In March, I sang the praises of Staying small, as each team focuses solely on its added value, and we share more as One Microsoft. If you agree that to go fast you must be small (which I do and you should), then shouldn’t Microsoft be much smaller? According to the 2013 Fortune 500, we have roughly 25 percent more employees than Apple and roughly 75 percent more employees than Google. That’s enormous! And these figures don’t include our new coworkers from Nokia!

To be fair, we are involved in more businesses than Apple and Google, but we’re still frigging huge. Do we need that many employees? No, we don’t. We are top-heavy and overstaffed. Should we break up into 650 companies of under 150 people each (plus 170 companies from Nokia)? No, we shouldn’t.

In spite of all the sloth and overhead associated with big companies, there are projects only they can do: Azure, Bing, Office 365 (including SharePoint, Exchange, and OneDrive), and a single-app model for Windows, Windows Phone, and Xbox, just to name some of our most recent work. No companies working on projects like these are small. The key for big companies is to focus on projects only big companies can do and to be the best big they can be.

Go big or go home

Why can’t a small company create Office 365? Yes, it’s a large suite of software and involves enterprise-scale services, but couldn’t a small company, or a collection of small companies, build up a similar offering over time? Basically, no.

Given enough time, 150 people could move a mountain with just picks and shovels, but that’s not realistic. The picks and shovels would get worn down in just a few years, and the people would die long before they could finish. Even if you had 100 groups of 150 people, you’d still need to equip them with more than shovels and coordinate their work. To do big projects, you must have heavy equipment (bulldozers and datacenters) and an organized army of people.

Building things like datacenters requires substantial capital up front. Maintaining datacenters requires a steady flow of revenue. Hiring and equipping an army requires substantial logistics, procurement, and capital. Directing an army requires vision, alignment, and coordination. None of this is easy, but it is essential to accomplish big things.

Eric Aside

Open source advocates would claim that hundreds of independent groups have moved many mountains in the open source community. While most open source projects are small, there have been several big ones. Those big projects were well coordinated, with a clear vision and alignment. They took years to become the size they are today, with much sacrifice from their primary contributors. Doing big things comes at a cost.

Because it is there

At Microsoft, we want to change the world—not once or twice, as some small companies have done, but many times across society. We want to move mountains. Unfortunately, moving mountains means lots of overhead and coordination. As a result—and this is key—Microsoft will never be competitive long term on small projects.

Small companies will always be faster and more frugal than big companies over time. They simply don’t have big-company burdens. Even a six-person team at Microsoft with full autonomy inherits the overhead of Microsoft benefits, staffing, administration, IT, facilities, networking, and procurement, not to mention the compliance and liability obligations of a large company. Any big company like Microsoft simply can’t compete long term with small companies on small projects.

Big companies still need to innovate, starting small on projects that may eventually be huge. But wasting effort in small areas that will never be big is a big mistake for a big company. We’re big in order to work on big ideas that change the world. Anything less for a big company is a fool’s errand.

Eric Aside

A company staying focused on its particular strength is known as the hedgehog concept. Even a company known for innovation, like Apple, uses this approach. You don’t see Apple shipping plumbing supplies. It ships beautifully designed electronic devices with simple experiences that Apple controls end to end. Microsoft ships integrated platforms that individuals and companies can personalize to enhance who they are and how they succeed.

Big projects often have small gaps that need to be filled in order to complete a customer experience. When the software ecosystem is mature, small companies fill those gaps and are rewarded for it. However, when the ecosystem is still nascent, a big company, like Microsoft, needs to fill gaps itself. That’s fine, as long as the big company steps aside as better-suited companies step forward to meet this need.

I don't even exercise

It’s not enough to work on big projects. We need to work on the right projects and execute them brilliantly in order to be successful. Who’s responsible for selecting the right projects and driving brilliant execution? Certainly, frontline engineers play a crucial role in ideation and doing the actual work, but in the end, it is leadership that sets the vision, drives alignment, and coordinates the work.

The success or failure of large projects rests on the effectiveness of their leadership.

  • Leaders must articulate a compelling vision.
  • Leaders must align all the small teams to that vision.
  • Leaders must coordinate all the small teams to work in harmony.

I could write columns on each of these topics, and in fact I have. The two primary ones are Vision quest and Coordinated agility. If you think focus isn’t necessary, and teams should just do it all, read You can't have it all.

Are we there yet?

How is Microsoft doing in its quest to change the world? We started off well—integrating computing into our society. However, for the last 15 years, our vision has been less compelling and our hundreds of coordinated, small teams have consolidated into unwieldy and slow behemoths.

Thankfully, we’ve got new leadership that is focused on speed, agility, and coordination. These leaders have talked about moving to a data-driven model of continuous improvement and continuous deployment. This model is enabled by staying small in our individual feature teams, while being big in our overall vision. So far, I like the message and I like the changes.

We can all be part of the transformation. Break your groups down into feature teams of five to 10 people. Free them to ship frequently and independently, iterating with Data-driven decisions. But, importantly, don’t let your small team work on small projects. Instead, align and coordinate your small teams with our big vision. We’ll win together when we switch to the speed of staying small, while embracing the power of being big.