Earlier this week, David Eaves kicked off a fascinating conversation with a post on TechPresident. Titled “Optimism, Fear and the Knight News Challenge,” it raises important questions about how open government work is supported and sustained. In particular, David focused on Democracy Map, one of two KNC finalist projects from friend-of-Sunlight Phil Ashlock. Democracy Map aims to improve U.S. citizens’ ability to determine who represents them at all levels of government. David argues that a subsidy from Knight to DM could threaten the business of companies like Cicero that are trying to solve the problem through a commercial offering. Once the Knight money dries up, will Democracy Map still be around? Or will it only last long enough to kill off Cicero?

Sunlight’s Eric Mill responded with a comment, prompting a fascinating response from Cicero’s Robert Cheetham. And yesterday Phil weighed in with a lengthy and compelling post that’s well worth a read.

Believe me: I know that sustainability is important. Technological improvements might mean that costs tend to fall, but it still takes money to keep data fresh and available. I’m supposed to be writing a grant proposal right now instead of this blog post, in fact (don’t tell Ellen).

But in this case I have to disagree with David. For one thing, the presence of an existing business is a terrible reason for not doing something worthwhile. Phil rightly points to the case of Intuit keeping the tax-filing experience miserable, but there are plenty of other examples, from why my public transport options for football games stink to why it’s so hard for governments to address climate change. If you had to sum up all of our government’s dysfunction, “fear of upsetting incumbent players” wouldn’t be a bad first pass.

Besides, in this case it’s not clear that Cicero actually can solve the problem. The data they’re publishing is an extremely valuable contribution. But Robert admits that Cicero doesn’t turn a profit. That’s not his fault — information about government is a classic public good. This is Econ 101 stuff, but it’s really important. I tweeted a link to this post a week ago; in one paragraph it neatly sums up the key points: public goods are available to everyone, and they don’t get used up. This makes them really hard to charge for, and as a result, the free market usually doesn’t produce enough of them. Cicero is selling a service and producing open government data as a byproduct.

But over the long term, you can’t make money charging for open data — not unless you start making it less open and therefore more scarce. Historically, that was the most common outcome when government relied on the free market to distribute its information. You can see that legacy today in the Elseviers, Thomson Reuterses and LexisNexii of the world. Those partnerships all did great work to organize and spread information in the contexts of their times. But technological possibilities and citizens’ expectations have continued to advance. The business models, information silos and paywalls built into these arrangements don’t.

Good work is being done to untangle those legacies, and of course many players in the private sector have embraced openness without reservation, Cicero included. And naturally, they aren’t the only ones doing good work on the open government data problem that Democracy Map aims to tackle — so are other vendors (Granicus comes to mind), other nonprofits (*cough* *cough*), and governments themselves. Establishing an expectation that our work will be shared openly means that the problem doesn’t have to be completely solved by any one party. It also means that if one player falters, someone else can step into their place. When it works well, it’s downright magical. Smart companies like BillTrack50 are using resources like our Open States project and the github.com/unitedstates repo to cut their costs. We’re giving away this data, so they can’t sell it. But they were never going to have much luck selling it anyway. They can build services and products on top of open data, though, all while avoiding having to reinvent the wheel. This is a strategy that clever startups have used over and over, from BrightScope and SEC data to Mapbox and USGS data to Yelp and health inspection data (David had a lot to do with this last example, in fact).

It’s true that this arguably doesn’t solve the sustainability problem. It spreads out the risk, but the issue is still there. But there’s really no solution to this other than convincing government to take on responsibility for providing an important public good. And we can! It’s happened before. It happened when Carl Malamud opened EDGAR. It happened when the Center for Effective Government (then called OMB Watch) got federal spending data online (full disclosure: Sunlight helped support this work). We can see it happening when state officials pledge improvements to their data in response to our own Open States scorecard. In some of those cases, like EDGAR, this exact debate occurred.

But it’s often true that before government embraces its responsibility to provide a public good, the underlying problem has to be shown to be solvable and the solution proven useful. It often takes a lot of people, approaches and ideas to achieve that — and resources. For our part, we welcome the help of everyone who wants to pitch in.