While most entrepreneurs dream of growing their startups into a beast like Facebook, most never reach global domination, get Hollywood treatment, and become worth billions of dollars like the social network. Statistically speaking though, successful startups are a rare breed.
One out of every ten startups do fail, eventually.
Bjoern Herrmann wants to crack the code of innovation. To discover what makes up the DNA of successful companies, he needs to understand why they fail. Herrmann co-founded The Startup Genome Report to survey 3,200 startups, a research initiative that used machine learning to identify the startup's type and stage.
It turns out, many companies that fail suffer from premature scaling. Don't blame age, location, experience, or education - none of those factors affect a company's chance of heading down the path of failure.
"[Premature scaling] helped us to explain about 74% of failures," he said to me. That might explain why so many startups do fail.
So what exactly is premature scaling? It occurs when one side of the business scales before other parts. This can happen when a startup hires too many people or messes up customer acquisition strategies. For instance, a startup that hires 40 people before they have customers is, well, tempting fate.
Startups are temporary organizations that are designed to evolve into large companies, Herrmann said. They have six stages: discovery, validation, efficiency, scale, sustain and conservation.
But most get lost along the way.
The team behind The Startup Genome also created a business accelerator called Blackbox, to use the data they find. The team added a tool, so entrepreneurs can track their own growth and use it to compare with other companies in the system.
Thinking about failure all the time must be nerve racking: Is Herrmann worried about his own startup failing? "I'll do my best to reduce risk and increase certainty," he said.
Cracking the code for failure, isn't as flashy as innovation, but at least, it can help entrepreneurs avoid repeating the same (premature) mistakes.