Experiment & Release Readout
An independent analysis of whether a release changed repeat behaviour, for which users, and with what limits on interpretation.
Independent analytics consultancy · London
Find the moments that make people stay.
Engine Nodecore examines retention curves, acquisition cohorts and in-app behaviour so product teams can separate a genuine habit from a noisy spike.
Flagship engagement
A retention review built around evidence, not a reporting ritual.
We audit event definitions, reconstruct comparable cohorts, inspect survival and return patterns, and trace the behaviours associated with durable use. Your team receives a concise decision brief, annotated analysis and a live readout.
Typical fit: consumer and B2B apps with live usage data, an established event stream and a retention question that has become difficult to resolve internally.
See scope and preparation
Three useful lenses
Work matched to the question in front of you
Instrumentation Quality Audit
A practical audit of event meaning, user identity, property coverage and release continuity before those defects distort a retention decision.
Retention & Cohort Review
A structured reading of who returns, when usage stabilises, and which early behaviours are associated with durable value.
How the review runs
From disputed metric to a shared reading
Frame
Agree the user, value event, observation window and decision the analysis must support.
Check
Inspect identity, event semantics, release markers and gaps before drawing a cohort.
Read
Compare acquisition dates, segments and behaviour sequences without hiding uncertainty.
Decide
Turn findings into named product questions, measurement repairs and testable follow-up work.
Client note
“The useful part was not another retention chart. It was learning that our apparent week-four drop was largely an identity merge problem, then seeing the remaining behavioural pattern clearly enough to change onboarding.”
— Head of Product, UK subscription app
Start with the knot
Send the question your dashboard cannot settle.
Tell us which users, period and decision matter. We will reply within two working days with fit, likely inputs and a sensible first engagement.
Submit an analytics brief