Reed Hastings

[ The move ]

In 1997, Reed Hastings co-founded Netflix as a DVD-by-mail subscription service. The competition was Blockbuster: 9,000 stores, $5B revenue, infinite optionality. Netflix's bet was that mailing customers a flat-rate subscription with no late fees would scale better than 9,000 retail leases. By 2002 they were public.

In 2007, with DVDs at 90% of revenue, Hastings split Netflix into two: a streaming service and a DVD service later called Qwikster. The market hated it. Stock fell from $300 to $63 in three months. He reversed the Qwikster spinoff, kept the streaming bet, and absorbed the bruises.

In 2013, with House of Cards, Hastings made the next bet: Netflix wouldn't just license content, it would produce it. Hollywood thought streaming was a distribution medium for someone else's catalog. Hastings turned it into the largest content studio in the world.

Netflix has 260M+ subscribers in 190 countries. Hollywood now follows the streaming model Hastings invented. Blockbuster filed for bankruptcy in 2010.

[ Why it was risky ]

Each transition cannibalised the previous business. DVD by mail killed the rental store. Streaming killed the DVD. Originals killed the licensor relationship. Each time the analyst class said don't, the customer base churned, and the stock fell. Hastings did it anyway because the alternative was being slowly displaced by someone else doing it to him.

[ What it looked like ]

[ EVIDENCE 01 / REED HASTINGS, TAKING TV TO THE INTERNET / CNBC 2015 ]

[ The numbers ]

From a DVD-by-mail experiment to the largest streaming and original content business in the world. Three platform transitions in 25 years, each one threatening the previous business model.

[ The lesson ]

The risk wasn't the streaming. It was killing the DVD that was paying for it. Hastings ran Netflix on the rule that the next chapter has to threaten the current one, or you're not actually transitioning, you're hoping. R.I.S.K. exists for leaders who treat their own success as the hardest thing to escape, and who are willing to break the model that's working in order to build the one that will work next.