One‑Stop R&D Collaboration Platform Construction and Practice at Hello

On October 26, the DevData Talks series invited Gao Mingguo, head of the R&D Efficiency Team at Hello Mobility, to share his experience in building a one‑stop industry‑product‑research (业产研) collaboration platform. The talk, titled “One‑Stop R&D Collaboration Platform Construction and Practice”, described how a series of efficiency solutions and best practices helped the company reduce costs and increase productivity.

The presentation highlighted four main sections: the challenges faced, the construction方案, the practical implementation, and future outlook.

Challenges

The company faced both management and quality/efficiency dilemmas. Management issues included difficulties for the corporate management team, business teams, and project management teams. Quality and efficiency problems were categorized into production‑research collaboration, delivery team, and efficiency‑measurement challenges.

Construction方案

A three‑step solution was adopted: standardization, online enablement, and digitalization. First, processes were standardized and then encapsulated as capabilities within an online system, allowing data to be collected and measured. Second, strategic alignment was achieved by defining business groups, strategic planning, goal linkage, and force‑monitoring to focus resources and monitor anomalies. Third, the platform supported project collaboration, distinguishing between “campaign projects” (large, strategic initiatives) and “non‑campaign projects” (regular support and bug‑fix work).

Key activities included:

Business grouping

Strategic planning

Goal linkage

Force monitoring

Practical Implementation

In fine‑grained operations, the team improved organization efficiency, project collaboration, continuous delivery, engineering capability, and production quality. Metrics such as release success rate (rising from ~94% in 2021 to ~97% in 2023), reduction of emergency releases and rollbacks, and various efficiency indicators (test‑delivery time ratio, per‑person demand delivery, demand participation) were tracked.

Quality control measures included a unified change‑issue reporting mechanism, rapid 5‑minute production issue aggregation for quick rollback, SLA‑based fault detection, and systematic post‑mortem improvement.

Summary

Key recommendations for effective measurement:

Data credibility: Metrics must be based on accurate, reliable data.

Talent availability: Ensure the organization has the necessary skills to implement AI or big‑data‑driven efficiency solutions.

Feasibility of measures: Proposals should be actionable rather than abstract.

Future Outlook

Having spent two years on cost‑reduction and efficiency improvement, the team plans to expand the platform’s scope, deepen business collaboration, and conduct more data mining to continuously identify and resolve process and efficiency issues.