The value of a good is much trickier to measure than measuring the cost.  Value involves considering a hypothetical question – what would your life be like without that good?




Economists commonly use two measures to assign monetary value to some good or service: the “compensating variation” and the “equivalent variation”. The compensating variation asks how much money we would have to give a person to make up for taking the good away from them while the equivalent variation asks how much money someone would give up to acquire the good in question. The term “consumer surplus” refers to an approximation to these theoretically ideal measures.

If we want to estimate “the value of Google search” we have to look at both the commercial and non-commercial aspects of search: users are searching for answers to questions (some of which are commercial in nature) and advertisers are searching for customers for (mostly) commercial transactions. So is useful to break the problem up into two pieces: the value of ads to advertisers and publishers, and the value of search results to users.

Suppose Google were to disappear tomorrow. In the first instance, advertisers and publishers would lose a lot of visitors to their web sites.  How much are those lost visitors worth to them? This is the question I tried to answer in the “value of Google” study. The tricky part is ascribing a value to the advertisers of those web site visitors, but it turns out there is a way to infer that value from advertising bidding behavior. This allows us to get a back-of-the-envelope estimate of the immediate loss in value from Google vanishing.

Of course, “if Google did not exist, man would have to invent it”. So we would expect that as the weeks went on, users would start to use other search engines, and advertisers would spend advertising dollars in different ways, and publishers would find other ways to get ad dollars for their web sites.

So the long-run loss in value would be substantially less than the immediate impact. Ultimately the lost value would be the difference between Google and the next best advertising alternative but that would be almost impossible to estimate given the available data.

Turning to the user side, we drew on the work of Yan Chen, Grace Young Joo Jeon and Yong-Mi Kim of the University of Michigan to estimate the value of online search in general.

Some of us are old enough to remember what life was like before search engines. We had to look through a pile of reference books to find the answers to basic questions. Even small questions, like how to spell a word, or whether it was likely to the rain the next day, required some effort to answer. Even trivia was hardly trivial: finding obscure facts involved substantial research.

So one way to measure the value of online search would be to measure how much time it saves us compared to methods we used in the bad old days before Google. Based on a random sample of Google queries, the UM researchers found that answering them using the library took about 22 minutes while answering them using Google took 7 minutes. Overall, Google saved 15 minutes of time. (This calculation ignores the cost of actually going to the library, which in some cases was quite substantial. The UM authors also looked at questions posed to reference librarians as well and got a similar estimate of time saved.)

I attempted to convert this time to dollar savings using the average wage and came up with about $500 per adult worker per year. This may seem like a lot, but it works out to just $1.37 a day. I would guess that most readers of this blog get $1.37 worth of value per day out of their search engine use.

When doing this calculation, it is important to take account of the fact that since the cost of getting answers is now so low, we ask a lot more questions. When getting an answer involved a trip to the library and 22 minutes to answer an average question, we only attempted to get answers to important questions. Now that it involves only a few minutes at a search engine to answer questions, we ask many more—and a lot less valuable—questions.

Said another way, we wouldn’t bother to even to go to the library unless we were willing to spend at least 22 minutes (on average) to find the answer. Now that it takes us only a few seconds or minutes to get an answer, we ask a lot of frivolous questions (along with the important ones, of course.)

Taking this effect into account involves estimating a “demand curve for questions” as a function of the “cost of getting answers”. I don’t know any serious research on this topic, so I made a rough approximation to that demand curve and came up with the $1.37 a day figure. It could be larger or it could be smaller, but I think that is the right order of magnitude.

There are many other ways to estimate the value of the internet and the services it provides. However, to the extent that they are based on current practices, they likely underestimate the long-term value of the internet.

It is now possible for everyone on the planet to have access to all the information humans have ever produced.  The barriers to this utopian dream are not technological, but legal and economic.  When we manage to solve these problems, we will be able to unlock vast pools of human potential that have hitherto been inaccessible.  In the future this will be viewed as a turning point in human history, and economic advances generated by global access to all information will be recognised as the true value of the internet.

Photo credit: PC World

Via The Economist