"Why not put the data out there to be used in new and innovative ways by developers who can benefit from it?" That leading question is from Michele Hovet, IT director of Arvada, Colorado, and vice chair of the Government Shared Services Council.
She's right to ask. Everyone who has looked at the operations of state and local governments agrees there are enormous opportunities to serve citizens better by leveraging technology, particularly by opening data.
A move towards open data also presents difficulties, not the least of which is increased costs. A natural response to such challenges is to apply sound management techniques and leadership, but that, too, has limits, particularly in public service.
The examples that follow should help you find the factors likely to lead to success in your own situation.
Great opportunity
Governments hold enormous stores of valuable data. Favorite descriptions are invariably anecdotal: the 9-11 hijackers were all known to law enforcement agencies, but the data weren't properly correlated; NASA is famous for "civilian spin-offs" enabled by sharing the results of its research; and of course any home-owner who has managed property taxes or researched deeds on-line rather than wait a turn in an inconvenient office is bound to be a fan of new digital possibilities for government data.
While visits to DATA.GOV peaked during coverage of the 2011 Congressional budget debate, insiders say that actual dataset use has grown steadily.
Government data management goes back to our country’s beginning. The United States Constitution, brief though it is, finds space in Article I to specify a decennial "enumeration" (our Census) as a crucial element.
Quantification of these values has lagged. Substantial businesses have long "re-purposed" what's available from court proceedings, census publications, CIA atlases, and agency scientific and commercial compilations. It seems plausible that release of, say, crime statistics in Cook County, or water flows of the Colorado River, will be valuable to someone. Which datasets are worth processing first, though?
No one knows. While the European Community has gone so far as to estimate "a €40 billion boost to the EU's economy each year" from an open data strategy, the analytic literature is skimpy. Specialists widely believe what European Commission VP Neelie Kroes and others have said: "Your data is worth more if you give it away." As with many other IT issues, though the people in the best position to make such measurements are too busy creating the future to invest time rigorously justifying it.
Data can be “opened” in different ways, some distinctly more satisfying than others. It was natural that the first wave of the move to on-line services largely reproduced existing workflows. Early project goals encouraged patrons to "self-serve" rather then process all operations through a clerk, and then to allow that access remotely. Projects were implemented as websites. Tom Downey, the director of the Department of Excise and Licenses for the City and County of Denver, describes how these transformations continue in even the most mundane operations. With the right data and calculations, he says, a liquor license can be automatically processed in regard to such requirements as distance from schools or churches, deficiencies in the application, and so on. Everybody comes out ahead, including the department clerks whose time is liberated to higher-value service.
As much as on-line applications need to improve, the emphasis has shifted recently from the application to the data. As Edward Felten and his co-authors put it in the academic language of their Government Data and the Invisible Hand, "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."
A related decision: Should an agency collect and publish all the data it can, or catalogue locations for data already in place? There's little reason to believe that government's comparative advantage lies either in application development or even publishing; it certainly accesses massive datasets, though. The real wins, as Hovet eloquently argues, are when people find un-planned uses for data.
Hovet also sees benefits to sharing in "inward-facing" operations: "Our intranet site is an open source social media, document management site. Staff can create their own groups, and share what they want via discussions or docs or comment with whomever they want. We are trying to push collaboration and tear down the departmental communication walls." More and more, these services are outsourced, sometimes to the cloud, and sometimes to peer agencies. One vision is that a community of government units in similar situations can share expertise: One city specializes in geographic information systems (GIS), another in mobile application construction, and so on.
Why not to open up
As exciting as the "Open Gov" and "Gov 2.0" movements are, there are also reasons not to open government data. Budget constraints — of expertise as often as finances — block many openness projects. Privacy limits publication of certain data. Most of all, openness threatens the standing of some government actors who profit from "gating" access to data.
Each of these forces can derail any particular "openness" project. In the aggregate, though, it's clear that the trend is for governments to share more and more data.
Sound management
Complexity, large opportunity, tangled requirements, many actors: This sounds like a situation that calls for a bit of management (or operations research) theory.
That's exactly what has started to happen. Success has already been claimed for Compstat with the New York Police Department, along with variations across the country on Six Sigma and Total Quality Management, Peak Performance, Management By Objectives, Good to Great, and even Lean Management. Brian Gryth, Manager of the Operational Support Team (OST) of the Colorado Department of State, emphasized in an extended interview how little of OST's focus is technical. While individual milestones often appear as technical achievements — availability of a dozen new datasets in a particular e-gov catalogue, for instance — the day-to-day work of OST lies mostly in sound management of governments’ service-delivery operations.
"Sound management" is a particularly daunting task in government, if only for the difficulty in identification of the customer. Does a local Motor Vehicles branch serve:
- the drivers queued up in the waiting room to renew their licenses (focus: quick, courteous service)
- the district-level career administrator to whom it reports (focus: operate within budget, and avoid scandal)
- the state-level political official responsible for the Division (focus: re-election), or
- taxpayers and citizens at large (focus: lower taxes, safety on the roads)?
It's widely-recognized by now that "you can only manage what you measure," but what should the key performance indicators (KPIs) be in government operations? It can easily happen, moreover, that a particular project is a technical success, and even meets all its objectives, but is shuttered because of an unrelated election upheaval.
Hovet, as an IT manager, thinks of her "immediate customers [as] the internal staff in the departments who really are the responders to citizens." She also needs to answer to taxpayers, out-of-town shoppers, and the City Council. To juggle these separate demands, she says, "We use the RACI method most days to make sure we are not forgetting anyone." RACI is a mnemonic for the responsibility matrix: Responsible Accountable Consulted Informed.
Limits to methodology
As technologists, our training is to seek system and solution. We orient to the "shiny objects" of breakthrough technologies and smoothly-running offices. There's no doubt at an abstract level that governments at all levels should "liberate" more of the data they hold. On the other hand, it's no benefit to get too far ahead of our customers, however defined. A security technology or Web design that's more sophisticated than users can afford to appreciate does no one any good. A site for paying traffic fines doesn't need to be engaging or awesome or fantastic; it just needs to do its job and send people on their way with a minimum of fuss.
No management methodology gives ultimate answers, particularly not in government. We remain at a stage that demands human leadership and sensitivity to balance all the different dimensions at play in such an apparently-simple matter as reporting where and when utility crews repaired water and sewage breaks.
What management methodology can do, though, is structure our conversation. "Peak Performance" and the other systems get us talking and airing out what's on our minds. They expose assumptions, and get teams pointed in more-or-less the same direction. For now, that's the best help for the Open Gov, Gov 2.0, and e-Gov movements.
See also:
- Problems and Opportunities in Government Data
- Federal Government Announces Cloud, Records Management Initiatives
- Brenda Michelson’s Active Information blog
- Sharon Fisher’s Policy Watch blog
