Security predicates in SAQL are the primary and most secure method for enforcing row-level security within a Tableau CRM dataset. They filter data at the query level based on the viewing user’s attributes (like role or a custom field), ensuring users only see rows they are permitted to see, regardless of how the dashboard is built. Inherited Salesforce security does not automatically apply, and dashboard filters are not secure as they can be modified by users.

For daily-refreshed executive reporting, replicating the necessary ERP data into Salesforce objects via a scheduled ETL process is the most robust approach. It provides a stable, performant data source for Tableau CRM datasets, allows for data transformation during the load, and avoids the performance overhead and connection issues associated with real-time queries (virtual objects or live APIs) against external systems for analytical workloads.

In Tableau CRM, creating indexes on dimensions and measures used in `group by` and `order by` clauses is one of the most significant performance optimizations for SAQL queries. It allows the query engine to quickly locate and aggregate data. While reducing data volume can help, optimizing via indexing targets the root cause of slow grouping operations. Adjusting timeouts or the dashboard layout does not address the core query performance issue.

The correct, sustainable approach is to modify the dataflow. Adding a transform step to perform a left outer join (`Augment`) between the existing data and the Custom Object on the Opportunity ID field incorporates the new discount data directly into the dataset’s schema. This ensures the data is unified, refreshable, and available for all lenses and queries. Creating a formula field for a cross-object lookup is not always feasible or performant, and dashboard blends add complexity and can impact performance.

A fundamental architectural difference is that Tableau CRM is a compute-on-query (or compute-at-visualization) system. Computed metrics, aggregations, and ratios are typically defined in the SAQL query or within the lens itself, and are calculated dynamically when the dashboard is viewed or the query is run. This contrasts with Salesforce reports where some summary logic is pre-defined. This offers great flexibility but requires a shift in design thinking.

To flatten a 1-to-many relationship and add an aggregated value (like SUM) from the child records to the parent, you use the `Augment` transform with a **Lookup** join type. You specify the child object, the join key (e.g., Opportunity Id), and define the aggregation (e.g., SUM(Amount__c)). This adds a new column (e.g., `Total_Payments__c`) to each Opportunity row in the primary data stream. A standard Left Outer Join would bring in all child rows, causing duplication of the parent records.

For complex, non-standard date arithmetic like business hour calculations that require awareness of holidays (often stored in a custom Holiday object), the most robust and reusable approach is to create a formula field on the source Salesforce object. Salesforce’s formula engine has built-in functions like `BusinessHoursDiff()` that can reference Holiday schedules. This logic is then computed once during the dataflow ingestion. Recreating this complex logic in SAQL or a dataflow transform is extremely difficult and less maintainable.

Tableau CRM supports writeback (or analytics insights) through a specific feature. This requires configuring an **Editable Grid** widget on the dashboard that is bound to a “writeback dataset.” This dataset is linked to a target Salesforce object. When users edit values in the grid, a dataflow can be triggered to write those changes back to the corresponding Salesforce records. It is not a simple inline edit of any lens and requires explicit configuration.

The recommended, supported method for migrating Tableau CRM (Analytics Studio) assets between orgs (e.g., sandbox to production) is to use **Analytics Templates**. You create a template from the source org (which bundles datasets, dataflows, dashboards, lenses, etc.), and then use either the API or the UI in the target org to import that template. Change Sets do not support Analytics components, and manual recreation is error-prone and not scalable.

Security predicates in SAQL are the primary and most secure method for enforcing row-level security within a Tableau CRM dataset. They filter data at the query level based on the viewing user’s attributes (like role or a custom field), ensuring users only see rows they are permitted to see, regardless of how the dashboard is built. Inherited Salesforce security does not automatically apply, and dashboard filters are not secure as they can be modified by users.0