Keeping Salesforce Data Up to Date in Tableau Cloud

The Salesforce CRM Connector only supports Extracts and does not support Live connections.
To keep data as fresh as possible, consider the following options.

Option 1: Increase Extract Refresh Frequency

By configuring an extract refresh schedule, it is possible to refresh data as frequently as every hour. It might be possible to shorten the interval further by combining multiple schedules or directly triggering the refresh jobs using the REST API. But be mindful of the Tableau Cloud site capacity when increasing the refresh frequency. Configure incremental refreshes to keep the refresh processing time as short as possible. Also note that capacity limits are not guaranteed, especially during peak periods. Scheduling is designed to start jobs within 15 minutes or less of their scheduled time. Refreshes too frequently may cause jobs to be skipped.

To further reduce refresh times, implement the following optimizations:

1. Hide unused fields: After bringing tables into the Tableau canvas, right-click any unused fields and select Hide to exclude them from the extract. (Fields used for relationships or joins cannot be hidden.)
2. Use Tableau Prep flows: Use Tableau Prep Builder or Tableau Prep Conductor to filter and join data in advance, minimizing the amount of data loaded.
3. Use Custom SQL: Use a Custom SQL query in SOQL (Salesforce Object Query Language) to restrict the retrieved data by filtering for a specific time period or specific records in the WHERE clause.

Option 2: Use Data Cloud (Data 360) Connector for Live Connections

The Data Cloud (Data 360) connector supports both "Extract" and "Live" connections. However, note that the following:

• Salesforce Platform data (such as Sales Cloud) must be integrated into Data Cloud (by creating a data stream) beforehand. To integrate real-time data, select Live as the connection type in Tableau and set the data stream frequency from Data Cloud to Streaming.
• Because queries are executed on the Data Cloud side, credits may be consumed. It is recommended to estimate credit consumption based on the expected data volume.