Bad Data is Limiting the Potential of Govtech
In the process of conducting several govtech founder interviews, a consistent theme has emerged. Every one of these companies is trying to solve a problem that originates with bad data. The data is either (a) not accessible at all, (b) accessible but unusable, (c) takes significant work to make usable, or (d) usable but outdated. Bad data is severely limiting the potential of govtech.
The open data movement was supposed to address many of these issues. By aggregating data across an organization into a single place, open data portals (in theory) would make the data accessible and available in a modern web format. But these portals have simply added another product between the end user and the system of record — a data broker of sorts. The data broker charges a fee to duplicate data already stored in other systems and makes it available in a single place. The outcome of this work isn’t higher quality and more usable data, it’s just public catalogues of the same bad data.
This bad data resides in legacy systems of record. When most of the SORs were installed, data wasn’t even considered an asset. The portal vendors rely heavily on a manual process, agency staff exporting data from the SOR and importing data into the open data portal. This leaves very little leverage for vendors of open data portals to dictate data quality and update frequency.
The issue of course; most of these legacy companies are closed systems. Many of them now offer data portals (displaying their own data) of their own (i.e. ESRI). They are orders of magnitude larger, control much bigger distribution channels, and have huge client footprints. There is very little incentive for them to normalize their data and make it easier to access by third parties. Their unwillingness to provide integrations pushed open data vendors into a strange existence. Many of them existed simply to solve a bug for SORs, inaccessible data.
In a perfect world, open data portals would operate more like Slack. Light on the functionality, heavy on the integration. But, growing competition in the space is driving companies to build more functionality instead of integrations. Need proof? Most of the companies that started as vendors of open data portals now offer dashboards and analytic tools. Companies that started selling dashboards and analytic tools now offer open data portals. Maybe someone should just make all of this shit work together instead of building more shit.
A possible alternative to a Slack-like solution would be a true operating system for government. Operating systems, at scale, provide a set of baseline conditions that ensure continuity in the user experience. Like Slack, operating systems create leverage through distribution. Build a big enough audience and leverage follows. When you bring an audience, most people are willing to abide by a set of rules. Those rules can be mutually beneficial to everyone involved. Will it be perfect, of course not. Will it better, seems likely.
Until this problem is solved, a significant portion of the money being invested in the space will continue to be allocated towards data cleanup. That seems like a waste of money and a waste of an opportunity.
Interested in getting a curated selection of the two most important things to read in #govtech each week? Sign up here.
