The data science team was coming up with cool ideas, but the engineers couldn’t implement these applications, as fast as they could produce them, simply because they lacked the tools to do it. That’s when they decided to create a tool that would give these teams a head start toward working more efficiently. That tool would become Yhat (pronounced Y-hat).
“We had a team of data scientists who produced new and creative ways of doing business and decision making, and engineering couldn’t keep up with the data science team. Austin and I were both working in the middle of this and watching all of this analytical work at OnDeck just sitting on shelves,” Lamp explained.
They figured there had to be a better way to get value from the work the data scientists were producing.
With their first product, ScienceOps, they developed a solution designed to help teams of data scientists work and communicate more effectively with one another as they built projects on top of popular data science tools like R and Python — but they didn’t stop there.
They also wanted to make it easier for the engineers and the lines of business to implement those applications to benefit the company by putting the ideas to work more quickly.
They kept their day jobs for a short time, but by June, 2013 they saw so much demand, they left OnDeck and started Yhat. For a time, they were working out of their Brooklyn apartment, but they soon out-grew that and moved into a shared office space in Manhattan where they are today.
The company began to ramp up in earnest in 2013 when they got a cool million in seed money, and they are currently part of the Y Combinator 2015 winter class.
Lamp says that, because they were already somewhat established and even had customers, they weren’t sure what YC could give them. As Lamp put it, “we were a lot further along than two hackers with an idea.” They worried that it might actually be a distraction, but “it’s been fantastic so far,” he says.
Both founders agree that it’s been a great experience and their batch teammates have helped them focus. They realize now that as new entrepreneurs, they didn’t know what they didn’t know and YC has helped them take a range of problems any new company has to deal with, and filter them objectively to concentrate on the ones that matter most.
For Yhat, a startup with a more mature idea, that has meant working on sales and prospects, whereas some of the newer startups might concentrate more on marketing and product development.
Yhat announced a second product called ScienceBox in June, 2014, which is a lower cost alternative, designed to help data scientists create and share data science projects.
Yhat has some interesting customers including a NASCAR racing team that is trying to figure out how to use data to get an edge by enhancing pit decision making.
“The NASCAR team sees [big data] as a new thing they know they should be doing, but they know they are behind, and are looking to get ahead of the curve,” Lamp explained.
Another client is Condé Nast. You might not think of a publishing company being data-driven but it has a team of 40 or 50 data scientists spread across a couple of different locations, so Yhat’s products appeal to them.
Lamp says ScienceBox has also been popular with data science professors as a way to deliver all the tools that participants need to take the class. Prior to using using ScienceBox, it took a lot of work to ramp up and get all of the students set up.
Lamp and Ogilvie acknowledge that, while they’ve been able to build a decent customer base, they may be a bit ahead of the industry where many companies have two or fewer data scientists. They admit that they’ve thought about that, but so far they have plenty of business for a company with just seven employees.
What’s more, Lamp said that data science is the fastest growing undergraduate major, and many folks in academia are leaving for lucrative jobs in industry. We may not be in the data science golden age just yet, but Yhat’s timing appears to be pretty good.
In case you’re wondering about the company name — I know I was — it actually comes from a German statistical measurement, kind of a data scientist in-joke.
Lamp says their target audience gets the joke, even if they have to explain it to everyone else.