What if recruiters could know, before posting a job listing, if the specific language in the job description will actually attract the types of candidates they want? Or if someone launching a Kickstarter campaign could submit the project description for analysis in advance to determine the prospects for success?
Those are the types of scenarios that will be enabled by Textio, an early stage Seattle startup founded by two Microsoft veterans: Kieran Snyder, a linguist and cognitive scientist who worked most recently at Amazon; and Jensen Harris, a 16-year Microsoft veteran known for his work leading the Windows user experience.
News of their startup emerged over the weekend, after Harris announced plans to leave his Microsoft job. Both Textio founders — Snyder is the CEO and Harris the CTO — stopped by the GeekWire offices in Seattle this morning to talk about their plans.
For people who track Windows, the timing is interesting. Harris’ departure follows Microsoft’s unveiling last week of a new user interface in Windows 10 that rolls back some of the key approaches introduced in Windows 8 — reintroducing the Start menu, for example.
So does his decision to leave have anything to do with that?
In short, the answer is no, Harris said. “Really, my decision to leave was based on the fact that I’ve learned what I needed to learn from Microsoft. The stuff that got shown within Windows 10 was stuff my team had worked on when I still working on Windows. So I wasn’t surprised. We knew that’s where we were heading.”
“It really is more about it being the right time for me personally to do something where I can be much closer to the product, being able to build something with rapid growth,” he said. “Microsoft is great because of the scale of the company. Just the number of default users you reach just by shipping some software. But the ability to create growth and to create a company from scratch, and have it be really successful, is a fire that I had burning in me.”
Harris and Snyder say they see an opportunity to build something significant with Textio.
“Although there are a few other companies that are doing text analytics, people need a computational linguistics degree to use them,” Harris explained. “They’re not easy to use; they’re horrible to use.”
Snyder’s work at Microsoft included creating language detection tools and natural language processing capabilities for Windows developers. Most recently, she worked at Amazon, building a team within the online retailer’s advertising organization. Her past work includes research to detect gender bias in performance reviews and in workplace conversations, but the text-analysis capabilities of Textio will go much further to include all sorts of documents and outcomes.
The company’s service will be able to take in all sorts of business documents — candidate interview feedback, marketing copy, customer service chats, product specs, etc. — and determine whether the text is resulting in the desired outcome for the company.
Snyder pointed out that text is the biggest output for many companies, so it only makes sense to find new ways to glean insights from those words.
“For any type of text, we want to be able to extract business insights,” Snyder said. “We’re exposing this as a service. It’s not a consulting model. That’s a key element of how we work, which also places the bar very high on user experience and document security.”
For example, one of its first demonstration projects involves analyzing publicly available text on the Glassdoor employment review site, correlating job descriptions with company rating to determine how those descriptions can result in hiring the right or wrong employees.
Based in Seattle’s South Lake Union neighborhood, Textio is in the process of making its first key hires. The company is bootstrapped so far but is talking with potential investors. The startup already has companies lined up as beta testers and plans to work with a handful of them over the next four to five months, before launching its initial product next year.