purple motes

Concern about too much government control over technologically limited and costly communication channels has been enormously significant historically. With the Internet revolution, governments can own and control communication channels without significantly lessening the opportunities for non-governmental bodies to do so. Governments that broadly disseminate government-created content do not preclude others from broadly disseminating other content. Vertically integrated government communication now carries much less political risk for the over-all communications industry. This fundamental change, it seems to me, favors more vertical integration in government communication with the public.

A draft of a new scholarly article makes the opposite argument. It declares:

If the next Presidential administration really wants to embrace the potential of Internet-enabled government transparency, it should follow a counter-intuitive but ultimately compelling strategy: reduce the federal role in presenting important government information to citizens. ... Rather than struggling, as it currently does, to design sites that meet each end-user need, we argue that the executive branch should focus on creating a simple, reliable and publicly accessible infrastructure that exposes the underlying data. Private actors, either nonprofit or commercial, are better suited to deliver government information to citizens....[1]

The idea essentially is to have more vertical disintegration in government communication. Government would focus on providing a large amount of detailed, machine-interpretable data that other organizations' technologies would search, aggregate, re-organize, and re-use. The anticipated benefit is more rapid innovation in the provision of information services to citizens.

Some efforts to promote vertical integration clearly are silly. The Yale Journal of Law and Technology (YJOLT) will publish the draft article quoted above in Fall 2008. The draft is freely and publicly available from the websites of SSRN and YJOLT. Yet on the top of every page of the article appears the bolded imperative "Do NOT cite." That literally implies that everyone can read the draft article but no one can discuss it. Many blogs have simply ignored the draft's pagely imperative (see, e.g., here, here, here, and here). One sheepishly declared: "it kindly asks us not to cite the draft, but - since it's out there for everyone to read - I assume a little quoting in a blog post like this is in order."

Wanting to respect the authors' wishes, I emailed them to ask if they would mind if I were to discuss their paper, cite it as a draft, and link to it. One of the author's responded graciously. He thanked me for my note, explained that YJOLT required the header, and welcomed me to discuss the draft and link to it. That's a good response. Allowing persons to discuss what they read increases the value of the time they spend reading. Moreover, the value of publishing an article in YJOLT isn't reduced by allowing discussion of the draft. Attempting to deny readers the freedom to cite a publicly available draft is an absurd product of an organizational silo-mentality. Fortunately, the specific issue is relatively easy to deal with in practice.[2]

The more general and important issue concerns supply incentives. With respect to government data, more important than the allocation of resources between government data infrastructure and government provision of data to individual end-users is the extent of investment in producing, cleaning, organizing, maintaining, and studying data. Government data collection typically is initiated to serve a narrow political purpose. Concern about specific statistics and the use of the data to produce specific reports drives investment in ensuring accurate reporting, finding and resolving data inconsistencies, and maintaining the data over time. A data collection effort that expands over time to serve diverse political interests within government has a better chance of enduring. To the extent that government data collection mainly serves non-governmental information intermediaries, governments will invest less in collecting data and ensuring high data quality.

Governments have significant advantages as suppliers of web content and services to end-users. Most adults know the names of the governments to which they are subject, have experience with those governments' services, and are concerned to make those services better. Governments typically spend little on user acquisition (many even aggressively discourage immigration) and relatively little on advertising and promoting themselves and their services. For example, U.S. federal government expenditure amounts to about 20% of GDP, but U.S. government advertising spending probably amounts to less than 1% of total U.S. advertising spending. Governments have a highly differentiated position within the space of user trust, and governments generate distinctive information flows. Eliminating governments from the ecology of end-user web content and services would waste their special institutional advantages.[3]

Stimulating end-user demand for government information is likely to make more government data available through information intermediaries. In academia, scholars who generate and share large amounts of data typically get relatively little academic credit, prestige, and status. Not surprisingly, only a small number of heroic academics pursue this unpropitious path. Even initiatives to require scholars to share data and algorithms necessary to replicate their published results have not been widely successful. However, the small share of scholars whose results attract considerable attention naturally generate demand for the data that they used. Moreover, these scholars then have some interest in ensuring that the data they used are widely available. The same dynamic is likely to be operative for governments. But the information flow is likely to be larger, because governments have a greater responsibility to supply demands for data and are less capable of controlling access to it.

Useful government data will get out one way or another. More important is to ensure that governments have an incentive to generate it.

Notes:

[1] From abstract of Robinson, David, Yu, Harlan, Zeller, William P. and Felten, Edward W., "Government Data and the Invisible Hand" . Yale Journal of Law & Technology, Vol. 11, 2008 Available at SSRN: http://ssrn.com/abstract=1138083

[2] I didn't try to contact YJLOT and get permission from YJLOT to cite the paper. When I'm not wearing my bureaucratic hat, I'm more concerned to respect the desires of human persons than those of corporate persons. That's particularly true when those desires seem to me silly or not in the public interest.

[3] As bright discussion of id. has highlighted, the distinctive characteristics of government also include distinctive forms of end-user political accountability.

Offer Comments

The proposed U.S. federal government budget for 2009 cuts funding for the American Time Use Survey. The survey program costs $6 million per year to administer for its full sample design. That seems to me to be a small amount to spend for high-quality, publicly available data very relevant for understanding value generation in the digital economy, long-term changes in media use, and key current issues in the communications industry.

You can sign a letter here to support restoring funding for the American Time Use Survey. Doing so won't take much of your time.

Offer Comments

The government of Washington, DC, provides real-time data feeds on crime, building permits, housing code enforcement, public space permits, and property registrations. The data includes geo-codes so that it can be easily mapped.

The terms of use for these important data resources state:

Neither the District of Columbia Government nor the Office of the Chief Technology Officer (OCTO) makes any claims as to the completeness, accuracy or content of any data contained in this application; makes any representation of any kind, including, but not limited to, warranty of the accuracy or fitness for a particular use; nor are any such warranties to be implied or inferred with respect to the information or data furnished herein.

So you can look at the data it provides, but the DC government makes no claims as to the "content of any data." In presenting these data, the DC government does not make "any representation of any kind." If that were literally true, then there is no data. If there's no data, then there's no liability. That's a clever way to share data to foster a more informed public.

In the terms of use for this blog, I had set out a less artful disclaimer. I've now appended to it essentially the above text. Read it and weep for our legal culture.

1 Comment

While making cars is a dying industry in the U.S., collecting car maintenance data seems to be a quite promising business. Jiffy Lube has 2,200 North American service centers servicing about 27.5 million customers per year. Jiffy Lube does "fast lube," i.e. quick oil changes for drive-in customers. But Jiffy Lube is also in the data business:

Jiffy Lube® also uses state-of-the-art computing technology to educate customers about vehicle maintenance services, share customers' maintenance histories across its network, and provide services that satisfy vehicle manufacturers' warranty requirements. This gives drivers the freedom to visit any Jiffy Lube® service center with the peace of mind that their records can travel with them. [from Jiffy Lube's History & Mission]

"Educate customers about vehicle maintenance services" means sell services in addition to oil changes based on collected data about a car's maintenance history. Those service are offered at a time when it's highly convenient for the customer to buy the services. The customer's car is right there at Jiffy Lube, ready to be served. That's a propitious action circumstance for personalized, relevant advertising, much like that of text ads displayed in the context of web search.

The automobile industry could do much more to develop its data businesses. Cars generate a large amount of performance data that could be downloaded at maintenance stops. Establishing open standards for such data and making it easy for car owners to grant anyone access to their car's data could enable considerable value in data services. Car service centers could sell a wider range of more accurately targeted maintenance services. Gas stations might sell personalized mixes of gas optimized for the car's driving pattern along with reports on fuel mileage history. Certification and evaluation services in the used car market would be more valuable with much additional car usage data beyond car mileage. Storing, sharing, and processing data is cheap and continually getting cheaper. Businesses that aren't thinking about how to create data businesses aren't learning from Google.

Note: Trust is important for creating value from data. "Don't be evil" makes particularly good business sense for data businesses.

2 Comments

Social networks are having difficulty generating advertising revenue (more discussion). Search is valuable for advertising because ads can help users get what they are seeking in their search. On social networking sites, users primarily engage in presence-oriented communication. If ads in those circumstances are even noticed, they are mainly an undesired distraction.

Social networks need to encourage their users to generate ad-relevant circumstances. Recommendations from friends is a major influence on purchasing behavior. The challenge is to organize "recommendations from friends" within social networks so as to add value to those recommendations and not undermine the non-instrumental quality of friendship. Doing this may require social networks to integrate into more field-specific sharing of information.

Consider a dining record service. Users would be provided with tools to build a database of what, where, and when they ate out in a restaurant. Users are likely to find this much easier than writing a restaurant review. Moreover, users are more likely to recognize the value to themselves of creating a record of their dining activities. It would be like the highly successfully LibraryThing, but instead of what you have in your library, this service would be about what you've put in your stomach.

A social network with a large number of users would a valuable asset for building a dining record service. Many web entrepreneurs could easily create a dining record service. But database scale (e.g. for identifying local patterns of taste) would be a key competitive advantage. Social networks could help build scale for a dining record service because dining together is such a socially significant experience. The total number of meals eaten together is probably one of the best indices of real friendship. Not inviting someone to dinner, and then plotting to make sure that they know they weren't invited, or that don't learn that they weren't invited, is ubiquitous social drama. Social networks have the sort of tools for enabling, sharing, and managing that social drama.

Dining record pages would be valuable action inventory for advertising. A user creating or reviewing records about where she or he has eaten would probably value advertisements about other places to eat. She or he might then also value diet ads. And exercise ads! The more you exercise, the more times you can go to that favorite restaurant of yours.

To be successful, social networks may need a more integrated and specialized information organization than widget-based applications can deliver.

Offer Comments