Imagine you want to understand a single customer fully. Part of their story lives in Salesforce. Part lives in your e-commerce platform. Part sits in a marketing tool, and a little more hides in a spreadsheet someone exported last quarter. Each system knows a piece, but none of them knows the whole person.

Before Data Cloud can help you with any of that, it needs the pieces in one place. That gathering of data is the very first job, and the tool that does it is called a data stream.

What a data stream actually is

A data stream is a connection that brings data from a source into Data Cloud. Think of it as a pipe. On one end is a system that holds data. On the other end is Data Cloud, ready to receive it. The data stream is the steady flow between them.

This is the CONNECT step in the Data Cloud journey. Nothing else can happen until your data arrives. You cannot harmonize, unify, or activate data that is still trapped in five different systems. So ingestion is not a small technical detail. It is the foundation everything else stands on.

Data Cloud can only work with the data you bring to it. Ingestion through data streams is always step one, and it quietly decides how good every later step can be.

Where the data comes from

Data streams can pull from many kinds of sources. The common ones for a beginner to know are:

  • Salesforce objects — your Accounts, Contacts, Orders, Cases, and custom objects. These often flow in through a ready-made connection, so it feels almost effortless.
  • External systems — databases, cloud storage, or other platforms like an e-commerce or support tool, usually through a connector built for that system.
  • Files — CSV files in cloud storage, useful for data that does not have a live connection yet.
  • Web and mobile events — signals from your website or app, such as a page view, a product added to a cart, or a button tapped.

Each source becomes its own data stream. You might have one stream for Salesforce Contacts, another for website events, and another for an order file. Together they start to paint the full picture of your customer.

Batch versus streaming

You will hear two words around ingestion, and they are simpler than they sound.

Batch ingestion brings data in on a schedule. Maybe every hour, maybe once a night. It is like collecting all the day’s mail and carrying it inside at once. This is perfectly fine for data that does not change minute to minute, like a list of products or yesterday’s orders.

Streaming ingestion brings data in almost the moment it happens. A customer abandons a cart, and the signal arrives in seconds. This matters when timing changes the value of the information. There is no point knowing about an abandoned cart three days too late.

Most organizations use both. You batch the slow-moving data and stream the time-sensitive data. You do not have to choose one for everything.

What happens after the data arrives

When a data stream brings data in, Data Cloud stores it in its raw form first. It does not immediately reshape or clean it. That comes later, in the harmonize step, where data from different sources is mapped into one shared structure so everything speaks the same language.

For now, the important thing is simply that the data has landed. A field called email from one system and EmailAddress from another are both sitting safely inside Data Cloud, ready to be made sense of.

I spent many years teaching, and one thing stayed true in every classroom: you cannot help a student you have not yet met. Data is the same. You cannot unify or act on a customer you have not yet brought into the room. The data stream is the door that lets them in.

A gentle word on getting it right

Because ingestion is the foundation, small decisions here ripple outward. Choosing which fields to bring in, naming your streams clearly, and picking batch or streaming thoughtfully will save you confusion later. You do not need to be perfect on day one. You can add streams and fields over time. Start with the sources that tell the biggest part of your customer’s story, and grow from there.

When you can see your first stream successfully pulling records into Data Cloud, something clicks. The abstract idea of a “single view of the customer” suddenly feels reachable, because the first real pieces are finally in one place.

Your next step

Now that you understand how data gets in, see the bigger picture and where it goes next:

Mustafa Aksu

Salesforce developer & ISV builder focused on Revenue Cloud, Agentforce, and Data Cloud. I write from real, shipped work.

See my work → Try Configra →