Data validation overview

Audrey1
Audrey1
in Blog
Image: Data Validation

Author: Audrey Neet is a Certified Master Anaplanner and Director at Tru Consulting.

Data validation is the process in which business users ensure that the data in Anaplan ties to their source system. It's an important step in preparing a model for go-live as it ensures the data in the model is accurate and the end users are confident in the data. If a model is released to end users without having the data validated, it risks end users losing confidence in the tool and incorrect information being used for planning.

Timeline

Data validation should be done after the Data Hub has been completed and QA-ed internally and before end users enter the model for UAT. Ideally, no end users would be in the model before data validation is complete. A project may attempt to use test data for end user UAT, but this often confuses users and prevents adequate testing.

Image of the timeline

Steps

To ensure smooth data validation, I recommend the below steps:

  1. The development team builds reports in the Data Hub that can be used to review and tie out all data files.
  2. The development team works with the process owner to put together test scripts for the data validation process. These test scripts should include directions on how to access the data validation reports and what needs to be validated within the report.
  3. The process owner identifies all SME's that will be involved in the data validation.
  4. The development team copies the existing Data Hub model to create a Data Hub specific to data validation. This step is important as it ensures any integrations running in the live data hub do not impact the data validation. Data validation should never be done in an environment where data is changing.
    1. This step includes pointing the pages used for the validation reports to the new Data Hub model that was just copied over.
    2. Ideally, this Data Hub validation model will be in deployed mode so that any changes needed to the data hub can be done in the live Data Hub and synced over.
  5. The development team loads the files to be validated into the data validation Data Hub. This can be done either manually or by running an integration into the system as long as the integration is pointed to the new model ID.
    1. It's important to save off a copy of the files loaded to ensure end users and developers have the data at the point in time of the load to compare back to. Some organizations prefer to run data validation directly from a frozen environment, which has the added benefit of being able to reload the data without interfering with the validation.
  6. The development team ensures all identified users have access to the Data Hub being used for data validation.
  7. The responsible end users go into the Data Hub being used for data validation and run through the test scripts to ensure the data ties out.
  8. The developers work on any issues that arise, reaching out to other stakeholders in the organization when necessary.
  9. Once all the data has been validated, data validation is complete, and the model copy used for validation can be archived.

Conclusion

Data validation can be an involved process to manage and execute, but the effort is worth it. A thorough data validation can catch issues in the Data Hub prior to end user UAT, leading to increased confidence in the platform.

Tagged:

Comments

Categories