Deconstructing the Accelerator: 5 Stages of Startup Discovery and Selection

In March, the Microsoft Ventures Accelerator in Seattle will announce its second class of startups—this time with a focus on digital work. These startups were chosen after a thorough jury selection process, preceded by a meetup tour across ten U.S. cities.

The Microsoft startups team and Microsoft Ventures Accelerator in Seattle Director Mukund Mohan canvassed some of the country’s top tech ecosystems in search of high potential startups. Kicking off the tour at Galvanize in Denver, the team visited several other startup hubs, including DevMountain in Provo, Utah, RocketSpace in San Francisco, NYU Entrepreneurial Lab in New York, 1776 in Washington D.C, 1871 in Chicago, Capital Factory in Austin, WeWork in Seattle, Real Office Centers in Los Angeles, and StartX in the Valley.

Finding groundbreaking startups that meet the selection criteria—which includes stage, team and theme—is no easy task. It takes a lot of planning and coordination; and even with its resources, Microsoft is not spared the dynamic and volatile nature of the startup ecosystem.

Here’s a behind-the-curtain view of the discovery process:

Stage 1: Identifying key partners

Our team first identified key partners across the ecosystem. As you can see from the tour details outlined above, we have formed partnerships in most of the established and up-and-coming startup cities. We’re always excited to work with our partners and we value their referrals and recommendations.

Stage 2: Aligning various internal teams to help support the mission

As a company, our footprint right here in the U.S. and around the globe, is huge. As such, whether you’re a startup working out of Grand Circus in Detroit or WeWork in Seattle, there almost always is a Microsoft technologist or advisor available to help you grow your business. Scouting high potential yet lesser-known companies is a challenge, so we rely on our teams scattered across the country for insider knowledge and input regarding up-and-coming startups in their regions.

Stage 3: Showing up and building relationships

Of course, the best way to find eligible startups is by meeting them in person, answering questions and exchanging ideas. The ten city meetups were filled with excitement, promising startups and lots of healthy conversations. This was a great opportunity for startups in the digital work space not only to learn about our accelerator programs, but also to connect with one another as well as engage with supporters from the community at large. To get a glimpse of what the City Tour was like with added insights into the value of such gatherings, check out the write-up on our New York meetup.

Stage 4: Finding the right fit

Several attendees asked during the meetups: “What are you looking for in a startup?” In addition to being an enterprise startup building in the Digital Work space, we were looking for companies with rock star teams, passion, a scalable business model, and a bold solution with the potential to transform the way we work today. Identifying companies for the accelerator is also about aligning with what the companies need from Microsoft. We take the time to understand what value we can bring to the table and explore whether there’s a true fit between the accelerator program and the startup.

Stage 5: Sifting through the noise

With several hundred applications, we formed a jury made up of employees with startup experience and outside experts to help select the top startups. A rigorous selection effort to meticulously review the stacks of applicants, combined with recommendations from our team, culminated in a full day of in-person discussions with our top finalists.

Stage 5: Sifting through the noise

It took three consecutive weeks of meetups in ten U.S. cities, over 1200 attendees, meetings with hundreds of entrepreneurs, and a full jury selection day to find a handful of top companies that would be a great fit for the program—and we’re eager to announce the second class.

Next week, we'll dive into the application data. Stay tuned!