Skip to main content

Validation Checks

Written by Kevin Jabbour

After you have created your query to explore the appropriate data, we recommend testing your data against the following data validation checks:

1. Dimension null checks

Ensure that there are no null values for registration_date, activity_date and any segment dimensions.

2. Monthly gross revenue check

Group by monthly granularity of date_activity and sum gross_revenue_primary. Check that the total values match your internal accounting and dashboards for monthly revenue.

3. Monthly marketing

spend check

Group by monthly granularity of date_activity and sum marketing_spend. Check that the total values match your internal accounting and dashboards for monthly marketing spend.

4. Daily active users check

Group by monthly granularity of date_activity and sum active_users. Check that the total values match your internal dashboards for DAU.

5. Check revenue only on active days

Check that no rows exist which have gross_revenue_primary and no active_users or active_payers (depending on which fields provided).

6. Check active_users >

active_payers

If both fields are provided, ensure that active_users >= active_payers. Check this at a segment level.

7. Check active_users

always smaller than

cohort_size

Check that active_users for all instances where registration_date != activity_date is less than active_users where registration_date = activity_date, i.e. cohort_size. Check this at a segment level.

8. Marketing spend only

provided by calendar

date

Check that marketing_spend is null when registration_date != activity_date as marketing_spend is provided as a calendar date time series and not at a cohort level by days since registration.

9. Active user value for

first day of any cohort

with data in

Check that day 0 (where activity_date = registration_date) has an active_user value if any rows exist with data for the registration_date in question. Check this at a segment level.

10. Check no activity data

before start of cohort

Check that no rows exist where activity_date < registration_date.

Did this answer your question?