How Data Thinking Transforms Modern IT Operations

Preface

In the previous article we described how to build a data‑metric system to achieve objective presentation of subjective events. Industry analysis shows that enterprises are completing their IT digitalization journey and demanding more efficient IT operation, making data thinking a critical factor for improving IT productivity.

1. Evolution of Operations Methods and Data

Based on the scale and complexity of enterprise information systems and the application of operation technologies, we can divide the evolution of operation methods into five stages: manual operations, process‑oriented operations, automated operations, DevOps, and AIOps . Each stage expands the scope of operation output from resource allocation to capacity management, continuous deployment, proactive problem prediction, and finally to user‑experience‑driven technical operations.

Correspondingly, operation data has developed through four stages: automated operation capability, platform‑based operation capability, data‑driven operation capability, and intelligent operation capability . Data‑driven capability establishes a basic data‑ecosystem, supports a data middle platform and visualization, and enables lineage and impact analysis. Intelligent capability leverages large‑scale data, big‑data, and machine‑learning techniques to create smart strategies that extend operation data to more scenarios.

2. What Is "Data Thinking" in Operations?

Data thinking means that every operation process and result can be verified and guided by data. It brings several concrete values:

(1) Data Bridges Business‑Service Chains

Business metrics allow management to quickly spot operational issues. For example, a drop in third‑party channel traffic without system faults may indicate a channel problem, a new feature bug, or a misconfigured switch. Monitoring data points help operations focus on the most critical issues.

Data also supports decision‑making in special scenarios such as resource scaling during major events or emergency system degradation, where data‑backed evidence can determine the optimal approach.

Furthermore, data enables cost review and post‑project evaluation by comparing expected and actual benefits, guiding resource investment decisions.

(2) Operations Personnel’s Data Perspective

No data, no work. Without real‑time, accurate data, operators cannot assess system status, allocate resources effectively, or deliver high‑quality services.

Data exposes problems instantly. Bugs, resource issues, code quality, and channel performance all surface through data anomalies.

Use data, don’t become data producers. Operations teams should consume data produced by data engineers, analytics teams, and business units, matching data to operational scenarios rather than generating it themselves.

3. How Operations Teams Implement Data Thinking

(1) Build an Operations Data Ecosystem

An operations data ecosystem aggregates all company data—resource data (datacenter, cloud, CMDB), system logs, business metrics (PV, UV, conversion rates, revenue), organizational information, and documentation. The data flow includes collection, storage, processing, and analysis, ultimately delivering actionable insights.

(2) Provide Data Usage Scenarios

Key scenarios include:

Knowledge Graph: Define operations entities and relationships (products, services, clusters, servers, networks, IDC) to form a unified semantic layer.

Data Middle Platform: Consolidate resource, alarm, performance, business, log, ticket, metric, and probing data, offering unified access, service catalogs, and visualization to break data silos.

Data Visualization: Present data through dashboards and interactive components, enabling rapid problem analysis, multi‑dimensional drilling, and report generation, thus digitizing operational expertise.

Future articles will discuss advanced scenarios such as unattended changes, automatic fault assessment, and predictive failure.

(3) Cultivate a Daily Data‑Review Habit

Operators should regularly monitor both infrastructure and business metrics, maintain sensitivity to data changes, and communicate findings with development, product, and business teams. This habit builds credibility and ensures proactive, data‑driven decision‑making.

4. Conclusion

Operations cannot thrive without data; in today’s era of lean IT and value delivery, data is the foundation that turns intuition into actionable insight.