This blog first appeared on Policy Exchange’s PolicyBytes.
Data is the single most vital ingredient for good policymaking and the effective delivery of public services. It provides policymakers with the evidence they need to identify, define, and find solutions to issues both large and small. For public services, having the right data enables the targeting of finite resources at areas of greatest need, and offers the chance to track and drive improvements in performance. Put simply, data is the fuel of government. All other things being equal, the more data the better.
It is true that government is not short of facts and figures. HM Revenue and Customs alone is thought to have more than 80 times as much data as the British Library. Yet the data at its disposal is still just a tiny fraction of the total it could harness if it could tap into the collective wisdom of the nation as a whole. People already contribute their knowledge to websites like Wikipedia; provide data to create Google’s live traffic maps simply by keeping their Android smartphones with them; and freely offer their ideas in debates on social media. What if that same kind of knowledge aggregation could be harnessed to make government better?
Thanks to technology, it can.
To date, government has failed to take advantage of the radical improvements that technology – particularly the internet and mobile devices – have brought to communications and data sharing. Below are outlined the key shortcomings of its current approach to data in public service delivery and policymaking.
The most significant development concerning the use of data in public services has been in open data – the provision of public sector information on free licence for individual or commercial use. The UK is a leader in the field. As of October 2013, more than 10,000 datasets had been released and listed on the data.gov.uk portal – the most comprehensive resource of its type in the world. Thousands more are available through other data portals, such as data.london.gov.uk.
Whilst there is much to celebrate in the UK’s progress on open data, it has one glaring shortfall: the flow of information is in just one direction. Citizens and businesses can receive data from government, but there is no official, standardised and automated mechanism to provide data to government. Imagine, for a moment, a group of app developers who wish to help solve a problem in their local community, such as graffiti, bike theft or potholes. Currently, they have no way of knowing what data their local council needs, or what format it would need to be provided in. Furthermore, there is no mechanism for delivering the data in machine-readable format to connect with the council’s IT system. And even if the group did manage to build a system that worked for one council, if they wanted to offer the same solution to another, they would be likely to have to create a completely bespoke system to meet each local authority’s IT needs.
A separate development in the use of data in public services has been the digitisation of transactions. Since 2011, the Government Digital Service (GDS) has been working to replace paper-based (or poorly-designed electronic) transactions with high-quality online services delivered according to the digital-by-default standard. Whilst these efforts are certainly making transactions easier to use, they have not fundamentally changed the type or amount of data that is provided by citizens. A paper form is now just a digital form. The transaction itself has not altered in substance. Data from citizens is not being used to assist the work of government.
Asking the public what they think has often been considered a dangerous political strategy. Governments that make – or which are perceived to make – policy by public poll are accused of being populist and failing to lead. Using referenda is frequently dismissed as being either opportunistic (politicians typically only hold them when they feel confident of the outcome), or downright chaotic, as demonstrated when Californians have voted for measures which are effectively mutually exclusive (e.g. slashing taxes whilst massively increasing spending in infrastructure). Public consultations, meanwhile, have not infrequently been accused of being shelved or ignored if they do not concur with government’s desired conclusions. As a result, it is hard to find examples of citizens being meaningfully engaged in shaping public policy.
These measures fail to have much utility because they ask the wrong type of question. Polls, referenda and consultations request individuals’ views on a subject. They ask citizens to express the level of their support for a particular measure, or to state their preference from a list of pre-set options. Gathering such qualitative responses may be helpful in revealing the strength of public opinion on a specific issue. But the pressing need for good policymaking is having ideas and information. A more interesting and potentially fruitful approach would therefore be to ask citizens to provide facts or answers to specific questions; to provide knowledge that government alone could not find for itself.
This idea taps into a growing field of study concerning the wisdom of crowds. In essence, there are certain scenarios in which knowledge gathered from the many can exceed the accuracy or completeness of that provided by the expert few. Wikipedia is the best known example of knowledge aggregation: thousands of brains together creating a body of information more vast than any individual could achieve. A second approach is akin to the ‘ask the audience’ lifeline in the game show Who Wants to be a Millionaire, which provides the correct answer over 91% of the time. In the right conditions, the majority (or average) answer given by a group of people asked to independently and simultaneously respond to the same question has a high probability of being right. A third option is to create futures markets where contract payoffs are based on real-world events, such as political or election outcomes. One such example is the Iowa Electronic Markets (IEM). During five US presidential elections covered by IEM, the average absolute error in the market’s prediction of the major-party presidential vote share (across the five days prior to the election) was 1.20 percentage points. Opinion polls conducted at the same time had an average error of 1.62 percentage points.
In short, there are many ways in which the collective knowledge of individuals can be harnessed to improve the completeness or accuracy of information. Yet, with the exception of a short-lived Conservative Party initiative in 2010, which offered a £1million prize to the person or organisation that could develop a platform to harness the wisdom of crowds, there have been few attempts to harness citizen knowledge to assist government. This is surely a profound missed opportunity for open policymaking.
In summary, government has used technology to streamline transactions and better understand the public’s opinions. Yet it has failed to use it to radically change the way it works. Have public services been reinvented? Is government smaller and leaner? Have citizens, businesses and civic groups been offered the chance to take part in the work of government and improve their own communities? On all counts the answer is unequivocally, no. What is needed, therefore, is a means to enable citizens to provide data to government to inform policymaking and to improve – or even help deliver – public services. What is needed is a Government Data Marketplace.
A Government Data Marketplace (GDM) would be a website that brought together public sector bodies that needed data, with individuals, businesses and other organisations that could provide it. Imagine an open data portal in reverse: instead of government publishing its own datasets to be used by citizens and businesses, it would instead publish its data needs and invite citizens, businesses or community groups to provide that data (for free or in return for payment). Just as open data portals aim to provide datasets in standard, machine-readable formats, GDM would operate according to strict open standards, and provide a consistent and automated way to deliver data to government through APIs.
How would it work? Imagine a local council that wished to know where instances of graffiti occurred within its borough. The council would create an account on GDM and publish a new request, outlining the data it required (not dissimilar to someone posting a job on a site like Freelancer). Citizens, businesses and other organisations would be able to view that request on GDM and bid to offer the service. For example, an app-development company could offer to build an app that would enable citizens to photograph and locate instances of graffiti in the borough. The app would be able to upload the data to GDM. The council could connect its own IT system to GDM to pass the data to their own database.
Importantly, the app-development company would specify via GDM how much it would charge to provide the data. Other companies and organisations could offer competing bids for delivering the same – or an even better service – at different prices. Supportive local civic hacker groups could even offer to provide the data for free. Either way, the council would get the data it needed without having to collect it for itself, whilst also ensuring it paid the best price from a number of competing providers.
Since GDM would be a public marketplace, other local authorities would be able to see that a particular company had designed a graffiti-reporting solution for one council, and could ask for the same data to be collected in their own boroughs. This would be quick and easy for the developer, as instead of having to create a bespoke solution to work with each council’s IT system, they could connect to all of them using one common interface via GDM. That would good for the company, as they could sell to a much larger market (the same solution would work for one council or all), and good for the councils, as they would benefit from cheaper prices generated from economies of scale. And since GDM would use open standards, if a council was unhappy with the data provided by one supplier, it could simply look to another company to provide the same information.
What would be the advantages of such a system? Firstly, innovation. GDM would free government from having to worry about what software it needed, and instead allow it to focus on the data it required to provide a service. To be clear: councils themselves do not need a graffiti app – they need data on where graffiti is. By focusing attention on its data needs, the public sector could let the market innovate to find the best solutions for providing it. That might be via an app, perhaps via a website, social media, or Internet of Things sensors, or maybe even using a completely new service that collected information in a radically different way. It will not matter – the right information would be provided in a common format via GDM.
Secondly, the potential cost savings of this approach would be many and considerable. At the very least, by creating a marketplace, the public sector would be able to source data at a competitive price. If several public sector bodies needed the same service via GDM, companies providing that data would be able to offer much cheaper prices for all, as instead of having to deal with hundreds of different organisations (and different interfaces) they could create one solution that worked for all of them. As prices became cheaper for standard solutions, this would in turn encourage more public sector bodies to converge on common ways of working, driving down costs still further. Yet these savings would be dwarfed by those possible if GDM could be used to source data that public sectors bodies currently have to manually collect themselves. Imagine if instead of having teams of inspectors to locate instances X, Y or Z, it could instead source the same data from citizens via GDM?
There would no limit to the potential applications to which GDM could be put by central and local government and other public sector bodies: for graffiti, traffic levels, environmental issues, education or welfare. It could be used to crowdsource facts, figures, images, map coordinates, text – anything that can be collected as data. Government could request information on areas on which it previously had none, helping them to assign their finite resources and money in a much more targeted way. New York City’s Mayor’s Office of Data Analytics has demonstrated that up to 500% increases in the efficiency of providing some public services can be achieved, if only the right data is available.
For the private sector, GDM would stimulate the growth of innovative new companies offering community data, and make it easier for them to sell data solutions across the whole of the public sector. They could pioneer in new data methods, and potentially even take over the provision of entire services which the public sector currently has to provide itself. For citizens, it would offer a means to genuinely get involved in solving issues that matter to their local communities, either by using apps made by businesses, or working to provide the data themselves.
And what about the benefits for policymaking? It is important to acknowledge that the idea of harnessing the wisdom of crowds for policymaking is currently experimental. In the case of Policy Futures Markets, some applications have also been considered to be highly controversial. So which methods would be most effective? What would they look like? In what policy domains would they provide most value? The simple fact is that we do not know. What is certain, however, is that innovation in open policymaking and crowdsourcing ideas will never be achieved until a platform is available that allows such ideas to be tried and tested. GDM could be that platform.
Public sector bodies could experiment with asking citizens for information or answers to particular, fact-based questions, or even for predictions on future outcomes, to help inform their policymaking activities. The market could then innovate to develop solutions to source that data from citizens, using the many different models for harnessing the wisdom of crowds. The effectiveness of those initiatives could then be judged, and the techniques honed. In the worst case scenario that it did not work, money would not have been wasted on building the wrong platform – GDM would continue to have value in providing data for public service needs as described above.
For anyone who believes that effective public service delivery and good policymaking depend on strong, empirical evidence, ensuring that government has access to the best data available should be a priority. Technology offers a means to source more information than ever before. Yet, to date, government has failed to harness its benefits. That is a missed opportunity.
Establishing a Government Data Marketplace offers an answer. By focusing on its data needs, and letting the market innovate in new ways to provide it, government could – for the first time – genuinely work in partnership with citizens, businesses and community groups to improve society. This would be a radical change – and offer a very tangible means for reducing the size of government. By creating a competitive marketplace for data, it could also help reduce the cost of sourcing information, and encourage the public sector to converge on commons solutions, rather than acting as hundreds of siloed organisations.
As for policymaking, GDM would provide a platform to experiment with initiatives to involve citizens in a far more meaningful way than is currently possible. No study has been conducted on a sufficient scale to find the optimum solution for harnessing the wisdom of crowds in the UK. GDM would enable the public sector to ask the questions it most needed addressing, and see if citizens could provide the answers. Successful initiatives could be improved and scaled, benefiting numerous areas of policymaking.
Technology has already brought about a data revolution that has benefitted individuals, businesses and other organisations. It is time that government took full advantage, too.
Follow Eddie Copeland on Twitter @EddieACopeland