Segment StrawberryA complete tracking-plan audit, run monthly instead of never
Which events exist in the plan and nowhere else, which are sent under two names, which sources point at a destination turned off in April. Ingestion calls wait for you.
The tracking plan says `Order Completed`. The iOS build sends `order_completed`. The web build sent both, for two weeks in March. Three dashboards downstream now disagree, and every explanation starts with somebody opening four tabs.
A companion reads every source, every destination and every tracking plan, then writes the differences out. The write side is the ingestion pipeline itself, track, identify, page, screen, group, alias and batch, and every send waits for your approval.
Two spellings of the same event
The companion lists your sources, lists the tracking plans, and reads what each source is configured to send, then lays the differences out: which events exist in the plan and nowhere else, which are being sent under two names, which sources are wired to a destination that was turned off in April.
The comparison is complete and repeatable. Run it monthly as a routine and the drift is caught in the week it happens rather than in the quarter someone finally investigates. Data and operations teams know this audit is overdue; what stops them is four hours of clicking.
What can a companion read about the pipeline?
Everything the Public API exposes to the token you connect: the workspaces on the account, the sources with their configuration, a single source in detail, the destinations, and the tracking plans. Enough to answer where this data comes from, where it goes, and what it is meant to look like.
The event data itself lands downstream, so “how many orders completed last week” is answered from the warehouse or analytics tool it flowed into. A companion working across both joins the two halves in one pass: the pipeline’s shape from Segment, the numbers from where they landed.
Why does sending an event wait for approval?
Track, identify, page, screen, group and alias calls become your analytics.
An event fired into a production source is in every downstream destination within seconds, in the warehouse, and in whatever dashboards someone built on it. There is no undo.
So each send waits for your approval, and pointing them at a development source while you work out what to send is the sane habit. The batch tool makes the size of what you are about to do visible: one confirmation, a payload you can read, a known number of events.
The diagnosis for the destination that stopped receiving
Which destination, when it was last configured, what the plan expects, and which source is sending something else. That written list is what this surface produces, and it is the part that takes a person four tabs and an afternoon.
From there the companion works the Segment console in the session you are signed into, with you watching, or hands the list to whoever owns the pipeline. Each item traced to the source or plan it came from, the ordinary data extraction discipline applied to your own plumbing.
Experience Strawberry for free
DownloadTrusted by fast-growing companies worldwide
Frequently asked questions
Strawberry is free to download and includes AI credits to start. Paid plans begin at $20/month. See pricing. · Reviewed · Canonical facts for AI agents