I often remark that it's impossible to establish the value of Google Search for the world economy. Google Search acts as the glue of the internet. Domain names and IP-adresses may point to servers. Google points to content. Most non-digerati don't care for bookmarks, adresses, del.icio.us etc. All they care for is Google. As Google doesn't charge for it's services and it makes the world more efficient in ways that can't be calculated it's value to the world economy can never be fully appreciated.
A way of establishing the value of an industry is by looking at the percentage of GDP it represents. Google's annual revenues are $22 billion. That may sound like a lot of money, but its tiny. That's less than 0.03% of the world's GDP . It's only 1/45th of the 1 trillion revenues of the mobile telephony industry. Per global internet user it's slightly over $1 per month. But here comes the problem with this method of valuation. Google's revenue is just from advertising, it's not from search at all. The two may be linked, but Google doesn't get paid for everytime you search for something.
So here is the conundrum and I'm not pretending to understand it fully or even at all. Google is offering a service, search, that nobody pays for. None of the benefits that a searcher derives from a search flow back to Google. Even the site that was found doesn't pay Google for the privilege. Google's value measured in GDP therefore is zero, nothing. It doesn't show up in official government statistics. Yet we all know the profound effect that Google has had on our lives. "Just f-ing Google it", or "Google is your friend" are expressions that are part of our lives. Google has made our lives more efficient in a gazillion different ways.
I've looked into account for price and quality changes over time in the calculation of GDP and they sometimes use deflators to account for differences in quality, differences in prices etc. However this does assume that there is an economic value for the service offered. If I've understood it correctly services like Google Search don't show up in such statistics as there is no price paid for the service. It's even more difficult, because the value of the production of a knowledge worker is equal to the value of the salary of the knowledge worker. The output of a knowledge worker therefore increases with the increase in salary over time.
So here you are, you're reading this so you must be a knowledge worker. You get linked to the Wikipedia, OECD documents and all this stuff I found using Google. Everything new you do in your work starts with a Google Search. One way or another everything can be found quicker and better than it could before Google was there and way better than before 1994 when the internet arrived to the general public.
As a knowledge worker if you lived in 1994 and had to write a report on the defects of the energy market of an EU nation, like I did in recent months, you had to rely on what your librarian could get you. Lexis-Nexis, some books, some government papers and interviews that was it. Finding the documents was the biggest problem. It might take you weeks to just get the documentation and there was no guarantee on whether you had found it all. Comparisons with other countries were even harder. And now what did I do... I just googled it. I'm not saying I couldn't have written a good report then, but I do know that in the first week on that job Google brought me to alot more information than I could have gathered15 years ago over the entire course of the job.
You may argue that the internet made the publishing of information possible and so I should credit the internet. And on the one hand I do. However, I believe that even in 1994 all information was to be had somewhere. Finding the head or document it was contained in however was hard. Without search engines, like Alta Vista in the old days, however getting to the data on the net was still hard. Less hard than without the internet, but still hard.
The impact of this is profound in a myriad of ways. The incredible speed with which new developments get disseminated are to an enormous extent the effect of Google Search. Building a new product in Silicon Valey used to be limited to the component available at Fry's Electronics. Now it is limited to the components you can find with Google. The ways of implementing a component used to be limited to your knowledge and skills. Now it is limited to what you can find of how other people have implemented it and how they solved their problems. This allows you to find communities of like minded people, cooperate, develop solutions together and the speed of innovation is increased drastically. A simple example is that it someone once estimated that the use of open source decreased the price of DSL modems by 5-10 dollars per modem.
All in all... the quality of our life has dramatically increased through Google Search. We've become better informed, make better choices and spend less time and energy on gaining knowledge. Time and distance have shrunk. We can know what's going on in Kenya, what room to rent in Bali, the way the Indian telecoms regulator deals with termination costs and that Gravity is a cool Twitter gadget for the Nokia E51. How much that's worth is impossible to calculate. Google Search may not show up in GDP stats. Not having it or a functional equivalent around would set us back to 1997.
UPDATE May 26th. Completely forgot to mention this is a similar paradox to the productivity paradox.
(thanks to Herman Wagter for pushing me to write down my thoughts)