Machine Learning Gets Hot, While Mobile Apps Cool off

While the payments business is caught up these days in working out strategies for mobile apps and social media, it turns out these aren’t the cutting-edge technologies they once were. Instead, fields like virtual reality, machine learning, and natural language are winning the spotlight.

That’s according to a study that looked at the frequency with which certain key words occurred over the years in the descriptions provided to investors by early-stage startups.

The study, conducted by New York City-based CBInsights, which tracks data on technology investment, examined prospectuses for venture-backed companies receiving their first round of funding between 2010 and 2016. Without being specific, the research firm says it scanned “thousands” of these documents.

In 2010, “machine learning” occurred in none of the prospectuses. By 2016, the term was found in 2% of them. Similarly, “artificial intelligence” registered at 2% this year, after a zero showing as recently as 2013, and “virtual reality” now is mentioned in 1% of the documents, after barely registering six years ago.

Those may not seem like big numbers, “but at the scale of thousands of companies, it represents a big jump in activity in the category,” CBInsights says in a blog post about its study.

Machine learning and related technologies are finding a place in payments, for example, with services that suss out fraud by rapidly absorbing patterns closely associated with fraudulent behavior and then zeroing in on those patterns across millions of transactions.

By contrast, mobile apps peaked at 30% just three years ago, and the term has now fallen to 24% as fewer startups, relative to all first-round, venture-backed firms, mention it. Instead, apps are now considered an entry point to other, more interesting technologies.

“While many startups may offer an app to access their service, the fact that they’ve built such an app is increasingly beside the point,” says CBInsights. “A mobile dimension to any successful service or software is taken for granted.”

Similarly, words like “mobile” and “online,” which once seemed so common in payments that they nearly held incantatory powers, are on the wane with startups. From a peak of 18% in 2012, “mobile” has fallen to 11%, roughly equal to its frequency in 2010. “Online,” meanwhile, has been on a steady downward course with early-stage tech companies, falling from 13% in 2012 to 7% now.

Even terms like “social media” and “social network” are on the decline among tech startups, despite the wild popularity of services like Facebook and Snapchat. The term “social” has dropped since 2010 from mentions in 11% of prospectuses to 4%, while “Facebook” has gone from 3% to 1% and “Twitter” from 2% to zero.

Interestingly, the word “electronic,” often used in not-so-distant years in terms such as “electronic payments,” is also disappearing among startups, with frequency falling from more than 2% to nearly zero now. “It seems that with an increasing focus on data mining, health, and software services, fewer companies describe themselves as working on hardware,” notes the CBInsights blog post.

Other hot trends include a number of technologies related to such fields as messaging and health care, the study found, while mentions of “email” appear to be going extinct. “It looks like social media, email, and mobile apps are ceding the stage to startups that focus primarily on healthcare, messaging, and artificial intelligence,” concludes CBInsights. “Bring on our chatty, robot doctors!”

—John Stewart