The Value of Data and Data Products: From Concept to Practice

Background

Since the rise of the big‑data concept in 2012, data has increasingly become the most important production material worldwide. Companies with strong data capabilities—such as robust data infrastructure, analytical skills, and a data‑driven mindset—can unlock growth.

Value of Data

Data is a mapping of the objective world, a digital representation of entities that helps humans understand, explore, and optimize the world. Like oil, data is a fundamental resource that fuels advanced industrial models (Internet, IoT) and drives human evolution.

From a business perspective, data creates value by increasing revenue, improving efficiency, reducing cost, and lowering risk.

Value of Data Analysis

Data’s value is realized only after processing and analysis. Analysis is a crucial step that extracts insights from raw data.

Limitations of Traditional Data‑Analysis Paradigm

The conventional workflow relies heavily on data analysts and follows a three‑step process: (1) write extraction code to retrieve data, (2) perform ad‑hoc analysis in tools like Excel or lightweight BI, (3) produce a report (often PowerPoint). This approach suffers from low decision efficiency, high marginal cost, and high decision risk due to analyst bias or error.

To overcome these issues, the author proposes productizing the analysis process so that the workflow and conclusions become reusable tools that empower more users to make data‑driven decisions.

Value of Data Products

Data products aim to lower the barrier for users to access data and to amplify data value. They consist of four modules: data collection & cleaning, computation & management, analysis & visualization, and mining & application.

Data products can be classified into two categories: commercial data products (offered to external enterprises) and enterprise data products (built for internal use, further divided into platform‑type and application‑type).

Core Advantages of Data Products

Amplify analysis value by enabling distributed, collective, data‑driven decision making.

Improve analysis efficiency, allowing many insights to be generated within the same delivery cycle.

Embed analytical capability as tools or documentation, preserving knowledge.

Reduce decision risk by aggregating diverse perspectives.

Lower analysis cost: once the product is built, marginal cost approaches zero.

Lower analysis threshold: intuitive design and visualizations reduce cognitive load.

Key Characteristics of Good Data Products

Accurate – reliable data sources and precise metric definitions.

Timely – data refreshed according to usage scenarios.

Comprehensive – cover all core business metrics and dimensions.

Easy‑to‑use – simple, guided workflows with continuous user feedback.

Practical Cases

The author shares experiences at JD.com, including the early use of Excel dashboards, the adoption of the JA visual analytics tool, and the development of the EasyBI agile BI platform.

As dashboard usage grew, a portal product named SMART‑A was created to provide one‑stop access to dashboards, improving efficiency and user experience. This portal inspired a generic dashboard‑portal platform used across departments.

Feedback from various teams (user growth, membership, product) highlighted the portal’s ability to quickly surface key metrics such as push volume, click‑through rates, and other core indicators.

Conclusion

Data products address the shortcomings of traditional analysis by amplifying value, improving efficiency, embedding capability, and reducing risk, cost, and barriers. The market potential for data products in China remains large.

References

Data Product Manager – Practical Advancement