How Zhongyuan Bank’s EPG Team Revolutionized DevOps with a Dual‑Mode, Multi‑Speed Delivery Model

Event Overview

On October 28, 2022, the GOITI 2022 IT Technology Leadership Awards were held in Shanghai, organized by the GreatOPS and DevOps Era communities. The ceremony awarded the Zhongyuan Bank EPG Process Improvement Team with the 2022 Banking QA Star Team Award.

Team Background

The Zhongyuan Bank EPG Process Improvement Team, established in May 2020, coordinates the bank’s R&D delivery management system, efficiency insight system, and process‑improvement mechanisms, aiming to empower data‑driven R&D efficiency and build an intelligent development management system.

Construction Plan

The team follows a “standardization‑online‑intelligence” roadmap, establishing a professional, refined, intelligent R&D management system and a continuous improvement loop.

1. Standardized Delivery

They designed a dual‑mode, multi‑speed delivery model that combines steady‑state lean and agile‑state iterative modes with four delivery speeds (armored vehicle, off‑road vehicle, ambulance, race car), supporting various project scenarios such as traditional large projects, agile iterations, and urgent demands.

2. Online Management Platform

The “Original Effectiveness Platform” integrates requirement management, development management, testing management, continuous integration, continuous deployment, and performance measurement into a one‑stop R&D‑Ops collaboration tool, enabling end‑to‑end online management and quality assurance.

3. Intelligent Performance Metrics

A comprehensive performance insight system measures both result and process indicators across the full lifecycle, providing multi‑dimensional data models for organization, team, and individual analysis, and supporting data‑driven improvement decisions.

4. Closed‑Loop Process Improvement

A data‑driven improvement mechanism at both organization and team levels follows a five‑step cycle (discover‑diagnose‑plan‑improve‑report) to continuously enhance delivery efficiency.

Challenges

Scaling the dual‑mode delivery framework across thousands of staff required extensive training, online governance, and QA checks. Data quality issues in performance metrics were addressed through unified definitions, standardized collection, and automated validation mechanisms.

Key Achievements

The dual‑mode, multi‑speed delivery model has been adopted by over 160 teams, supporting more than 1500 staff and 300 systems, and delivering over 3000 business demands annually. The performance insight platform reduced product iteration cycles from two months to two weeks, improved delivery quality by approximately 38%, and cut security vulnerabilities by over 30%.

Industry Recognition

The team earned multiple certifications, including three‑level DevOps continuous delivery assessment, advanced DevSecOps ratings, and industry‑leading evaluations for its performance measurement tools.