Computerworld

A Google 'stalker' deconstructs the secrets to its success

A self-described 'Google stalker' examines the secrets to its success and explains how you can emulate it

Educators used to follow the auto industry because that's where all the lessons came from, says Bala Iyer. Then it was Microsoft. Now it's Google. In this month's Harvard Business Review, Iyer, an associate professor of technology operations and information management at Babson College, looked deep into Google's DNA to discern what makes it an innovation machine. Iyer talked with Kathleen Melymuka about what he and co-author Thomas H. Davenport discovered.

Let's talk about some of the key attributes that you say contribute to Google's success. The first is "strategic patience." What do you mean by that?

Look at their mission statement: "to organize the world's information and make it universally accessible and useful." This requires a long period to execute, and they're not kidding. CEO Eric Schmidt says it's probably going to take 300 years. When a company has such a powerful long-term mission, that sets it apart from the others. They go in knowing they're in for the long haul. Every action they take, you can see that.

In these days where quarterly results are everything, how can other companies emulate this?

Google won't play by those rules. Google changes the rules by this mission statement. They can take actions they feel are good for the long term. Google won't even give [Wall Street] guidance for their quarterly results.

This is a basic thing in business: You want to make sure you have a powerful mission statement. You can give people a lot of freedom to do what they want to do and be productive and innovative when you have a mission statement like this and allow people the freedom to operate within it. Companies do try, but sometimes the mission seems like it's separate from the company.

How does Google's infrastructure support innovation?

Everybody thinks that when a company is based on the Internet, you [get a] free ride on the Internet's infrastructure. But this company put billions of dollars into building things on top of the Internet. Google has its own operating system that works on top of the Internet and is based on Linux. It customizes the Internet to its advantage.

For example, Google deals with huge amounts of data. Data centers are very important to them. They need to be able to add them as necessary. When Google plugs in a data center, they can make it operational within eight hours, which is a phenomenal capability, but they have it because they made this investment. When they introduce new products and services, no one questions whether their infrastructure can support it.

I imagine this is something almost impossible to copy.

The lesson is not about building another billion-dollar infrastructure, but about how much you could customize to your own advantage. You could still layer your own system on top of the available public infrastructure. Isn't that what Salesforce.com is doing? Amazon's cloud, eBay's infrastructure -- that's what it is. The platform should be built on top of the available Internet. The idea is to put your own secret sauce on top of it to support your ecosystem.

This platform on top of the Internet even supports Google's product development, because they can test-drive products on their own infrastructure and can allow third parties to write products, which we're now calling mashups.

Let's talk about those.

Stacks have always been a big part of our industry. And the value is migrating up the stack. These guys are creating higher value. An example is Housingmaps.com. Somebody thought it would be a good idea to combine Craigslist and Google Maps to look at properties for rent. You get permission from Google, and that's a mashup.

Page Break

When you build, you have to pick a platform, and each has own API and languages and so on. If I'm an independent software developer, what languages should I be learning today? Google's APIs, Amazon's, eBay's, Salesforce.com's. I should keep track of the number of mashups being developed on these platforms so I can learn these APIs and build my own mashups that run on them. That is a good career path. You've got to watch the evolution of these platform companies and see which are the next operating systems, if you will.

You write that innovation is literally built into job descriptions at Google. How does that work?

Managers are required to spend 70 per cent of their time on core business, 20 per cent on related but different projects, and 10 per cent on anything else. [Technical employees spend 80 per cent of their time on core business and 20 per cent on projects of their own choosing.] It doesn't have to be on a daily basis; they can chunk it or spread it out.

Sounds good, but if you don't track it, it's just a meaningless statement. But these folks track it. That 20 per cent time has produced phenomenal products, including Gmail, AdSense and Google News. When you have creative people and you make this part of the job description, it's a great fit.

That would seem to be something any company can do, if they commit to it.

Every company should do it. You may think it's goofing off, but that's where the best ideas come from: time to sit down and actually think. People are so caught up in delivery that we don't have think time. Everybody who wants to add value knows it comes from our free time.

Next, Google cultivates a taste for failure and chaos. How so?

Failure is not considered to be a bad thing. They encourage people to make new mistakes. If people are not feeling stretched, they're not trying hard enough. In this business where innovation is happening on a day-to-day basis, unless you push the envelope, you're not going to be leading. There's a story about how a senior executive made a big mistake that cost several million dollars, but [Google co-founder] Larry Page told her he was happy she made the mistake. Because if they're not making mistakes, they're not taking enough risks. They would rather try and fail than sit back and say, "What if we had ...?"

And finally, Google uses data to vet inspiration. Is that backing away a little from failure and chaos?

We didn't want readers to feel that Google is totally chaotic and there are no processes in place. Actually, they have very good processes. They listen to ideas. Ideas get vetted on an internal discussion board. Once ideas come through, people get to present them to the top team -- 15 minutes per idea. And they don't care about what you think; they want data to back it up. So first, you alpha-test your product, find out what users are telling you, and then you present that. They're very brutal internally, and a lot of ideas get killed. But then you feel confident that the right ideas are being pushed up.