Mastering Game Product Data Analysis: From 5W2H to AARRR Metrics

Data analysis is a strategically vital part of product operations; from macro to micro analysis, mining surface data to uncover product issues is a required skill for every operator.

First, let’s look at common analysis methods:

5W2H analysis: What (what do users want?), Why (why do they want it?), Where (where do they get it?), When (when should we act?), Who (who is the target?), How much (how much to give?), How (how to implement?).

Example: (what) users want premium gear! (why) because they need to boost power; (where) gear drops from a boss; (when) we run the event during the National Day holiday; (who) all players; (how much) boss drop rate set to XX; (how) the event uses a monster siege format.

This is a demand‑translation form; for products, data must support decisions, and one element should not drive the whole. Starting from a big‑picture view, refine analysis based on overall data trends, using techniques such as comparative analysis, cross analysis, correlation analysis, regression analysis, clustering, and more.

If a game shows high download numbers but low registration rates, possible reasons include server login issues, a cumbersome registration process, or recent network failures.

If a game’s data has been solid and then suddenly drops, causes might be reduced marketing effort, user‑lifecycle limits, or competitive pressure.

True data analysis is less about the raw data and more about analytical ability; data serves as a benchmark, while analysis drives behavior and change.

Multi‑dimensional (or dimensional) analysis means product data should not be confined to the product alone; in the broader entertainment context, it must combine product, market, and user perspectives. Data analysis serves business goals, and business relies on users and the market. In short, it involves purposeful data collection, organization, processing, and analysis to extract valuable insights that optimize the product, attract more users, and increase revenue.

So how to analyze? General thinking:

Why analyze? To understand large fluctuations in paid‑user metrics such as month‑over‑month or year‑over‑year changes.

Who is the analysis target? Identify whose data is fluctuating—total paid amount, paid‑user details, etc.

What effect do we want? By analyzing paid users, find problems and solve them to boost revenue.

What is needed? Gather total paid amount, number of paying users, payment frequency, and the distribution of paying users across tiers.

How to collect? Directly query the database or have developers export the data.

How to organize? Once data is retrieved, structure reports on payment tiers and frequency.

How to analyze? Perform comprehensive analysis—correlation, user saturation, competitor promotions, possible issues in the payment system, or loss of novelty.

How to present? Identify problems such as high churn among veteran payers or low conversion for low‑tier users, then visualize with charts.

How to output? Convert the findings into a business‑value report that can persuade developers, planners, and execution teams, turning knowledge into productivity.

These steps form a systematic analysis framework; further refinement includes building user models for different groups, such as churn models, churn characteristics, recharge models, and so on.

Next, we break down common data using the AARRR model:

Acquisition, Activation, Retention, Revenue, Referral.

The figure shows basic AARRR metrics; we summarize historical data as follows:

Daily New Users (DNU): Number of users who register and log into the game each day, measuring channel contribution and quality.

One‑Session Users (DOSU): New users with only one session, indicating channel promotion quality, initial conversion, and onboarding friction.

Daily Active Users (DAU): Users who log in each day, reflecting core user base and trends across product stages; can be broken down into new‑user conversion, veteran activity, and churn.

Weekly/Monthly Active Users (WAU/MAU): Users who log in within the week/month up to the reporting date, measuring user scale, product stickiness, and lifecycle trends.

User Activity Ratio (DAU/MAU): Measures user stickiness; calculated to assess engagement and active days.

Retention: Day‑1, Day‑3, Day‑7, bi‑weekly, and monthly retention, indicating user adaptation and channel quality.

Payment Rate (PUR): Proportion of paying users among active users during the period, evaluating the effectiveness of payment prompts and conversion.

Active Paying Users (APA): Number of users who successfully paid during the period, reflecting paying user scale and system stability.

Average Revenue Per User (ARPU): Revenue generated per active user during the period, assessing channel quality and overall game earnings.

Average Revenue Per Paying User (ARPPU): Revenue per paying user, indicating paying level and trends.

Average Session Length (TV): Average time users spend in the game during the period, measuring stickiness and activity.

Lifetime Value (LTV): Value contributed by a user over their lifecycle, measuring profit contribution of user groups and channels.

Customer Acquisition Cost (CAC): Cost to acquire an effective user, aiding channel selection and marketing spend.

Return on Investment (ROI): Comparison of investment versus return, evaluating promotion profitability, channel selection, and marginal effect of paid traffic.

Finally, a common analysis method: DuPont analysis.

These points provide an overview of data analysis; a rational approach is needed because definitions and algorithms vary across companies. Avoid blind or crude analysis; combine multiple data sources, define a clear analysis plan, and adjust based on insights.

Remember that refined operational data analysis requires clear thinking; a confused mindset leads to chaotic results. Stay calm, organize your thoughts, sketch a data framework, then act—this cultivates rigorous logical analysis skills.