Funnel analysis notes
A record of conversion changes across key steps, with the paths or groups that merit follow-up, for the people framing the product question.
User growth · Funnel analysis · Experiment design
Funnel analysisProduct instrumentation andexperiment design
I’m Kevin Zhao. My current responsibilities are to monitor funnel data for a core product line, process product event-tracking data, and design A/B tests. The results support product iteration and decisions.
User growthFunnel analysisExperiment designProduct instrumentationSQLPython
About
A typical working context
My day moves between the product itself, event data, and the analysis that follows.
A funnel movement can sit beside a tracking change in the same piece of work. I may need to review the product steps while organising event fields, firing rules, and measurement definitions.
Experiment work enters at a different point: turn a product idea into a test plan, define the measure, use SQL and Python to work with the data, and prepare the result for a product decision.
Current work
Senior Data AnalystShanghai Weiguang Network Technology Co., Ltd. · Full-time
Education
East China Normal UniversityApplied Statistics · Master's degree · 2018–2020
Work
What I deliver
The work produces three kinds of material for product discussions, data records, and iteration decisions.
Instrumentation definitions
Event fields, firing rules, and measurement definitions for the people working on product changes and data collection.
Experiment plans and results
The hypothesis, primary measure, and organised result for the people deciding whether a product iteration should continue.
Notes
A question worth separating
I use each note to examine one disagreement from product analysis rather than repeat a description of the role.
Retention and repeat purchase need different cohorts
Defining populations, starting events, windows, and mature observation periods for activity retention and repeat purchase.
Read the noteUseful segments are not the smallest segments
Choosing actionable, interpretable user groups with stable samples, starting from the analysis question.
Read the noteA pre-launch checklist for an A/B test
Reviewing decisions, assignment, measures, duration, and stopping rules before a product experiment begins.
Read the note