Journal · 8 November 2025
Instrumentation debt in Series A apps
Series A is when an App Analytics dictionary stops being a founder’s notepad and becomes a second product that nobody staffed. Events were added to “just see.” Properties were duplicated because two contractors did not share a spreadsheet. A warehouse was bolted on. The board still asks for a single retention number.
Pipeline Nodecore’s House Commission is often this scene: a two-week document sprint that refuses to rebuild the warehouse. Founders are disappointed, then relieved. The debt is in the names, not in the pipeline brand they bought.
How the debt compounds
Every unnamed event is a permission slip for the next unnamed event. Growth hires arrive with a vendor they liked at the last company and add a parallel taxonomy “just for experiments.” Engineering treats analytics as non-blocking. Six months later, Funnel Grammar students discover two paywall events. The funnel was a collage.
Identity debt is quieter. Anonymous users, shared devices, and a CRM identifier that only attaches on iOS will make every cohort slightly fictional. You cannot hire your way out of that with a prettier dashboard.
What we ask teams to freeze
Before the next analyst hire, freeze the qualifying action for activation and the event names that feed it. Write owners. Put a kill date on anything deprecated. Stop adding experiment events that bypass the contract. That freeze is unromantic and it is the only reason a new hire will not spend their first quarter archaeology.
We do not tell Series A companies to buy a bigger platform. We tell them to stop collecting souvenirs. Retention Atlas can wait. The dictionary cannot. If you want the longer literacy, Cohort Signal is the twelve-week version of this freeze, with homework and a public readout so the rest of the company hears the limitation list at least once.