Skip to main content

Optimal Data Specifications

To get the most value from Customer Analytics, we recommend sending held data for all users, refreshing your data daily, including key demographic information, and providing 6–12 months of data. Here’s why:

  • Sending held data for all users, not just a subset. Limiting held data to a subset of users (for example, digital only) leads to less impactful analysis. Analyzing held data for all users enables us to give you results that are based on the entire scope of your business rather than just one part of it.

  • Refreshing your data at least daily. Data changes rapidly. When you refresh your data daily, we can present you with more accurate analyses about things like conversions on Messaging campaigns. To learn more about sending held data, refer to Sending Held Data.

  • Including key demographic fields. The more information, the better. Include these fields for more precise analytics; note that some fields apply only to customer relationship management (CRM) integrations:

    • Date of birth. Knowing the user’s age enables us to provide analyses based on age-related life events like retirement, buying a home, or marriage, empowering you to recommend the right service at the right time.

    • ZIP code. When we know the location in which money is being spent, we can identify when a user might be moving or traveling, giving you the advantage of knowing when to extend relevant offers such as relocation assistance or rewards cards.

    • Alternate ID (CRM integrations only). Users may have alternate identification numbers in addition to their external GUID. Including all of their IDs enables us to match all of their records between the Customer Analytics and CRM platforms, ensuring comprehensive analytics for them.

    • First and last name (CRM integrations only). Having the user’s name means we can personalize their experience. It also means we can add their name to their profile if it’s missing from the CRM.

    • Email address (CRM integrations only). Similar to how we use alternate IDs, we use email addresses to match a user’s records from the CRM to Customer Analytics, ensuring all their records are associated with them.

  • Providing 6–12 months of data. For optimal prediction analysis, provide 6 months of data for life events such as retirement, job loss, or marriage, and 12 months of data for attrition events such as how many users might leave for another institution. We can still calculate features with smaller datasets, but the prediction performance is suboptimal.