Skip to main content

Data Health

Combine scheduled refresh monitoring, data-quality checks in datasets/recipes, and validation of source permissions and field mappings.

Freshness Checks​

Set up scheduled app refreshes so the dashboard is always rebuilt from the latest connected data, and run them outside business hours when possible.

In practice, you should track the last successful refresh time, refresh duration, and whether all dependent assets finished in order, because CRM Analytics app refreshes can run recipes, dataflows, connectors, models, and prediction definitions together.

Missing Data Detection​

Use recipe steps or dataset preparation logic to identify nulls, incomplete rows, and unexpected blanks before the dashboard consumes the data. For dimension columns specifically, CRM Analytics can help fill missing values when appropriate, but you should still flag the original null rate so you know whether the source is degrading.

Good checks include row-count comparisons between source and dataset, null percentage by key field, and required-field completeness for fields like owner, stage, date, or account.

Troubleshooting Flow​

When a dashboard looks wrong, first confirm the source data exists, then verify the refresh completed, and finally check whether the issue is due to permissions, field visibility, or caching.

CRM data stream failures can be caused by missing field or object permissions, and refresh history is the place to inspect errors and download error details. If the underlying data is present but charts still lag, browser caching or delayed refresh behavior may be the reason rather than a broken dataset.

A useful troubleshooting sequence is:

  1. Confirm the source object or warehouse table has the expected records.
  2. Check the last successful refresh and its logs.
  3. Compare row counts before and after transformation.
  4. Validate field-level permissions and integration user access.
  5. Inspect dashboard filters, bindings, and date ranges.
  6. Test whether the issue persists in Explore or only in the dashboard.

Common Root Causes​

The most common causes of stale or incomplete CRM Analytics data are refresh scheduling gaps, permission issues, field-level security mismatches, failed joins or filters in recipes, and cached dashboard views.

If the issue only affects one widget, it is often a binding, filter, or data-shape problem rather than a full dataset failure.