This is likely due to how Looker (and most analytics tools) calculates unique Users and Sessions when aggregating over different date ranges. When you query a full year, the Users metric is deduplicated across the entire year, so a user active in multiple months is only counted once. When you query just a single month, deduplication happens only within that month, which can lead to different totals. The same applies to Sessions if your session definition spans midnight or overlaps date boundaries. To confirm, check your LookML model for how these measures are defined and consider using a daily granularity export with post-aggregation in your analysis tool to ensure consistent counts.