Field notes · 2026-05-03
Why account-type labels distort switch counts
If personal, team, and organisation accounts share events, your switch rates will lie. A short checklist for cleaner definitions.
Account switches are only meaningful when 'account type' means the same thing in product, billing, and support. Many teams reuse a single user ID across personal and organisation contexts, then wonder why switch rates look unstable.
Write a one-page glossary before any analysis. Define personal, team, and organisation accounts. State whether a workspace invite creates a new account or attaches to an existing one. Note what happens when ownership transfers.
Next, list every event that looks like a switch: type change, merge, split, ownership transfer, and plan change that also alters account entitlements. Mark which events should enter the switch study and which should stay out.
Finally, sample twenty recent journeys. If two analysts disagree on whether a journey counts as a switch, your definitions are not ready for an audit.
Clean labels do not require new software. They require agreement, documented once, and referenced in every later cohort brief.