Where the ancients sought alchemy, today’s investors wish they could accurately predict the success of startups. Now one company thinks it has built an algorithm that can make that wish a reality.
Instead of guessing, consulting, or offering advice based on years of experience, Growth Science is trying to take the disruption theory to a whole new level. “As far as I know, there isn’t anyone else in the world who predicts startup survival or failure for a living,” says CEO Thomas Thurston. “At least not in any quantitative way, it’s all guesswork and intuition out there.”
Business model simulation as a service doesn’t have a recognizable acronym–like SaaS–because it isn’t a thing, and no one else does it according to Growth Science, except them. “A few people use simulation software to model things like cash flow, supply chain, et cetera, but nothing remotely similar to what we do,” says Thurston. “But nobody out there simulates business models, just us.” The company had previously only been doing this for huge, Intel-size, companies, but the newest and most important change has come with automating the processes in order to bring costs down to become accessible on the startup level.
About 80% of Growth Science’s process is now automated. “There’s a kind of data that is very valuable to predicting business survival or failure that’s related to technology trends, even for low-tech products. We used to harvest all that information manually–for example, human researchers–which was slow, expensive and difficult,” explains Thurston.
“Similarly, we’ve found about 80% of the predictive value for a startup has to do with externalities–market, customers, competitors, et cetera. Only about 20% of our algorithm looks at the startup itself,” says Thurston. “That’s a big surprise for most people. Also, the team only accounts for around 12% of our equation. 88% of what our algorithms look at has nothing to do with the team whatsoever, whereas most VCs list the team as the number one thing they focus on when investing in a startup.”
Most of the research for the algorithms used to predict success (or failure) come from the year Thurston spent working with Clayton Christensen at Harvard Business School. Christensen has been known for his disruption theory which tries to explain why, and which, companies are able to disrupt whole industries. Growth Science uses this data science to look at three key areas: the likelihood of the business surviving or failing, the growth expectations, and what changes can be made, in the case of failure, to increase the odds of success. While the exact algorithms are highly protected, Thurston did mention a few companies it’s found favorable, such as Dropbox, Tango, Indow Windows, Practice Fusion, and CloudFlare. All of which he described as successful disruptive businesses.
In case it’s not clear from past examples, including the one set forth by the book Moneyball, using certain data as a prediction tool isn’t about giving the decision making process over to machines, it’s about trying to expand the way we think and use every tool available. Is it an unfair advantage to some, or does not using all the data an artificial limitation?
This is also going to affect the way startups think about their own business. It can be an intimidating obstacle to know there’s a company like Growth Science out there capable of calculating the odds of success before ever launching a product into the market. Thurston, of course, describes Growth Science–and I think all similar types of ventures–as a benefit. “Think about it this way–everyone uses CAD to simulate new products and tweak them in a virtual environment before building a ‘real’ one. Why not do the same with businesses? In other words, the business is basically a prediction factory.”