Google & Product Management

I was privileged to spend an evening at Google Haifa with:

• Yossi Matias, Head of Google’s Israel R&D center
• Marissa Mayer, Google’s VP of Search Products and User Experience
• Noam Nisan, Google research scientist
• Some Googlers demonstrating some of their products (Sadly I didn’t catch their names — anybody who knows them, feel free to chime in)

The understated descriptions for Yossi, Marissa, and Noam are from Google’s page for this event, and remind me of Danny Sullivan introducing Matt Cutts by saying that “Matt is a senior engineer at Google. Which is all you need to know about him.” Dr. Matias is a former air force pilot who published over 100 research papers and holds 20 patents. Dr. Nisan is an expert in algorithmic game theory. Marissa Mayer has led product management efforts on Google search products since 1999 and was the youngest person to ever make Forbes’ list of most powerful women.

What most stood out to me was Google’s approach to product management.

• A Googler mentioned that Google has never been good at starting with a business plan and building the technology to address those needs.
• The same Googler mentioned that meetings are usually short at Google because Google’s product managers were all very smart & technical and understood the issues right away. He contrasted that to the other companies in which he worked where he said most product managers were idiots.
• There were 4 rooms where Google engineers were holding court, describing the products they were working on and answering questions. The passion and pride were impressive.
• Dr. Matias discussed how Google doesn’t have traditional product management. In addition to Google’s famed policy of letting engineers spend 20% of their time on pet projects, they have a large role in defining the features on which they spend the other 80% of their time.
• They don’t have any project managers. He said they don’t use Gantt charts, by which I think he meant they don’t use any scheduling tools. They do have deadlines, but the deadlines don’t seem to be driving the pace of development.
• Within engineering, the team leadership positions are fairly fluid. One person will lead one project, and then a different member will lead the next project.
• It’s largely geeks working together to design and develop products and features that they want to develop and use. Their passions, skills, and professionalism drive the pace.

They were trying to recruit developers, so they may have been playing up the sides that sound like developer heaven, and perhaps showing off their happiest and most passionate developers. It did all sound great, but it’s important to understand some of the harder things that make this work:

• Test-driven development. They had papers put up in the bathroom educating about test-driven development’s importance and best practices.
• Multivariate testing environment exporting boatloads of data.
• Data-driven decision making.

In this sense Google is the ultimate engineering organization. Their product development process seems built on the metaphor of the compiler. Developers are given more say to develop what they think will work because ultimately their creations have to pass this super-compiler, which checks not just for syntax errors but also whether or not the code breaks some other functionality or harms some business metric.

As a product manager, I’ve been repeatedly humbled by our own data-driven systems. It’s our nature to assume we’re usually right. Maybe you are. Testing has proven to me that I’m not.

Google’s Udi Manber told Business Week that they ran 5,000 experiments last year, and probably have 10 experiments for every successful launch.

On a perhaps related note, at SMX Advanced Bing’s Yusuf Mehdi pointed out that Bing’s mission was to help people accomplish their tasks, while Google’s mission was to “organize the world’s information and make it universally accessible and useful.” Bing focused on people, Google focused on data. One quick data point: Googling “Google’s mission” got me directly to what I was looking for. Binging “Bing’s mission” got me nothing, so I’m quoting Mehdi from memory. Focusing on people sounds like a great idea, but focusing on the information may actually be a better way to give people what they want.

Peter Drucker spoke of “management by objectives and self control.” Google takes this a step further by giving testing and data a central role in management. I wonder if most engineers would rather be managed by a compiler or by a person. It’s nice that Google has icons like Marissa Mayer to put a human face to it all. Google’s system is admirable, IMO, and probably quite efficient. Still, it may be best that Google’s product managers are generally engineers by training. This approach may be best suited for us tin men.

More about Google and its competitive environment:

• Google & The Facebook Fantasy
• Microsoft, Mehdi, and Matt Cutts
• The Google Definition Link and Answers.com