Google User Privacy

Keith Sharfman —  19 February 2006

I have some questions about Google’s reluctance to turn over user search data to the government. (For background on this story, see here.)

1. Why does Google gather this data to begin with? Wouldn’t the best way to protect user privacy be not to save this information in the first place? Why does Google need to know the identities of those who use the Google product?

2. Does Google release the information it collects to private third parties such as advertisers? If so, why the public-private distinction? Aren’t the same privacy interests implicated in both the private and public contexts? Why not ask the users themselves when they contract with search engines about the circumstances under which their data may be disclosed to third parties? In the early days of the Web, I published a paper, “Regulating Cyberactivity Disclosures: A Contractarian Approach,” 1996 University of Chicago Legal Forum 639, arguing that disclosure of a user’s “cyberactivities” to private third parties should be permitted as a default matter, so long as users have the affirmative right to opt for non-disclosure. Additional “public disclosure” language could easily be added to such opt out clauses.

3. If Google and other search engines such as Yahoo continue to collect user data, then shouldn’t there be a market opportunity for a search engine to break away from the pack and offer users the guarantee of privacy by simply not collecting user data to begin with? If this were to happen, it would be interesting to see which data collection procedure will end up becoming the norm. If enough users vote with their keystrokes in favor a “no collection” regime, one could imagine all the search engines feeling some pressure to adopt such a policy. But it’s equally plausible to think that many users will not care very much one way or the other, in which case users who are idiosyncratically private will have the “no collection” option but the majority of other users will stay with search engines that continue to collect search data.

6 responses to Google User Privacy

  1. 

    I would argue that this is indeed a barrier to entry. Google was able to gradually improve its search engine because it started in an environment where search engines were much less developed, so it was possible to benefit by making improvements incrementally. Modern rivals, who no longer exist in such an environment, would not be able to do the same thing and would instead have to make all the improvements at once to be able to compete.

  2. 

    Ken, I thought your initial reference to size Comment 1, First para.) was referring to demand side economies of scale, or network effects. Apparently not.

    But, this seems to be a different argument: Google’s investments into optimizing its search engine and providing a better product are themselves a barrier to entry. Again, I respectfully dissent. The theory that the rivals cannot engage in the same sorts of investments doesn’t seem to make sense either. Any competitive advantage earned by these sorts of investments are earned on a level playing field. These are certainly not barriers to entry in the Stigler sense (i.e. costs faced by entrant but not by incumbent), no more than lower costs, lower prices, or superior product quality would (or should) be.

  3. 

    “Big” is of course a vague description. Google is “big” in a relevant sense not only because it indexes a lot of pages, but because it’s used a lot of time and money trying to optimize its search engine. Rivals have not done so to the same degree, so in practice any rival, even if it already indexes a lot of pages, still faces a barrier to entry. (And this isn’t a network effect, either. A network effect happens when having more customers makes it more useful to each customer. This is not directly about the number of customers, but the money and research pumped into the search engine.)

  4. 

    Yes, I think Josh has it exactly right. Ken is correct that there are barriers to entry associated with creating a search engine. But as Josh says, this would not preclude an existing engine from adopting a “no collection” policy. And if users in fact prefer search engines of this sort, that will likely become the majoritarian default. Advertisers won’t pay much to a search engine that can’t retain its users!

  5. 

    Ken, what about current rivals in the search engine market? If consumers value such a privacy guarantee, could not an incumbent create a competitive advantage by delivering what consumers want? It seems like you would need more than a mere networks effects story to conclude that consumers would value such a privacy guarantee, but none of the competitors can gain by offering it. Any other thoughts on how to get there?

    You seem to making a related point in the second paragraph, but I am not quite sure what you mean there. You write that “if satisfying advertisers produces more income for search engines than satsifying users, search engines that satisfy users might not be able to compete no matter how many users refuse to use privacy-invading search engines.”

    Without more, I do not think this makes economic sense. Google benefits from ads because it has a large consumer base (which means advertisers are willing to pay more for this access). If a bad privacy policy reduces the size of its base, of course, its ad revenues will decrease. In other words, a search engine that simply ignores consumer preferences will not receive a free ride.

  6. 

    shouldn’t there be a market opportunity for a search engine to break away from the pack and offer users the guarantee of privacy by simply not collecting user data to begin with?

    No. First, creating a search engine has a huge barrier to entry. You can compete with a big restaurant by starting a small restaurant, but you can’t do the same with a search engine. A search engine is more useful the bigger it is.

    Second, the user data is valuable to Google because, as you point out, it’s worth money to advertisers. If satisfying advertisers produces more income for search engines than satisfying users, search engines that satisfy users might not be able to compete no matter how many users refuse to use privacy-invading search engines.