Optimizing Demand Analysis Process for Youku Mobile App: An Agile Coaching Case Study

Author bio: Zhang Yinghui (Wenju), Alibaba Agile Coach, has supported teams such as Taobao Live, Youku, Intelligent Marketing Platform, and has guided agile theory implementation across Alibaba’s departments and the RDC R&D collaboration platform.

Background and Goal

In December 2016, together with the Mobile Taobao PMO, I joined the “Wind Forest Fire Mountain” project to help Youku quickly integrate into Alibaba’s R&D system, focusing on optimizing the demand analysis process. This article briefly describes how research, design, implementation, evaluation, and continuous improvement formed a closed loop to solve Youku’s problems.

To understand Youku’s situation, the PMO and I interviewed product, design, development, testing, and project manager roles from the Youku core team. They reported the main pain points in the demand analysis stage:

Target setting for demand analysis process optimization:

Provide a lightweight, standards‑heavy, data‑driven demand management solution that drives team improvement through metrics and offers transparent end‑to‑end management from creation to release.

Design Solution and Implementation

Youku’s core team already had a demand analysis process with priority ranking and review mechanisms. Based on the feedback and the characteristics of the Youku mobile app, and borrowing from Taobao’s experience, I designed an improved demand management scheme.

Bi‑weekly Iteration Timeline (applicable to regular iterations, not special projects)

Differences Compared with Existing Scheme

Add product planning phase

Each quarter a product planning meeting is held, with business owners attending. The agenda includes reviewing the previous quarter’s business data and goal achievement, and planning the next quarter’s business objectives and tactics. Core demands for the next three months are planned and designed around these goals.

The main purpose is to solve the “unclear planning” pain point, creating top‑down alignment and focusing on business objectives.

During demand grooming, interactive sketches are required. TLs and key stakeholders attend the grooming session. Before the meeting, product members enter the demand into Alibaba Cloud RDC and provide a demand outline; the product team aligns on priority. In the meeting, product members present demands by priority, and the audience asks clarifying questions. The outline must at least clarify demand value, technical feasibility, and main interaction flow.

The goal is to get product owners to move forward early, communicate early, and avoid one‑sentence or oral placeholder demands.

From centralized TL‑only review to distributed review involving frontline members

After the grooming session, TLs decide the delivery scope and assign personnel, updating the information in Alibaba Cloud RDC. Product members invite relevant frontline members and key stakeholders to organize their own demand reviews. TLs attend critical reviews; frontline members attend reviews relevant to them.

Youku’s functional teams consist of 102 people across product, design, development, and testing. Previously, frontline members did not participate in reviews; TLs relayed information. By moving from large meetings to small, focused sessions, frontline members participate in demand design and discussion, understand the demand better, raise issues and risks early, and TLs can focus on critical demands.

Demand must meet start‑work criteria before entering development

The start‑work criteria are drafted by the Youku product team and approved by TLs. They are configured into the Alibaba Cloud RDC demand template; new demands only need to fill the template.

All demands unified in Alibaba Cloud RDC

Tracking the entire demand lifecycle in RDC is the prerequisite for data‑driven management. A unified tool also strengthens collaboration and reduces communication cost.

Effect Evaluation

After the solution was rolled out in January 2017, I interviewed two product members, one designer, and one developer from the Youku core team, and organized a version summary meeting on Jan 20, 2017 with TLs and frontline representatives.

Key feedback from interviews and the summary meeting:

Improved sense of rhythm

Timing points are clear and each stage’s deliverables are well defined.

Simpler, clearer process, avoiding redundant steps

Frontline staff understand demands better and spot issues early; distributed reviews improve efficiency; product moves forward earlier, reducing backlog.

Tool introduction increased efficiency

RDC is convenient, improves information sync; the unified demand management tool boosts efficiency; linking demands with bugs aids problem location.

Continuous Optimization

After the demand process optimization was implemented in the Youku core team, other teams and departments also requested help. From February to March 2017, I assisted other Youku product technology teams and the advertising team in rolling out the new demand analysis process: demands are unified in Alibaba Cloud RDC, achieving transparent end‑to‑end management.

Example of core metrics for the Youku product technology line (four report images below):

(Note: For data security reasons, the clear version of the data is not provided.)

The four reports cover demand count, total lead time from creation to release, newly created defect count, and average defect close time for the period March 1‑31, reflecting quality, efficiency, and responsiveness. After the reports were generated, business teams analyzed them, identified problems, and took improvement actions.

Examples

Case One

A team found that the demand analysis stage was unusually long. Investigation showed that some demands were ready, but development capacity was full, so they waited for scheduling, and the waiting time was counted as analysis time. The team added a scheduling status after analysis, making the workflow reflect actual work and helping identify bottlenecks.

Case Two

In March 2017 a business line delivered 16 demands but introduced 1,030 defects. The team identified two main reasons: overly large demand granularity and inconsistent understanding of demands among testing, product, and development. Improvement actions included splitting demands into appropriate granularity and involving testers in demand reviews to ensure a shared understanding.

Conclusion

As a member of the agile coaching team, I strive to translate industry best practices into Alibaba‑specific R&D models, helping teams improve quality and efficiency, capture success cases, and embed them into tool platforms.

Guided by agile principles, we helped the team establish a stable iteration rhythm and used transparent process data to drive continuous improvement. In the functional‑based Youku team, we avoided rigid constraints, prioritized product and time slicing, and after three months of stable bi‑weekly iterations, a cross‑functional team prototype emerged, laying a solid foundation for a feature‑team transformation.

Additionally, I closely collaborated with the RDC product team to continuously improve reporting functions, providing convenience for other teams’ ongoing optimization.

Click “Read original” to experience the same demand management functionality as Youku!