Mid‑Platform Strategies: Lessons from Supercell, Zero‑Cost Reuse, and Hidden Risks

Around 2015 the term “mid‑platform” (中台) quickly became popular in the Chinese Internet industry. Many large tech companies launched mid‑platform strategies, but by 2024 the hype has faded. This article reflects on why the trend rose and fell, using Supercell’s success as a case study and discussing the benefits and pitfalls of heavy reuse.

1. Supercell’s Miracle – In 2015 Supercell’s games such as Clash of Clans and Boom Beach achieved massive success. The company generated $1.5 billion in net profit with fewer than 200 employees, organizing development teams (cells) of about 7 people. Their advantage stemmed from a strategic rule: all low‑level capabilities must be built on a shared, reusable platform, minimizing migration costs and maximizing incremental value.

2. The Essence of Mid‑Platform: Zero‑Cost Reuse – The article uses two analogies: a Tesla factory scaling production and a developer needing a thread‑safe map. In both cases, the “raw material” is reusable infrastructure that should be adopted without extra cost. High reuse reduces labor costs and accelerates delivery.

3. Hidden Risks of Reuse

Reusing a library is cheap; reusing a mid‑platform often requires extensive communication, integration effort, and creates strong dependencies that can outweigh the benefits.

Conway’s Law: organizational structure influences technical architecture. Large, separate mid‑platform teams may have goals misaligned with business units, leading to slow response and “boundary” friction.

Latency: mid‑platforms are not directly exposed to market feedback, causing delayed reactions that can make the platform obsolete before business needs are met.

4. Some Ideas – The Supermarket Analogy – The author compares a mid‑platform to a supermarket manager deciding whether to add a new product (e.g., imported crabs). Questions about cost, lead time, and customer retention mirror the challenges of adding new capabilities to a platform. The core dilemma is whether business should drive the platform or the platform should drive business.

5. Change and Constancy – The only constant is change. Mid‑platform construction works best for stable, low‑change domains where reusable components can be abstracted. Successful platforms require deep domain expertise, strong modeling ability, and a clear understanding of the intersection between business and technology.

In conclusion, the article invites readers to share their own mid‑platform experiences and join the discussion.