The idea of E-government initiatives is to make it easier for citizens to transact their business with their governments. This is surely a good idea, but it carries with it several problems including endangering the long term preservation of government information.
Take, for example, Adobe’s new product, The Adobe Digital Enterprise Platform for Customer Experience Management (CEM), which it hopes will attract government agencies:
- New Adobe platform would personalize an agency website for individual users, by Joseph Marks, NextGov (06/20/2011).
A new Adobe product unveiled Monday would allow a federal agency to tailor its website and customer service operations to specific citizens based on their geographic location and, perhaps, their past contact with the agency.
This sounds attractive in a lot of ways. It promises better customer service, personalized information, and faster access to relevant information.
There are, however, several problems if this approach is used exclusively.
Citizens or Customers?
Adobe says that its software allows agencies to “stop making a distinction between customers and citizens.” Surely all of us would like to know that our “customer experience” with the DMV would be as easy and straightforward (and brief!) as our experiences with the best commercial web sites. Companies like Amazon have made their fortunes not because they offer better products, but because they make it easier to find and buy those products. Wouldn’t it be nice to have have government agencies’ web sites work as well as the best commercial web sites? Wouldn’t it be great if government agencies could shrug off the old, cliched unfriendly-to-users image, and create new, user-friendly, customer-centered web sites?
It would, of course. But the problem is that, when we visit an agency web site, we are not always the “customer” of that agency. We are more often citizens seeking information than we are customers engaging in business-like transactions.
And that is the beginning of the problem. Treating citizens as customers can jeopardize our privacy, make it harder for us to find the information we want, and make it harder to preserve government information for the future.
There are, of course, big privacy issues if governments start replacing the dissemination of information with the personalized transactions of e-government. Citizens should be able to search, browse for, read, and use government information without the government tracking and recording each individual’s every search and use. The e-gov interaction between citizen and agency requires just such tracking, however. Adobe, for example, says of its product that it would provide “an instant, unified record of a customer’s interaction with a company or agency, regardless of where or how that interaction is happening.”
There are certainly occasions when citizens want and need to personally interact with a government, but we probably all hope that we don’t have to do this frequently. Going to the DMV to renew a license, or applying for a grant, or filing tax returns are not what we do (or want to do!) every day. These are the exceptions to our interactions with governments.
Most of our interactions with governments are about looking for information that the government has gathered, or compiled, or created as part of its mission. Whether it is about proposed legislation, or existing regulations, or the location of flood-plains, or the population of a city, or the latest economic indicators, or how to manage agricultural pests, the government does not need to know who we are and what we are looking for in order to deliver the information we need.
If governments replace the anonymous delivery of information with e-government “customer-based” services, we will lose our ability to read (or even look for information) privately. (If you are not convinced that privacy is important, see Privacy: “I have nothing to hide”.)
Filering out what we want.
It seems almost counter-intuitive to say that personalization of web sites would make it harder to find what we need. Surely, personalization is designed to make it easier to find what we want, isn’t it? Take the example mentioned in the NextGov article above: Imagine you live in an area that has just been damaged by a flood or a hurricane and you go to the FEMA home page. Wouldn’t it be great to have the website “know” where you live and immediately show you links to specific services available and relevant to you? It would, but this example is not typical of all our interactions with government and therein lies a problem for relying only on customization.
Those who examine how people use the web have long understood and documented that customization of search results and browsing can do more to limit our understanding than enhance it. See, for example, Nicholas Negroponte or David Weinberger in 1995, and J.D. Lasica or Cass Sunstein in 2001.
And now Eli Pariser has written a book (The Filter Bubble: What the Internet Is Hiding from You) that documents how “the hidden rise of personalization on the Internet is controlling — and limiting — the information we consume.” Pariser says that, when we are seeking information, “personalization” silently filters out relevant content as it tries to predict what we “want.” (See Pariser’s excellent TED talk for more details and examples: Lunchtime listen: Eli Pariser on filter bubbles.)
Do we want governments to favor “customers” who require “personalization” over citizens who are seeking information? I worry that such an approach will likely lead to government web sites that silently filter out relevant search results in an attempt to show you what the government (or Adobe) thinks you want. If we do not know how this process works and if we have no control over whether or not to use this functionality, we will end up not knowing if we have found what would be most relevant to our information needs. This would be bad. As we know, If It Is Too Inconvenient, I’m Not Going After It. Citizens seeking a broad array of information are not the same as customers wanting to buy a single product. Governments delivering a cornucopia of information are not the same as businesses trying to persuade customers to buy the shiniest, newest, highest-profit-generating product.
How do you preserve something that you can’t get?
As we move to the delivery of government information through dynamic web sites (whether “customer” driven or not), we face an increasing problem of preserving that information because we have no direct access to the information that needs to be preserved.
In order to preserve information (even digital information), we require an “instantiation” of that information — a digital object to preserve. In the past, information was instantiated in physical books, pamphlets, maps, journals, posters, and even microfiche, CD-ROMs, and DVDs. Although some government information today is instantiated in PDFs and spreadsheets and even static web pages, government information is increasingly instantiated in databases that are not directly visible to users. When libraries (or GPO) cannot get copies of these databases, they cannot preserve them.
Websites use those databases to present selected information to users who visit web pages or who request information through searches. “Customer-driven” web sites will ensure that two people who make the same query or who visit the same URL will get different information. (To use the FEMA example again, if I live where there was just a flood and you live where there was just a hurricane and we both visit a customer-driven FEMA web site, I’ll get flood information and you’ll get hurricane information.) Web harvesting will be insufficient for preservation under these conditions.
The essential problem here is that, if agencies see their information mission as one of processing transactions with individuals rather than one of creating and delivering preservable instantiations of information, it will be difficult if not impossible for digital preservation to be complete or accurate or successful. Gertrude Stein might say of the lack of access to preservable digital objects, “There is no there there.”
At FGI, we are not technological determinists. We don’t believe that the existence of software such as Adobe’s CEM will inevitably lead to loss of privacy, harder to find information, and the inability of libraries to preserve digital government information. As noted above, there are circumstances where use of such software could yield better service and make it easier for users to find the information they need. We believe that thoughtful management of digital technologies can result in easier access and new functionalities without sacrificing long-term, free, public access to government information.
We also know, however, that technology is political and that sometimes organizations make bad technological decisions for apparently necessary reasons. Our concern is that agencies are under pressures that could easily lead to bad decisions and that software such as Adobe’s CEM could make it easier to make bad decisions. Specifically, agencies are under pressure to reduce the number of government web sites and to streamline existing websites using new technology-based plans to improve their customer service at the same time that budgets are under increasing stress, open government initiatives are being reduced drastically, and GPO is being hit by big budget cuts.
Our concern is that these pressures will result in bad decisions. We worry that agencies will not add new, much-needed functionality to existing web sites, but will instead replace a citizen-centered model with a customer-center model. We worry that such a substitution will result in a loss of privacy, a loss in information-based functionality in favor of product-based functionality, and that all of this will make it even harder than it is already to preserve digital government information.
We agree with OMB Watch that information is a customer service and hope that agencies will keep this in mind when they make their information-technology decisions. But we worry that hard-pressed agencies will not.
We believe that those (including GPO and FDLP libraries) who are interested in preserving government information should address the task of preserving the databases of government information that drive dynamic web sites. The information behind even customer-based transactions needs to be preserved; the transactions themselves do not. A model for this already exists with Census data. The Census Bureau has been able to provide a dynamic, database-driven web site and, at the same time, provide the databases behind the web site as preservable digital objects.
Preserving databases is a more complex task than preserving monographs or PDFs (see for example The Preservation of Databases, by Kevin Ashley, Vine, 34 (2004), 66-70), and agencies that personalize the delivery of information will have to ensure that personal information is kept separate from agency information, but database preservation and privacy protection can be accomplished. But to do so will take an active commitment.