APD Agile Improvement Case Study of a User Product Center Team

After the previous APD (Agile Product Development) case sharing with a dealer team, many colleagues requested more examples; this article introduces a user product center team's APD agile improvement case, hoping the practical experience can help teams facing similar challenges.

1. Team situation before improvement

The team consisted of 1 product manager, 6 developers, and 2 testers, primarily responsible for implementing three‑end automotive business features.

The main problems before improvement were:

Frequent insertion of ad‑hoc tasks after weekly scheduling, disrupting development rhythm.

Unreasonable task breakdown, causing some tasks to be untestable after completion.

Excessive and high‑retention bugs.

Developers tied to fixed business modules, lacking flexibility to respond to multiple product demands.

Unclear collaboration among product, technology, operations, and external business units.

2. Team improvement stage results

The improvement started in April 2018 and, after six months, the team achieved stable, staged results.

Key improvements include:

Daily scheduling and a standardized morning meeting better handled external changes.

Task granularity reduced to less than two days, making tasks independently developable and testable.

Strict TDD execution solved the high bug count and omission rate.

Team capability became full‑stack, allowing members to handle diverse needs and work autonomously.

Figure: Throughput

Figure: Demand delivery cycle

Figure: R&D process efficiency

Figure: Bug count

Metrics after the staged improvement:

Monthly throughput increased from about 50 to around 100, more than double.

Demand delivery cycle shortened from 5 days to 2 days, a 60% improvement.

Monthly bugs reduced from an average of 50 to about 2.

3. Team improvement insights

The team's pain points were frequent insertion of product demands that disrupted development rhythm, leading to slow delivery and poor quality. By analyzing the three APD R&D models, we introduced the APD stream development model to enable fast, high‑quality delivery.

Key practices applied from the stream model and APD practice library include visual value stream, explicit process rules, WIP (Work‑In‑Progress) control, bug‑board visualization, workflow management, and an evolved TDD approach.

Visual Value Stream , which makes workflow, tasks, and scheduling information visible in real time, reducing reliance on verbal communication and improving team communication efficiency and accuracy.

Explicit Process Rules , defining DOD for each stage to ensure effective execution of the R&D process and avoid inefficiencies caused by unclear standards.

WIP Control , preventing the start of too many tasks so team members focus on current work, deliver quickly, and then start new tasks, achieving rapid delivery and early user value.

Bug Board Visualization , managing the full lifecycle of bugs from creation to resolution, monitoring resolution time to encourage rapid fixing and improve quality.

Workflow Management , using a daily 10‑minute stand‑up to confirm startable demands, sync progress, and identify bottlenecks, interruptions, or blocks for efficient management.

TDD , an evolved test‑driven development where test cases are written before development starts, reviewed across product, R&D, and testing roles, and used as acceptance criteria, improving requirement quality, understanding, and significantly boosting delivery quality while shortening cycles.

In addition to these common practices, the team also employed daily backlog tracking, risk assessment mechanisms, and weekly technical lead meetings to continuously improve and sustain fast, high‑quality delivery.