect the data stored in Data Lake Objects and verify the records and attributes that have been ingested.
For example, after ingesting customer data, we can use Data Explorer to check:
- Customer records
- Customer IDs
- Names
- Email addresses
- Other available attributes
It is particularly useful during development and POC activities when we want to validate whether our data ingestion is working as expected.
5. Query Editor
Query Editor provides another way to explore and analyze the data.
You can think of it as a more developer-oriented way of investigating data.
Instead of relying only on the visual interface, developers can write queries to retrieve and analyze records.
This becomes particularly useful when you want to:
- Investigate specific records
- Filter data
- Analyze larger datasets
- Validate data during development
- Perform more detailed data research
So, at a very high level:
Data Explorer → Visual exploration of data
Query Editor → Query-based investigation of data
6. Data Transforms
Once data is available inside Data Cloud, we may need to transform or manipulate it before using it for other purposes.
This is where Data Transforms come into the picture.
Data Transforms allow us to perform different types of data-processing operations.
One powerful use case is joining data from multiple Data Lake Objects.
For example, imagine that customer information comes from two different systems:
System A → Customer DLO A
System B → Customer DLO B
Now suppose we want to combine information from both sources into another data structure.
Using Data Transforms, we can perform operations such as joins and other transformations to prepare the data according to our business requirements.
In simple terms:
Data Transforms help us prepare, combine, cleanse, and transform data for further processing.
7. Data Model
Imagine that after ingesting and transforming data, we have a large table containing everything:
- Customer information
- Product information
- Sales information
- Transaction information
- Order information
Keeping everything in one large structure is generally not a good data architecture.
Instead, we would normally want to organize the data into multiple logical entities with appropriate relationships.
For example:
Customer
↓
Order
↓
Product
This is where the Data Model comes into the picture.
Data Cloud provides a data model that allows us to organize our data into appropriate objects and relationships.
We can work with the available standard data model objects and, where appropriate, extend the model using custom objects.
A well-designed data model is important because it provides a structured foundation for using the data across Data Cloud.
8. Data Mapping
Once we understand and define our Data Model, the next question is:
"How do we tell Data Cloud which incoming field belongs to which Data Model attribute?"
This is where Data Mapping comes into the picture.
For example, suppose our incoming Data Lake Object contains:
Customer_IDCustomer_NameEmailPhone
We need to map these incoming attributes to the appropriate attributes in the Data Model.
Data Cloud provides tools that make this mapping process relatively straightforward, often using a visual drag-and-drop experience.
In simple terms:
Data Mapping connects the ingested data structure with the appropriate Data Model attributes.
9. Identity Resolution
One of the most important challenges when combining data from multiple systems is duplicate customer records.
For example, suppose the same customer exists in three different systems:
System A
Arun Kumar — arun@email.com
System B
A. Kumar — arun@email.com
System C
Arun K. — +91XXXXXXXXXX
Technically, these may appear to be different records. However, they could actually represent the same person.
This is where Identity Resolution becomes extremely powerful.
Identity Resolution uses configured matching and reconciliation rules to identify records that may belong to the same individual or entity and create a unified representation.
The major advantage is that we don't necessarily need to build complex custom code for this matching process. Data Cloud provides functionality to configure and manage identity resolution rules.
In simple terms:
Identity Resolution helps Data Cloud identify and unify records that represent the same customer or entity across different data sources.
10. Calculated Insights
Now imagine that we have 100,000 customer records, and customers have made multiple purchases.
The business may ask questions such as:
- How much has each customer spent?
- Which customers have spent the most?
- What is the total purchase amount per customer?
- How many purchases has each customer made?
This is where Calculated Insights can be very useful.
Calculated Insights allow us to perform calculations and aggregations on our data.
For example:
Customer A → ₹50,000 total purchase
Customer B → ₹1,25,000 total purchase
Customer C → ₹15,000 total purchase
Here we are effectively calculating an aggregated value based on customer-level purchase data.
In simple terms:
Calculated Insights help us derive meaningful metrics and aggregated business insights from our data.
11. Segments
Once we have meaningful insights about our customers, the business may want to group customers based on specific criteria.
For example, a company may want to divide customers into:
- High-Value Customers
- Medium-Value Customers
- Low-Value Customers
The company can then use these groups for targeted business activities.
For example:
Customers who have spent more than ₹1,00,000 → High-Value Customers
The company could then target this segment with a specific promotional campaign.
This is where Segments come into the picture.
Segments allow us to define groups of customers or other entities based on specific criteria and attributes available in Data Cloud.
In simple terms:
Segmentation allows businesses to create meaningful groups of customers based on their data.
12. Activation
We have now reached one of the most interesting stages of the Data Cloud journey.
We have:
- Connected external systems
- Ingested data
- Stored and explored the data
- Transformed the data
- Mapped it to the Data Model
- Resolved customer identities
- Created calculated insights
- Created customer segments
Now the business wants to take this data or audience to another system and actually use it.
This is where Activation comes into the picture.
For example, suppose we have created a segment called:
High-Value Customers
The business may want to send this audience to an external destination for marketing, advertising, personalization, or other business purposes.
Data Cloud provides activation capabilities to send data or segments to supported activation destinations.
In simple terms:
Activation allows Data Cloud to make data or audiences available to external systems where the business can take action.
Bringing Everything Together
Now let's look at the complete journey at a high level:
External Data Sources
↓
Connectors
Establish connections
↓
Data Streams
Ingest data
↓
Data Lake Objects
Represent ingested data
↓
Data Explorer / Query Editor
Explore and validate data
↓
Data Transforms
Transform and prepare data
↓
Data Model
Organize data into a meaningful structure
↓
Data Mapping
Map incoming data to the Data Model
↓
Identity Resolution
Unify customer identities
↓
Calculated Insights
Calculate meaningful metrics
↓
Segments
Create meaningful audiences
↓
Activation
Send data/audiences to supported destinations
Final Thought
At first glance, Data Cloud (Data 360) can look like a collection of many different features and terminologies.
But once you understand how these components fit together in a sequence, the overall picture becomes much easier to understand.
The most important thing to remember from this post is not the definition of each individual term, but the overall data journey:
Connect → Ingest → Store → Explore → Transform → Model → Map → Unify → Calculate → Segment → Activate
In the upcoming blog posts, I will take each of these concepts one by one and explain them with real-world examples, Salesforce Data Cloud configuration steps, sample data, and practical POC scenarios.
