Assessing the social value produced by founding Google


80,000 Hours has outlined some reasons tech entrepreneurship could be a particularly promising career path. One relevant factor is that the technology sector is a candidate for a sector of the economy that produces significantly more social value than its total earnings. Some reasons for this are:

  • Anecdotally, people report that they benefit substantially more from certain technologies than they pay for them. For example, Google provides services to Google users at the very low cost of unobtrusive advertisements, and Google users benefit substantially relative to this cost.

  • Technological innovation has been a large driver of economic growth, and economic growth helps people who haven’t been born yet, who don’t pay for past technological innovation.

As the beginnings of an attempt to estimate the social value of tech entrepreneurship relative to other high impact careers, in June 2013 I made a very rough attempt to estimate the value that Sergey Brin and Larry Page generated by starting Google. I attempted to estimate the value based on a subjective judgment of how much people would have been willing to pay to use Google historically, and a subjective judgment of how soon a counterfactual replacement to Google would have emerged had Google not existed.

Upon reflection, I think that my analysis is sufficiently crude so that it shouldn’t be given nontrivial weight. This is not unexpected — in writing the post I was soliciting feedback on my thinking rather than aiming to make a cogent argument. I benefited from feedback from Carl Shulman.

My main conclusion from this work is that I think that it is possible to give meaningful and nontrivial upper and lower bounds on the expected social value of a given tech company, but that doing so will be a very time consuming project that requires examining the question from many different angles.

Here are some additional considerations, that emerged in the course of discussion with others and upon further reflection, which will be relevant to future attempts to make such estimates:

  • If Google hadn’t been created, and a counterfactual replacement to Google emerged ‘n’ years later, it would initially resemble “Google except n years shifted forward,” but it’s unclear how long and to what extent the ‘n’ year gap would persist, as opposed to the counterfactual replacement “catching up” to Google over time.

  • It can be difficult to remember what the Internet was like when Google was started. Carl Shulman pointed out that the Internet had only 31% penetration in the developing world in 2000, and that at the time there were 100 times fewer searches relative to present day search rates. Note that the medium penetration in the developed world suggests that there may have been a sizable population of Internet users who were just barely willing to pay for the Internet at all, and which correspondingly wouldn’t have been willing to pay for improvements to search.

  • People used the Internet both at home and at work. The value of being able to use Google at work could have been higher or lower than the value of being able to use Google at home.

  • In order to be confident that customers would be willing to pay more than they do for a given product, one needs an explanation for why the company doesn’t charge them more for that product. There are explanations for why there’s some gap between what people were willing to pay and what Google was able to charge — for example, the transaction fee of paying for a cheap product (e.g. in terms of time spent paying) can make the de facto price of a good higher than what customers would be willing to pay for it in the abstract. But there’s a lot of uncertainty as to how large the gap is. There’s also uncertainty as to whether Google could have been charging more but chose not to. One key question for further investigation is that of whether tech companies provide their users with value out of proportion with what the users pay for their products, and if so, why they don’t charge more for their products.