The opportunities for European startups to access government open data via APIs to create innovative business models are yet to materialize. That's the read-between-the-lines message from a panel session hosted today at Internetdagarna, a two-day Swedish cloud industry event that will tomorrow see a full day's program hosted by NordicAPIs.
The panel, with speakers working either in or alongside ministries in Denmark, Germany, the UK, the Netherlands and the European Union presented current progress reports on open data infrastructure. None of the country progress reports demonstrated strong engagement with entrepreneurs, innovators, or citizen groups to make use of open data, once it is available. Many governments are still grappling with internal agency reluctance to open data, lack basic consensus on provisioning, and have not yet identified many positive impacts from their current open data efforts.
On the plus side, key policies are now in place that could be leveraged by individual European governments to push ahead with an open data agenda, and low-hanging fruit is becoming clearer for entrepreneurs wanting to commercialize around open data. Unfortunately, many of these opportunities are at this stage based around infrastructure and business development of the associated open data ecosystem, and, as yet, not directly related to using government open data APIs as components to create an innovative business product or service that could solve industry or community problems.
Panelists included Dr Malte Beyer-Katzenberger (Policy Officer at the European Commission), Cathrine Lippert (Special Advisor on Digitization in the Danish Ministry of Finance), Ton Ziljstra (a consultant working on open data projects in the Netherlands), Daniel Dietrich (from Open Knowledge Foundation's German office), and Richard Stirling (UK's Open Data Institute).
While each of the panelists showed a strong knowledge of the open data frameworks in their European areas, and all had an obvious passion for the benefits of government open data, the actual progress they were reporting showed a muddied, lackluster level of progress that is hopefully set to improve in 2014. Two of the biggest issues that most of the panelists shared included:
- An inconsistent focus on infrastructure
- Lack of focus on private sector partnerships.
An inconsistent focus on infrastructure
Several of the speakers mentioned inconsistent open data infrastructure efforts in their home region. Denmark, for example, had built an open data catalog in 2009 that has not been maintained since. The complexity in trying to create a shared momentum across government agencies in which there will not be a one-size-fits-all solution is one of the biggest barriers the national government faces in moving forward with open data. However, Cathrine Lippert's role as an open data evangelist in the Ministry of Finance is an innovation in itself, creating a new type of role within government to encourage open data solutions. Lippert believes "Open data is bound to happen, but it is in our interests to make it available faster and better." While machine readable (i.e. API-based) open data is recommended, it did not sound like a strong priority amongst current government efforts.
The Netherlands shares several of the same problems, particularly around an initial move towards open data that has resulted in a sense of disillusionment with little impact following open data releases. Ziljstra gave the example of one city government that had created an open data portal several years ago. After the initial investment, the government was disheartened enough by the lack of response that they did not continue funding its operations. Now, advocates push forward with freedom of information (FoI) requests in a community-led drive to get new data added to the portal, as FoI is funded from a different bucket of money than the previous (now non-existent) open data infrastructure source. In the Netherlands, FoI is taken more seriously by government agencies as an impetus to release data. Even when open data is made available, finding how to access it can be a significant barrier. In one test, it took 2 hours to locate an online source for postcode open data: not an uncommon experience for a national data portal made up of up to 90% dead-links.
Germany's open data problems also centered on a lack of consistent policy frameworks across the country and across various tiers of government. It is hoped that European Council directives will create a new push for an open data agenda, but to date, previous EU policies have had little weight internally in the country. Additionally FoI legislation on releasing data is riddled with exceptions, according to Daniel Dietrich, who has been analyzing the open data policy landscape in Germany for the Open Knowledge Foundation.
Lack of focus on private sector partnerships
The economic impact arguments around leveraging open government data are yet to translate into funding policies that provide incentives to startups to innovate and create new commercial solutions using open data. Despite evidence that open data supplied by the Dutch Meteo weather agency has sparked 100 new jobs and generated 35 million Euros in new tax revenue, or how the Finnish small and medium enterprise sector grow 15% faster than their counterparts when given access to geospatial data, or even how the UK National Health Service could save 200 million British Pounds a year by analyzing open data on prescribing practices, there is little sharing of stories that demonstrate how private partnerships can help make use of government open data to spur new economic activity. Many of the countries represented showed high scores on several measures of open data accessibility (for example, as measured by the Open Data Index), but all scored equally poorly on the Open Data Barometer which calculates civic and business impacts from government open data supply.
When questioned by ProgrammableWeb, few of the panelists could point to any national government funding incentives that empower startups to create new business models based on open data. However, the new G8 Open Data Charter has given European member countries the policy weight now to work towards removing transaction fees when providing open data to potential users. In many countries, open data supply has been charged at the full cost of provision rate, whereas the new G8 Charter and associated directives promote the use of a marginal cost of dissemination. Dietrich says that in Germany, at least, charging for the full cost of making open data available has been "why businesses have not taken up the promises of open data" to date. The marginal cost now means that governments must cap what they charge for making the data available to match the cost of the specific data they are supplying, and not factor in the total costs of production, as has often been the practice in the past.
What was surprising in hearing several of the panelists was the lack of knowledge exchange between government agencies and private sector companies who are dealing with similar problems when implementing API strategies. For example, Lippert and Ziljstra both spoke of the need to match current government planning goals to open data sources in order to better make the argument for open data supply within government circles. It is not unlike comments made by AT&T's Laura Merling about how, in order to implement an enterprise-wide API strategy, her team needed to locate the subject matter experts in each of the enterprise's business departments. The lack of engagement between private and public sector around open data limits the sharing of these sort of best/effective practices. It is a bit like how Kin Lane was talking about techniques to wrap government open data into APIs in a workshop at API Strategy and Practice last October, while Jason Lobel was doing the same with businesses in the first wave of data-related products from SwiftIQ. There is plenty that both parties can learn from each other when they are talking in the same room together.
Low hanging fruit
For startups looking for government initiatives that provide seed funding or startup incentives to use government open data via APIs to create a new business opportunity: you probably need to wait a bit longer.
Open data is not considered an industry for government to invest in just yet, beyond the occasional civic hackathon. One interesting model was from a Dutch regional government that tendered out to the private sector seeking open data solutions to traffic congestion problems is an all too rare example at present, it seems.
But there may be room for entrepreneurs with an open data mindset to create new business in Europe loosely related to open data. Dr Malte Beyer-Katzenberger from the European Commission highlighted new efforts to build a Pan European open data portal, which is "strongly in need of visualization tools". API developers with expertise in web-scraping or in wrapping open data formats into APIs may find there is an increasing amount of work as governments tend to upload open data in any format, mostly not machine readable, to their initial portals and catalog-based sites. MapBox has had success with this sort of business model. They have been able to clean open data, re-build it in more useful ways and then sell it back to government, for example. Those with a keen knowledge of what data is available, or who can help private industry - or even city governments or department agencies - to map open data against policy goals may also be able to hang their shingles a little higher. For now, other potential market opportunities seem to be centered more on providing training on open data.
While on some levels it was disappointing to realize that much of the rhetoric about the new possibilities that open data will create are still some way off, the good news for those with an entrepreneurial zest is that the scope for designing new business models and creating new market opportunities are only just beginning.