Hey everyone! I'm new to the data side of our marketing team and I've hit my first big snag. 😅 We're comparing a core metric—let's say "monthly active users" from a specific campaign source—between our Snowflake data warehouse and our CDP (Segment). The numbers are different, sometimes by like 5-10%, and it's causing a lot of confusion in our reporting.
Can someone help me understand where to even start looking? I know the basics of both systems, but I'm not sure how to trace the discrepancy step-by-step. Our company is B2C SaaS, mid-market size, with around 2 million monthly website visits.
Here’s what I *think* I know so far:
* Our data warehouse gets data via dbt models that transform raw event data.
* Our CDP (Segment) is the source for a lot of this raw data, but it also does its own processing.
* The metric definition *seems* the same on both sides ("MAU from campaign X"), but maybe it's not?
I'd love a beginner-friendly walkthrough on how to debug this. Specific things I'm wondering:
* What are the most common reasons for these kinds of discrepancies?
* Should I check timestamps, deduplication rules, or the exact event definitions first?
* Are there tools or simple SQL queries you'd recommend to compare the two data streams?
Really excited to learn from your experiences. Even pointing me to a good guide or checklist would be amazing!