Why this matters
In Excel, you used XLOOKUP to bring a product name into every sales row. In Power BI you rarely do that. Instead you keep tables separate and connect them with relationships. That connected set of tables is called a model. A good model makes reports fast, measures simple and filters work everywhere. Modelling is the second skill area of Microsoft's Power BI Data Analyst certification, and interviewers often ask about it.
Facts and dimensions
Most business data splits into two kinds of tables:
- Fact tables record events: each sale, each payment, each training attendance. They have many rows, with numbers (amounts, quantities) and codes that point to other tables.
- Dimension tables describe things: products, branches, staff, dates. They have one row per thing and the details you filter and group by, such as Product name, Category, Region.
For a shop network:
Sales(fact): Date, Branch code, Product code, Quantity, AmountProducts(dimension): Product code, Product name, CategoryBranches(dimension): Branch code, Branch name, Region