Finding 44K members missing from the CRM

Membership counted from engagement came to about 112K people, while the CRM said about 68K. Both numbers fed reporting, and nobody could say which was right or why they differed.

I helped move the CRM to EveryAction, working on data integration, Snowflake reporting and data governance across teams. On the membership gap, I traced the cause, quantified the impact and set priorities for the fix.

1. 1.Define

   Write down what counts as a member in each system, including the date window.

2. 2.Count

   Count members in each system with its own definition.

3. 3.Join

   Match people across systems and sort every person into: in both, only in one, or in the other.

4. 4.Trace

   Follow the people missing from the CRM back through the pipeline to where their records stopped flowing.

5. 5.Prioritize

   Quantify the impact and agree the order of fixes with the teams that own the systems.

• Data integration and Snowflake reporting for the move to EveryAction, with data governance across teams.
• A person-level reconciliation between engagement-based membership and the CRM’s count.
• The root cause: a disabled Adestra-to-Salesforce sync, so new members stopped reaching the CRM.
• An impact estimate and a prioritized fix plan.

Reconcile people, not totals

Two totals can match while the people inside them differ. Sorting every person into in-both, only-here or only-there shows where the gap really is.

Agree the definition before arguing about the number

Both counts were right by their own definitions. Writing the definitions down first turned an argument about numbers into a question about data flow.

• Each count was correct under its own definition, so the first step was separating definition differences from missing data.
• The cause wasn’t in either system’s reporting but in the connection between them, so the trace had to follow records across systems.

• The gap was explained person by person, so the size of the fix could be checked against the people it should bring back.

• Described without employer data: no records, queries or system internals.

Most RevOps and data roles spend real time on exactly this: two systems that disagree, a migration, and a sync that quietly broke. Finding the root cause and getting it fixed is the job.

Snowflake, SQL, EveryAction/VAN, Salesforce NPSP, Adestra