CRMA Academy

Build a Baseline Dataset with Dataset Builder

Use Dataset Builder to create an opportunities dataset from Salesforce data. Start from the lowest-grain object, walk relationships up to Account and User, pick fields, and let it write the dataflow for you.

Build a Baseline Dataset with Dataset Builder

Now we build a real Salesforce dataset. Dataset Builder is the original guided tool: it walks the object relationships for you, figures out the join keys automatically, and writes its instructions into a dataflow. It's the friendliest way to stand up a baseline dataset, which is exactly why I still hand it to people learning the platform.

Remember the plan from the grain lesson: this dataset is about opportunities, so we start with Opportunity and look up to Account and User.

Dataset Builder reads from the Connect/Sync layer, not from Salesforce directly. So unlike a CSV upload, the sync has to be in place first. If an object isn't synced, you won't be able to select it.

Kick it off

From Analytics Studio: Create → Dataset → Salesforce Data. This is the middle option — you're not uploading a CSV, and you're not building on top of an existing dataset (that path leads to recipes).

You'll be asked which dataflow to use. Every org ships with a default Salesforce dataflow, and you can add to it or create your own. In production you'll typically have a small number of named, scheduled dataflows and you'll know exactly what each one populates. For now, use the default. Click Next and Dataset Builder opens in the Data Manager tab.

Start from the lowest object

Dataset Builder now shows every object it can reach. Our three are Opportunity, Account, and User.

  1. 1

    Select Opportunity first

    This is the core object and the grain. Starting here is what lets you reach up to Account and User. Start with Account instead and you'd be stuck — it has nowhere to look up to.

  2. 2

    Add Account via Relationships

    Hover the Opportunity node, click the plus, open Relationships, and choose Account. It's a lookup — it attaches account fields to each opportunity.

  3. 3

    Add User (the owner)

    From Opportunity's relationships, add the User table to bring in the opportunity's owner. Now you have three connected nodes.

The grain holds: every row is one unique opportunity, enriched with the account it belongs to and the user who owns it. No opportunity is counted twice.

Notice what you didn't have to do: you never told it that Opportunity's AccountId equals Account's Id. Dataset Builder read the relationships straight from the Salesforce schema and wired the keys for you. That automatic key detection is the whole reason it's beginner-friendly.

Pick your fields

With the relationships in place, add the fields you need from each object. In the walkthrough that's roughly:

From Opportunity

Amount, Close Date, Created Date, Name, Stage, Won, Type — the core measures and dimensions, plus the two IDs it carries automatically.

From Account

Account Name, Type, Industry, and Billing State (or zip) — the firmographics you'll slice opportunities by.

From User

The owner's Full Name and photo — so dashboards can attribute and personalize.

Once you click Next, you can't come back to this screen. Forget a field and you'll have to edit the dataflow's digest node instead (more on that next lesson). So before you move on, double-check every field you need is selected. Measure twice, click Next once.

When your fields are set, choose an app (the private app is fine while you learn) and click Create Dataset. Dataset Builder pushes all of that into the dataflow and kicks it off. In the next lesson we'll watch it run — and handle the errors that new training orgs love to throw the first time.

This is a members-only lesson

Sign in free to read the full lesson, track your progress, and take the quiz.

Sign in to continue

Discussion

No comments yet — be the first to start the discussion.

Grain, Lookups & Joins

The one concept that makes or breaks a CRM Analytics dataset: grain. Understand grain, why you start from your lowest-level object, and how lookups and joins pull related data onto each row without multiplying it.

Run the Dataflow

Execute the default dataflow to build your dataset, monitor it in Data Manager, and fix the most common first-run error — objects that haven't been synced yet — by running Connect manually.