Tuesday, December 06, 2005

Data Based Decision Making

Topic: Management Improvement

Acumen visits Google:

As a first step, we hope to collaborate with interested Googlers to find better ways to learn what works around the world. Identifying powerful solutions to poverty that are useful to people in different settings, and that are market-driven, scalable, and sustainable, is our greatest challenge. Second, we're hoping to strengthen how the world measures both social and financial returns to investments in delivering critical goods and services to the poor. Like Google, we hold a deep belief in the power of measuring everything we can.


Google has done a fantastic job of using data to make decisions. In fact so much so, that some think they may go overboard trying to find an algorithm for everything. My dinner with Sergey:

It was a classic Google moment. Your S.A.T. score was the measure of your intellectual capability; your GPA represented the numerical summary of your ability to execute on that potential. Your value to Google could be plotted using those two data points.

Sergey's desire to reduce every decision to an equation would cause me a fair amount of frustration in the years to come. While it forced a discipline on me that was likely lacking in my career up to that point, it also went against my deeply-held conviction that some things are not expressible simply by deriving the correct algorithm. A lot of engineers at Google would dispute that with religious conviction, though they might admit that deriving the correct algorithm would be "non-trivial."


I believe you can't measure everything that is important. I also believe in most organizations the amount of stuff you can't measure usefully and realistically is quite a bit higher than it is for Google. Having highly intelligent, skilled and experienced people who can derive complex formulas effectively does greatly expand the effective use of measurements.

Still there are limits, and those limits are much lower for most organizations that have neither, thousands of phd level mathematicians, rocket scientists, software engineers etc. nor a anything approaching Google's percentage of such people.

Still I think we will benefit from the innovation that will continue to take place at Google. The are making great strides in using data to inform their decision making process.

No comments: