Ex-Yammer trio launches Mode: ‘GitHub for data analysis’

A trio of data-analysis veterans — all most recently part of Yammer’s analytics team — has launched a startup called Mode that hopes to do for analytics what GitHub has done for code. The company has raised $550,000 in angel funding, with Yammer Founder David Sacks pitching in the lion’s share.

Mode Co-founder and CEO Derek Steer says there’s a problem right now in the world of data analysis: Scripts are being written, used once and then vanishing into the ether. That might be fine if that data or those results would never again be useful, but that’s rarely the case. Most of the time, the isolated nature of data analysis just means the next person doing something similar has to reinvent the wheel.

At a gaming company, for example, someone might run an analysis on user behavior, use it to develop a new product or inform some decision, and then move on. If someone in another department need to do a similar analysis two months later, she’d probably never know the previous work existed and would have to write her own script from scratch. Mode “makes any one person’s work available to a broader group of people,” Steer said.

Derek Steer's LinkedIn profile pic.

Derek Steer

Thus Steer’s GitHub analogy and his other analogy to academia, where new research typically builds on previous research by adding a new, valuable piece on top of it. That’s what Mode does — surfaces the work that’s already been done on a specific analysis, saving all the underlying data, code and versions so people can see what’s been done and take what parts they need.

Like Github, mode will be free for individual developers and data analysts, and anyone can access community-wide projects. The company plans to make money by charging companies who want to use Mode as the layer for finding internal analyses that are still being stored in corporate databases.

Steer is quick to point out, though, that Mode is not about sharing dashboards or other actions that often constitute collaboration among business users. In fact, Mode isn’t even targeting those users. It’s targeting the people neck-deep-in-data who need to work with languages like R, SQL and Python, and who need to see below the surface of chart into what data was used and how it was created.

I asked Steer if he’s concerned that targeting data analysts and data scientists isn’t too narrow a strategy, but he’s confident that community of users is both large enough — and growing fast enough — to support a company like Mode. Actually, he added, the whole concern over whether there will be enough technically savvy data analysts is a mindset rooted in the realities of a few years ago. (Although, I would note, it’s one that inspired a whole lot of innovative startups trying to democratize complex data analysis.) And, he said, there will always be people who need to do cutting-edge work or otherwise get down and dirty with data at a level most business-user-focused tools can never touch.

If you look at all the questions on StackOverflow, Steer noted, there are more about SQL than anything else. The D3.js data visualization community has as almost many members as does the Rails programming framework. Plus, universities and other training programs are putting a greater emphasis on these skills and churning out even more people who possess them.

“To say that there aren’t that many of these people,” Steer said “… is not a very forward-looking perspective.”