Google

Google's dominance didn't come from better algorithms alone - it came from organizational innovations that maintained breakthrough capacity even at scale. The 20% time policy (2004-2013) represented voluntary caloric restriction: engineers spent one day weekly on self-directed projects. This slack produced Gmail, AdSense, Google News, and Maps Transit - products worth tens of billions. The APM program (2002) operated as biological stem cells: hire brilliant graduates without specific expertise, rotate them through products for two years, then allow specialization based on fit. This maintained undifferentiated capacity that could become whatever the organism needed.

Optimization killed both. Post-2013, OKR-driven quarterly goals made 20% time impractical - if you can't measure it, you can't get promoted for it. The policy quietly disappeared. APM intake shrank as headcount growth slowed. Meanwhile, Google became the tech industry's primary talent source, with extensive hiring from Google importing its practices across competitors and contributing to industry-wide homogenization. The company's data network effects - billions of queries refining algorithms - created self-reinforcing dominance that no longer required innovation to maintain.

The lesson: breakthrough mechanisms are the first casualties of optimization. 20% time and APM programs are overhead when measured quarterly. But they're investments in adaptability. Once dominant, companies optimize what works and starve what creates breakthroughs. By the time the pipeline runs dry, it's too late - you've already trained your organization to kill anything that doesn't ship this quarter. Google ran this experiment in real time: innovation abundant when maintaining slack, breakthrough drought when optimizing for efficiency.