How to Accelerate Effective Innovation Loops in Internet Product Development

He Mian, author of *Lean Product Development* and member of Alibaba R&D Efficiency Division, presented the theme “Accelerating Effective Innovation Loops, Empowering Internet Product Innovation”. He began by contrasting waterfall and agile development, noting that agile’s iteration is still limited to the engineering phase and does not fully address business innovation.

He argued that product delivery is not a linear path from requirements to testing; it starts with product planning and continues after testing with validation and implementation. Limiting iteration to the development stage leaves a larger “waterfall” at the business level, preventing true acceleration of feedback and value creation.

To truly speed up innovation, the entire value stream must be visualized—from business goals through product discovery, engineering, and verification—so that bottlenecks become visible. He illustrated this with a story about a drunk searching for his keys under a streetlamp, emphasizing that looking only where the light shines (the obvious) misses the real problem.

Referencing Don Reinertsen’s *The Principles of Product Development Flow*, he highlighted that the real constraint is often stagnant demand, not idle resources. Accumulated demand at any stage (analysis, testing, release) blocks value flow.

He introduced the practice of “lighting up the critical points” by visualizing the end‑to‑end demand flow, using Kanban boards that show demand cards (blue) and their decomposed tasks (yellow) across lanes for front‑end, back‑end, and external modules. This visualization reveals where demand stalls and where capacity is limited.

Two key cycle times were defined: the delivery cycle (from demand selection to release) and the development cycle (from ready‑to‑develop to deployable). Shortening these cycles indicates smoother demand flow.

He emphasized building in quality (built‑in quality) by defining clear input standards for each stage—e.g., well‑defined user flows, identified dependencies, and breaking large demands into two‑week or one‑week items.

Metrics are essential: measuring delivery and development cycle times, release frequency, lead time from code commit to production, and defect trends. Proper metrics guide teams to focus on the right goals and avoid misleading KPIs.

The talk also covered product growth engines—stickiness, paid acquisition, and virality—explaining that a strong stickiness engine must precede the others to ensure users stay long enough to be monetized or become advocates.

He presented a mapping from “why” (business goal) to “how” (levers) to “what” (features), stressing that the “how” is often the hardest part and must be continuously validated through feedback loops.

Finally, he summarized the acceleration loop: illuminate problems, ensure smooth value flow, focus the team on key objectives, align features with those objectives, and close the feedback loop to continuously improve product innovation.