The blanks in the Decomposition Tree visual typically represent null or missing values in the dimension field being used for analysis (in this case, income level from the DimCustomer table). Even though there are no orders without a CustomerKey in FactInternetSales, the income field (YearlyIncome) for some associated customers might still contain nulls, causing orders from those customers to appear under the "(blank)" category.To confirm:
Run a query like SELECT COUNT(*) FROM FactInternetSales f JOIN DimCustomer c ON f.CustomerKey = c.CustomerKey WHERE c.YearlyIncome IS NULL to count the affected orders (or distinct SalesOrderNumber if your measure is a distinct count).
If that returns 170, that's the source.
The 600 customers without orders are unrelated here—they don't contribute to the visual because there are no orders from them to decompose. The visual only breaks down existing orders (totaling 25,164), so the blank specifically ties to orders where the customer's income is unspecified. If you've confirmed no nulls across the entire dataset (including YearlyIncome), double-check the data model relationships or any filters/bins applied to the income field. If nulls exist in YearlyIncome, consider replacing them with a placeholder like "Unknown" in Power Query to avoid the blank category.