Thursday, February 15, 2007

A fork in the road for Google.org?

When I worked in Russia for the World Bank, I got to know a lot of government officials who had been rising stars under the old Soviet system. One day in 1995 I was having drinks with a guy who used to work at Gosplan, the Soviet central planning bureau whose job it was to determine all the goods and services that would be produced in the country each year - and where they would be distributed and consumed.

I asked him, "Konstantin, I am really baffled by this central planning system you had. Tell me - how did it work?"

"Badly!!" he exclaimed in reply. "Very badly. It was a terrible system that was horribly inefficient. In retrospect, it was crazy to think that we should be making all those decisions. In fact, we had to keep hiring smarter and smarter people --and working longer and longer hours--just to keep the system from sinking under its own weight, which of course it eventually did...."

I laughed and had another vodka. But about three years later after I was asked to co-lead the Corporate Strategy Group at the World Bank, Konstantin's words began to haunt me. I realized that the World Bank was structured like a central planning agency. The idea at the Bank was literally to "hire smarter and smarter" experts and make us work "harder and harder." And we were responsible for allocating approximately half of all the official aid resources to developing countries.

And yet, no matter how many experts we hired or how hard we worked, we "experts" suffered from three big problems: 1) lack of adequate information about local conditions and what people in developing countries really need; 2) lack of "bandwidth" to process all possible solutions to problems and challenges - there are only 24 hours in a day!; and 3) lack of feedback - it was very hard to understand the impact of our policy advice and project investments, so it was hard to adapt and improve.

At that time, I began to think about how market economies differ from central planning and why they are more successful. The core answer is because they distribute information, decision making, and resources among millions of decentralized people and companies. And the emergent nature of markets mean that successful innovations encouraged and replicated.

In the US and other market economies, there is no "World Bank" making decisions about what needs to be produced or where it needs to be distributed. Instead, there is Adam Smith's invisible hand at work: "[The merchant] intends only his own gain, and he is...led by an invisible hand to promote an end which was no part his intention." The invisible hand has its shortcomings, and Adam Smith himself talked about the importance of government in regulating certain aspects of the economy and even producing some goods. But no other system discovered to date comes close to the invisible hand in driving innnovation, creating wealth and addressing people's needs.

So what does this all have to do at Google? Google describes its mission as "organizing the world's information." But I would describe what they do as "revealing the underlying organization of the world's information." If you search for "dog" on Google, for example, there is not a central planner at Google deciding which search results you "should" see. Instead, Google's algorithms (roughly speaking) display the dog-related sites that have the most relevant links from other highly ranked sites. Google thus "reveals" the content and connections of all the users of the web itself.

The remarkable thing is that as soon as Google launched several years ago, people quickly learned that Google search results were superior - often far superior - to results from other sites that used a centrally planned approach where editors determined what search results should be displayed.

The power of aggregating these user activities and links by Google is remarkably similar to Adam Smith's invisible hand.

So it is strange to hear some of the reports coming out of Google.org - Google's social good arm - implying that they are going to take a much more traditional approach to their philanthropy and social investments. Although their plans appear to be far from decided, it looks like they may choose a sector and a small set of countries to "work on." Presumably, a relatively small set of (very smart) people will analyze the situation, talk to people, design a program, etc. And then the program would be "carried out." The problem, as the Soviet Union demonstrated (and as Bill Easterly masterfully analyzed in his book "The Elusive Quest for Growth"), is that this approach does not work.

Although we at GlobalGiving are often referred to as "an eBay for philanthropy," in a sense we are "organizing the world's information about global philanthropy." Our eventual goal is to have all bona fide philanthropy projects around the world listed and "organized" on our site according to algorithms we are refining over time. Google has both the platform and reach to accelerate this dramatically. Will they seize the moment?