How the UK Government Saved £4 million in 15 minutes with Open Data
This story tells how a senior UK government official used open data about government spending to save £4 million pounds in 15 minutes.
It shows that one of the biggest beneficiaries of government open data programs can be government itself. This may seem surprising as government already has access to its own data. However, in reality, there are many obstacles to effective information sharing within government - so that even if data exists, friction may prevent it “flowing” to where it is most needed and most valuable.
We emphasize this point that government itself is a major beneficiary of open data because it is distinct from - and complements - the more frequently cited benefits around innovation and transparency driven by third-party use in businesses and civil-society. As such, it provides a powerful additional reason for government to pursue an open data agenda.1
In 2010, the UK began publishing government spending data as open data At first in the form of the central government budgetary database (COINS) and then with the commitment to release detailed transactional spending data on an ongoing basis which was announced at Open Government Data Camp in November 2010.
The commitment to publish departmental spending at the transactional level on a monthly basis - that is individual purchases from suppliers, was a major step beyond what any other government had done before. It positioned the UK as a global leader in fiscal transparency. At the same time, when the decision was made the government frankly accepted that they did not know the exact benefits that would arise.
Fast forward three and a half years. Liam Maxwell, the UK Government CTO is talking with a CIO-level colleague from another government department. In conversation they mention that they have recently purchased an expensive market research report for senior IT decision makers. These reports, whilst valuable, are very expensive. Liam is a man intent on getting maximum value for money in government IT - especially in a post Great Recession world. These market reports are expensive, are other government departments buying these same reports he wonders? Perhaps it is even possible that different departments are even buying the same report and thereby spending money unnecessarily?
It is worth taking a moment here to reflect on what would have happened in the “old”, pre open spending data days. Liam would have had a simple choice. On the one hand, he could simply have dismissed this question from his mind and moved on to the many other things he had to do. End of story.
On the other hand, he could have decided to pursue it. After all, the Government’s Treasury department has had a central budgeting and reporting system for many years. However, this system, like most government financial management platforms, does not track information at a transactional level. Instead, departments will report spending against particular finance codes and projects. At best, Liam would be able to look something like total spending on marketing if there was a specific code for that. So that would be a dead end, even if he were able to access the central budgeting system - which is well known to be complicated and requires training to manage and use.
His next, and only remaining, option would be to write to CIOs or other relevant officials across government and ask them to tell him what market research reports they purchased and at what cost. How would this go? Liam would need to write the request - or task a staff member to do so. This would then get sent out. Likely weeks or months would go by as this request worked its way to the top of the pile in other departments. Some departments might lose the request or never get round to responding - some might not even want to respond. Ultimately, weeks or months later Liam would be lucky to have a partial picture of how things stood.
Reflecting on this option, it’s likely that Liam would quickly decide to save himself the time and energy and go with the first option: forget it and move on - perhaps filing it mentally in the drawer of “things to look at again later when I am less busy”.
So what happened differently?
To return to our story, in 2010 the government had started to publish transactional data openly. Thus, when Liam Maxwell went looking it was already all online and included spending on expensive market research reports. There was still a problem that each government department published their data in separate monthly CSV files. This meant you would have to search across hundreds of different spreadsheets to find the data. However, because the data was open, another organization outside of government had taken the role of “infomediary” and had aggregated all those files into one conveniently searchable whole on the OpenSpending website.
It was therefore a matter of minutes for Liam Maxell to do a quick search across all government spending and find out which departments were buying which reports and to see whether there was duplication. And guess what: there was. Total estimated savings from eliminating that duplication: just over 4 million pounds.
This story shows how a senior government official could use open data to turn a question into insight in minutes - and save the government money in doing so. But it is not just a powerful story, it also illustrates several key points about the potential and use of open data:
- Data must be connected to a concrete problem or opportunity to generate value. This is obvious, but it is worth reiterating. This story provides a specific, powerful example of how this can work in practice.
- Uses are unexpected: what open data can be used for and where it will generate value is often difficult to foresee. When the UK government committed in 2010 to release detailed transactional spending information it almost certainly did not anticipate the use reported here.
- The use was inside government: that is, it was a government official who used the data and took action. This illustrates a key point that one of the biggest potential beneficiaries of government open data is government itself. This may seem paradoxical since government already has the data but it really isn’t: governments are large and complex and taking data from where it resides to where it can generate insight is hard. Open data can make a huge difference to creating a frictionless data ecosystem inside government.
- Infomediaries matter: the raw open data released by government would not have been useful since each department published data in dozens of different files. Thanks to a third-party platform - OpenSpending - this data had been aggregated and made easily searchable.
- The “many minds principle” for government: the many minds principle adapted for government reads: “the best thing to do with your data will be thought of by another department”. By opening up data we reduce friction in data sharing, accelerating “time to insight” and “time to action”. A government minister can go from a question to an answer in just a few minutes generating a basis for actionable, concrete change that saved government money.
Personal data should generally not be published as open data unless there are strong public interest arguments for making this information available as part of the public record (e.g. in the case of information about company directors in company registries, or information in parliamentary records). Much government data is personal in nature, for example, individual tax records or social security records. This personal data would not be shared as open data. ↩