Billions of dollars have been invested in the rise of data science and machine learning as mainstream disciplines in the world of business, one of the most exciting tech trends of the last (and next) decade.
In the enterprise, many of the applications of data science and machine learning ultimately produce a prediction: which customers are the most likely to buy? Or churn? Which transactions are most likely to be fraudulent? What part of town is likely to place the most food deliveries tomorrow afternoon?
However, powerful though it may be, there is one thing machine learning generally doesn’t tell you: once you have a prediction, what do you do with it? For example, once you have predicted high demand for food delivery in a certain part of town, how do you decide which delivery team member to dispatch where and when, to optimize for efficiency and maximize revenue and customer satisfaction?
Enter decision science. While the term has not crossed over to mainstream consciousness like its data science cousin, decision science has been around for decades. Also often known as Operations Research, it encompasses a variety of advanced analytical methods and quantitative models to help with decision-making and efficiency, including simulation, mathematical optimization, queuing theory, etc.
Decision science covers *many* use cases across industries, basically any scenario where optimization is required: pricing (e.g., airline tickets), routing, assignment (which Uber driver should be sent to which customer), network optimization, financial engineering (asset allocation, risk management), bidding, etc.
Many hypergrowth tech startups, which often are on the right side of the future when it comes to deploying the best tools and processes, have already combined data science and decision science. DoorDash, for example, has made heavy investments in machine learning infrastructure (see our chat with Alok Gupta, its head of Data Science and Machine Learning), and also has a full Operations Research team, working on assignment optimization problems (see this post from their engineering blog for example).
So why hasn’t decision science become ubiquitous yet, across startups and larger companies? To date, the tools that have been available on the market have been helpful but clunky, as they were designed to be operated by Operations Research PhDs, a small group of power users. Those tools were also designed several cycles ago in terms of overall software innovation.
The time has come to take decision science to the mainstream, and this is exactly what Nextmv (pronounced “next move”) is doing. The company offers developer tools that put the power of decision science into the hands of regular software engineers who have no Operations Research background. Their API and cloud product plays nice with modern software architecture, and abstracts away the complexity of the underlying science.
Today, I’m excited to announce that FirstMark is leading Nextmv’s Series A, just a few months after being involved in their seed round. The fact that we, alongside existing investors Dynamo and 2048, decided to double down on the company so soon, while it still had almost all of its seed money in the bank, is a testament to accelerating momentum and the magnitude of the opportunity for the company. It is also, and most importantly, because of our conviction about the utmost quality of the team, including co-founders CEO Carolyn Mooney and CTO Ryan O’Neil, who previously built the decision science team at Grubhub.
Of course, Nextmv is hiring! A number of positions are open across engineering, sales and others (all remote).