Consistency as the Essential Path to R&D Efficiency

The speaker argues that consistency is a prerequisite for improving research and development (R&D) efficiency in large‑scale software organizations. Consistency reduces entropy, mitigates random complexity, and keeps communication and knowledge‑transfer costs low.

Three pillars are required to inject consistency into a system:

Process mechanisms – standardized workflows that align with market competition and team capabilities.

Tool platforms – infrastructure that enforces the same process rules across teams.

People capability – engineers must follow shared coding standards, design paradigms, and documentation practices.

The speaker introduces the “Dao, Fa, Qi, Shu” framework (Methodology, Rules, Tools, Practices) to operationalise consistency. The methodology (Dao) provides a common philosophy; the rules (Fa) turn that philosophy into concrete standards; tools (Qi) and practices (Shu) then implement those standards.

Key concepts discussed include:

Entropy law – without external forces, a complex system becomes increasingly disordered.

Random complexity – arising from divergent fixes, ad‑hoc processes, and unmanaged dependencies.

Engineering Productivity Competence (EPC) – a competency model with one outcome dimension and twelve process dimensions that defines “what, why, and how” for each activity.

Change List Cycle Time (CLCT) – the time from code commit to production, proposed as a north‑star metric for early‑stage efficiency improvement.

Practical examples such as code readability training, “Everything as Code”, and pipeline‑as‑code illustrate how consistency can be achieved in code, configuration, and infrastructure.

Ultimately, the speaker urges organizations to adopt a systematic, consistent approach to processes, tools, and people, using the Dao‑Fa‑Qi‑Shu model and metrics like CLCT to sustain productivity as they scale.