Who this review is for
This engagement suits a product, growth or insight team with a live app, at least several months of behavioural data and a consequential retention question. Typical prompts include why two acquisition cohorts diverge, whether onboarding completion predicts later use, and whether a product change altered repeat behaviour rather than merely shifting clicks.
It is not a replacement analytics stack, an outsourced dashboard queue or a promise to prove causation from observational data.
What is included
- A 60-minute framing session to define the user unit, value event, return cadence, cohort basis and decision deadline.
- A quality check covering identity continuity, timestamp handling, event definitions, release markers and obvious tracking gaps.
- Retention curves and cohort tables using appropriate calendar, rolling or bounded definitions.
- Segmentation by agreed dimensions such as acquisition source, plan, geography, first-use path or account type.
- Behaviour sequence analysis around one or two hypotheses.
- A written decision brief, reproducible query notes and a 75-minute team readout.
Delivery sequence
1. Question and access
We narrow broad concerns such as “retention is down” into a falsifiable comparison. Your team provides read-only warehouse or analytics access, a tracking plan, release calendar and the names of people who understand product changes.
2. Data confidence
Before comparing cohorts, we test whether users and events remain comparable across the period. Material caveats are raised immediately; severe identity or instrumentation defects may make repair work more valuable than premature interpretation.
3. Cohort reading
The analysis examines shape as well as headline percentages: early fall-off, return cadence, plateaus, reactivation and the denominator behind each view. Segment cuts are chosen from the question, not produced as an indiscriminate catalogue.
4. Decision session
The final conversation distinguishes observed facts, plausible explanations and unanswered questions. Each recommendation names an owner, evidence need and next decision.
Timing, location and preparation
Allow three to four weeks from access approval. Work is delivered remotely across the UK, with an in-person London readout available by arrangement. Your data must be lawfully collected and shared under least-privilege access. Please remove unnecessary direct identifiers.
Fee
Reviews start at £6,800 plus VAT. The fixed proposal reflects the number of products, identity models, data sources, cohort comparisons and stakeholder sessions. Data engineering repairs, production tracking implementation and ongoing dashboard maintenance are excluded unless quoted separately.
Next step
Submit a short analytics brief with the user population, time window and decision you need to make. We will confirm fit and likely preparation within two working days.