How Tencent Game Ops Evolved: From Standardization to Intelligent Operations

From the "Four‑Transformation" of Operations Services

Last year, Tencent Game's operations leader Liu Qitong described the next‑generation game operations model based on four pillars—standardization, automation, service‑orientation, and productization—to improve service planning, tool building, user experience, and product decision support.

We share how Tencent Game operations have been implemented over the past years, progressing through three stages:

Stage 1: Focus on standardization;

Stage 2: Emphasize service‑orientation;

Stage 3: Move toward intelligent operations.

Pre‑Transformation Situation

QQ Game hosted hundreds of mini‑games with millions of daily active users, generating frequent change requests that overwhelmed a small five‑person operations team, leaving the team passive and unable to handle the load.

Multi‑Role Division in Operations

To address this, the team restructured operations into four distinct roles:

Operational Execution: Handles repetitive tasks;

Business Operations: Focuses on basic operational work;

Planning Operations: Manages project planning and requirement design;

Development Operations: Develops operational tools.

Collaboration Among Roles

Business Operations define repeatable tasks as atoms ;

Development Operations build platform pages for atom execution, turning daily work into UI‑driven actions;

Planning Operations compose these atoms into flexible workflows;

Operational Execution solidifies the processes and operates the tools.

Through continuous tool building, automated execution now exceeds 90% of requests, freeing operators from repetitive work.

The Challenging Path of Standardization

Early standardization faced issues such as limited BlueKing platform APIs, high refactoring costs for legacy games, and initial serial processing performance.

After six months of adjustments, the platform matured, offering richer interfaces, flexible steps, and customizations, enabling standard integration for all basic Tencent Game operations.

Standardized tools have executed over 20,000 tasks across publishing, server scaling, and custom scenarios, eliminating waiting and incomplete tasks, equivalent to saving 700+ workdays.

How Service‑Oriented Operations Were Forged

Service‑oriented operations shift from reactive to proactive problem solving, driven by data and closed‑loop services.

Using DNF (Dungeon & Fighter) as an example, online recovery time after a major release dropped from 12 hours in 2013 to under one hour in 2022 through continuous improvements.

Key improvements included:

Optimizing patch download speeds via CDN and pre‑download mechanisms, raising average per‑user speed from 157 KB/s to 1.2 MB/s;

Refining strategy deployment using active‑time distribution metrics for precise targeting;

Real‑time monitoring of player disconnect rates and rapid bug fixes before peak periods.

The team categorizes service scenarios—login, download, version, user experience, and cost—and defines clear metrics for each, forming a reusable framework for other businesses.

Exploring Intelligent Operations

Vision: "Leader of Game Intelligent Operations"

With standardization and service‑orientation in place, Tencent Game now leverages abundant business data to enable intelligent strategies, aiming for fully automated, closed‑loop operations.

Key considerations for intelligent ops include:

Collecting and cleaning business data to identify core signals;

Applying complex scenario analysis for automated decision making and precise push without human intervention;

Ensuring all services execute in a closed loop;

Using big‑data insights to influence product decisions while balancing cost and service quality.

Conclusion

Tencent Game's journey from standardization to service‑orientation and now toward intelligent operations demonstrates how systematic tooling and data‑driven practices can dramatically improve operational efficiency and business value.