This solution addresses both of the client's goals: performance and accuracy. By creating an "Unknown" record in the Browser Table and replacing NULL values in the Site Visitor Table with the new "Unknown" BrowserID, every record in the fact table (Site Visitor) now has a corresponding entry in the dimension table (Browser). This allows the Referential Integrity setting to be changed to "All Records Match." This setting is more performant because CRM Analytics can make optimizations during query execution, knowing that a match for every record is guaranteed. This approach also ensures an accurate count of all site visitors, as those with unknown browsers are now explicitly categorized instead of being represented by a NULL key.

A. This is the current configuration. It does not implement any changes to improve dashboard performance as requested by the client.

C. While this option correctly cleanses the data by adding an "Unknown" record, it fails to capitalize on the performance gain by not changing the Referential Integrity setting.

D. Changing the setting to "All Records Match" while NULL values exist would cause data loss. The join would act as an inner join, dropping all site visitor records with a NULL BrowserID, leading to an inaccurate visitor count.

---

1. Salesforce Help: Join Data in a Recipe. This document explains the "Referential Integrity" setting within CRM Analytics Recipes. It states

"For better performance

select All Records Match if you’re certain that every key in the left source has a matching key in the right source." This supports the performance improvement aspect of the correct answer. The document also clarifies that "Some Records Match" should be used when a key might not have a match

which is the case with NULL values

confirming why option D is incorrect.

Source: Salesforce Official Documentation

"Join Data in a Recipe

" section on Join Keys and Referential Integrity.

2. Salesforce Help: Data Integration Guide - Best Practices. While not a single page

data integration best practices consistently recommend cleansing data to avoid null foreign keys. Replacing them with a default or "sentinel" value (like an "Unknown" record ID) is a standard data warehousing technique to ensure data integrity and enable more efficient join operations. This principle underpins the logic of creating an "Unknown" record to allow for the "All Records Match" optimization.

Source: Salesforce CRM Analytics Documentation

principles of data preparation and optimization.