Choosing and Analyzing Operational Metrics for Product Success

Common Pitfall: New operators often begin with an event‑driven plan (e.g., "create an H5 page because others have good ones") and then try to infer outcomes, ignoring the primary goal such as user retention. Without first improving retention, any additional traffic is wasted.

Correct Approach: Identify the most urgent objective (e.g., increase user retention), analyze why retention is low, and then design targeted actions that directly address the problem. Operations should be goal‑driven, not activity‑driven.

What to Focus On: Beginners tend to look at generic metrics like PV/UV for web or cumulative users/DAU for apps, assuming they reflect product health. However, these numbers often do not indicate true value for a product like an API marketplace.

Defining Meaningful Metrics: For an API store, the core value lies in API calls, not page views. An active user should be defined by actual API invocation behavior, and the definition of "active" (e.g., number of daily calls, continuous usage period) must be carefully considered.

Decomposing Metrics: Once the key metric is identified, break it down. For API call volume, the formula is:

调用量 = 调用人数 * 调用API个数 * 单个API调用量 * 持续调用时间

This shows that increasing call volume can be achieved by improving the number of users, the number of APIs each user calls, the calls per API, or the duration of usage. Operators can then focus on the variables they can influence.

Improving Specific Variables: Example – enhance API stability (monitoring, rating) to reduce downtime, which encourages users to keep calling. Complement data analysis with user feedback to identify the most impactful factors.

Data Consolidation: After selecting metrics, assess whether they are high or low and set growth targets by comparing industry benchmarks, competitor data, and historical internal data. Use tools like industry reports, ALEX, Baidu Index, and financial statements, while maintaining a consistent internal record of metrics over time.

Operational Data Types: Separate daily operational data (core health indicators) from event‑driven data (campaigns, ads). Distinguish page‑level data sources (internal categories, search, external promotion) and track each source’s contribution.

Practical Tips: If a needed metric cannot be captured, verify whether the current analytics tool supports it; if not, request a development enhancement from the engineering team.