United Kingdom's TheyWorkForYou
Taking the Long View
by Becky Hogge
Brief View 5 MINS
Full Story 20 MINS
is a parliamentary monitoring website that has been running for more than 10 years in the UK. It provides an accessible, searchable version of the official record of proceedings of the UK Parliament, as well as the devolved Northern Ireland Assembly. It also provides analysis of the voting records of elected and non-elected representatives. TheyWorkForYou predates open data reforms in the UK by a number of years, and is included in this report as a window onto the long-term impacts of outputs based on public data. One of TheyWorkForYou’s long-term impacts appears to be encouraging parliamentarians to vote less with their party and more in the interests of the people they represent. It is also important to consider the potential monetised time savings delivered to the site’s users—many of whom are civil society groups and journalists. These might well be in the millions of pounds a year, although methodological constraints mean it will probably always be impossible to put an exact figure on them.Read More
- The greatest impact of TheyWorkForYou may be on Parliamentarians themselves. mySociety suspects, and some data confirms, that Parliamentarians have changed the way they go about their work in response to TheyWorkForYou’s vote monitoring and analysis tools, both by turning up for more votes and rebelling against their party more often.
- Just as TfL’s open data policy saves commuters’ time, saves time for its users—many of whom are civil society groups and journalists. As well as expecting websites with a social/political mission to achieve long-term positive social/political impacts by themselves, we can also expect them to save time for people trying to achieve those impacts through other means.
- Although it may look on the face of things like postcodes have nothing to do with Parliamentary monitoring, in fact postcode data is one of the key datasets that drive . Early on in the website’s history, mySociety had access to postcode data via a licence from Ordnance Survey. It now uses open data on postcodes provided by Ordnance Survey and the Office for National Statistics. But mySociety’s lack of access to Royal Mail’s Postcode Address File (see page 12) means it cannot deliver accurate information to all its users on who their elected representative is.
- mySociety views maintaining as a “residual”, and hence low priority, activity, and now focus most of their work on helping international NGOs. The long-term sustainability of projects like is brought into question by this case.
Proportion of users surveyed who say they are using it as part of their job: 30%
Proportion of users surveyed who say they are getting information on their elected representative for the first time: 21%
“I don’t think the political classes have got used to the idea of news or information that sticks.”
“It is a hundred times easier to search than Hansard itself. I’ve spent two hours on Hansard before coming here and have just found what I want in under a minute.”
provides accessible, searchable data on the members and proceedings of Parliament, the main legislative body of the United Kingdom, as well as the Northern Ireland Assembly. Previous versions of the website also offered information on the members and proceedings of the Scottish Parliament. provides a wide range of information, including members’ voting records, speeches, and registered interests. The front page of the website invites users to answer the question “Does Your MP represent you?” by filling in their postcode to access analysis on the way their constituency’s MP votes and see their latest appearances in Parliament (See Figure 9 and Figure 10).
was launched in June 2004 by a group of volunteers “who thought it should be really easy for people to keep tabs on their elected MPs, and their unelected Peers, and comment on what goes on in Parliament”. Individually, the original volunteers had already developed a number of civic-minded websites. Since 2006, has been run by mySociety, a nonprofit social enterprise based in the UK that develops web platforms that “give people the power to get things changed”.
The data that powered was scraped from Hansard, the official record of parliamentary proceedings, published on Parliament’s own website. launched despite the fact that this activity constituted a copyright infringement: The volunteers did not have the right to reproduce Hansard, which was covered by Crown Copyright. Later on, and in cooperation with some of the volunteers, click-use licences were developed at the Office for Public Sector Information (OPSI) which among other things legitimised the site’s activities.Read More
- Postcode and constituency boundary data, for mapping people to their
- People, for each of the legislatures covered
- Transcript data, for each of the legislatures covered, and
multiple different types of data in that (e.g., actual debates, written
answers, future business, committees, etc.)
The vote analyses are important, but are fundamentally just derived
from the transcript data that includes the voting information. I’d possibly add our own dataset of everyone signed up for email alerts, as the daily emails we send based on information gleaned from the other datasets does presumably help power the site in a way.
Over 10-plus years operating , Somerville has seen changes in how this data is provided. In terms of postcodes and constituency boundaries, since 2010, has been using Ordnance Survey’s CodePoint Open open data product to match people’s postcodes to their constituencies in Great Britain, and the Office for National Statistics’ postcode product to perform the same function for users living in Northern Ireland. The latter source does not conform to the open definition, since it is provided for non-commercial use only. Before these products became available, mySociety had access to postcodes via a licence from Ordnance Survey.
The people data has been constructed over the years from various data sources, and mySociety makes it available under an open licence.
The transcript data for the Northern Ireland Assembly is accessed via the assembly’s AIMS open data portal launched by the Northern Ireland Assembly in 2012. Since 2014 the development team of the official UK Parliament website, , has also been making its data available via API. Despite this, still scrapes transcript data from Hansard every morning to populate the site with the latest proceedings of the UK Parliament, rather than access the data via the API.
Tom Steinberg, outgoing CEO of mySociety, explains that continuing to scrape official websites rather than taking advantage of new open data APIs is mainly down to prioritising resources at mySociety:
If we’ve not used some of it [open data] it’s quite often because the cost of us rewriting the software to use it is just not worth it when we can carry on screen scraping.
The original scraper that provided with transcripts of proceedings in the Scottish Parliament was developed by a volunteer, Mark Longair. But for the last couple of years ’s Scottish site has not been updated. Somerville explains:
[The] Scottish Parliament changed their site a couple of years back and we haven’t had time since to update the scraper, and not had anyone volunteer to help out, so [it’s not been] updated since then. Don’t think they have an API of any sort as yet.Read More
played a key role in forming the UK’s policy around open data. In 2007 Tom Steinberg was commissioned to co-write an influential review of the opportunities presented by opening up public sector information, a move he says traces directly back to his involvement with :
Back in the Labour era, Number Ten [i.e., the Prime Minister’s office] was willing to talk to us as a group of people because TheyWorkForYou existed. Why were they willing to talk to us? Because they used it, they used it to look up their own parliamentary data and they couldn’t use the Parliament website at the time because it was too [poorly designed]. And so unarguably TheyWorkForYou was a key to get in through the door of Number Ten. And it led pretty directly to the Power of Information review.
One of the most interesting things about is how central postcode data is to its operations. mySociety was part of the range of voices calling on Ordnance Survey to free its data before the 2010 CodePoint Open release, particularly boundary data, which is crucial in identifying a user’s elected representative.
The way mySociety uses postcode data immediately personalises the experience of using TheyWorkForYou, and lowers the barrier to accessing the workings of democracy in a country where research shows only 22% of people can name their elected representative. This observation underscores the utility of postcodes across a vast range of online applications.Read More
The most recent in-depth research into ’s user base was published by Tobias Escher of the Oxford Internet Institute in 2011. It showed that the site receives between 200,000 and 300,000 visits every month (see Figure 11—the peaks in May 2009 and May 2010 are respectively due to the MPs’ expenses scandal, when many newspapers published links to MPs’ pages on , and the publication of an election quiz that matched voters to Parliamentary candidates based on each of their stated political beliefs, developed by volunteers in the run-up to the 2010 elections).
Escher reports that about half of ’s audience are regular users, and about one in five use it every month. 30% of respondents to a survey of 903 site users developed by Escher say they use the website for work, and usage patterns concentrated in UK business hours substantiate this. About 2% of visits come from IP addresses that indicate the user is working in the Houses of Parliament, and a further 2-3% indicate the user is working for the government. Workers at the BBC accounted for 0.5% of visits in 2010.Read More
Tom Steinberg thinks that the site’s impact on elected members is probably greater than on any other of its stakeholders:
I don’t think the political classes have got used to the idea of news or information that sticks. They’re very habituated to the idea that there’s a bad news story today and it’s gone tomorrow. All politicians can cope with that. However, a Wikipedia page or a TheyWorkForYou page … they’re sort of permanent. They mean that if you’re going to go and meet someone who doesn’t know you, in the future, then you have every reason to suspect that that person will find out about you from this. So what these static pages say I believe influences what politicians do.
Several anecdotes support the claim that Parliamentarians are becoming increasingly aware of how their activities are reflected on . In 2006 The Times newspaper speculated that some MPs were making interventions in House of Commons debates simply to keep their “appearances” tally on sufficiently high, an event which prompted mySociety to rethink how it represented MPs’ activity. Steinberg relates a story he heard from someone who once had a meeting with an MP cut short so that the MP could go and vote in the House of Commons. When they asked whether the vote really mattered, the MP replied “They all matter since TheyWorkForYou”. In 2008, Gordon Prentice MP complained to the leader of the House of Commons that did not cover MPs’ work on committees and therefore gave a “distorted impression” of the efforts MPs make at Westminster.
Philip Cowley, Professor of Parliamentary Government at the University of Nottingham, has published data showing that “rebellion” (that is, MPs voting against the party line, or—specific to Cowley’s data—government MPs voting against government policy) is on the up in British politics. An Economist article from 2012 cites this data and argues that vote-tracking websites are part of the reason why, as they allow voters and constituency parties to keep better track of their MPs:
One ringleader of ’s huge Conservative rebellion over a referendum on EU membership says that it was “incredibly easy” to persuade MPs to join because they were already under intense pressure from their constituency associations. Gone are the days when an MP could vote with the government, then sign a contradictory early-day motion or two to muddy the waters, he adds. These days it is “harder to bluff”.
In recognition of the site’s important accountability role, most of the new resources mySociety now devotes to go into improving voting records and voting analysis. A recent grant from the Joseph Rowntree Reform Trust saw mySociety add about 60,000 more pages of vote analysis to the site.
The impact on the site’s other users, be they civil servants, journalists, campaigners, or simply citizens interested in the workings of Parliament, is less visible. But Steinberg does not doubt that it is real:
I have no doubt that in some way [TheyWorkForYou] must oil the wheels of tons and tons of campaign groups and people who are kind of just generally in the political world … if it saves them five minutes here, that’s five minutes they can spend on something else …. However, what I can’t do is say “Here is someone who said ‘I was going to lose my campaign, and then I used TheyWorkForYou and then I won it’”.
For these communities, ’s impact is probably best understood through the same lens as the impact apps derived from TfL’s data have on London’s commuters: It saves them time. One respondent to Escher’s survey notes:
It is a hundred times easier to search than Hansard itself. I’ve spent two hours on Hansard before coming here and have just found what I want in under a minute.
Crudely, then, one might try and follow Deloitte’s analysis of TfL and assume that 30% of the 200,000-300,000 monthly visits to are work-related. If every one of these visits represented a 119 minute time-saving (and since time wasted sitting in traffic can be compared to time wasted looking for something and not finding it) one could use the Department for Transport’s average value of working time (£34.12 per hour) to conclude that mySociety was saving this group of users roughly £5,075,350 worth of time every month, or just under £70 million worth of time each year. Taking a more realistic assumption, say that each visit represented Steinberg’s 5 minute time saving, annual time savings delivered are just over £2.5million.
The analysis is flawed, of course. Unlike the Deloitte analysis, which takes advantage of TfL’s data on Lost Passenger Hours, we cannot assess how many hours working users of might have spent trawling through Hansard and other information sources, not finding what they were looking for. It is included here simply to remind readers that although they may not make great stories, incremental time savings are an important and entirely valid form of impact to consider.
contributes to my ‘current awareness’ of what is being said about diabetes and insulin (access to treatment, statistics, etc.) and other related health topics (for example organ transplantation and stem cell research as that’s very relevant for diabetes and its complications too). The email alerts and RSS feeds mean the information lands rather helpfully in my intray …. It’s a great site—thank you.
mySociety’s team also receive private messages of thanks from policy experts, elected and unelected members, third sector workers, and others working in public affairs.
While there are clearly TheyWorkForYou users who are already politically active and who are also [organised] in groups, the online survey shows that one in five users (21%) has neither been politically active (online and/or offline) within the last year nor been a member of any group and importantly, has got information for the first time on what his or her representatives are doing through the use of TheyWorkForYou.
has also played a key role in helping other Parliamentary Monitoring Organisations (PMOs) around the world. Three years ago it changed its mission, and it now works internationally to support partners who deploy its technology in countries around the world. The code that runs is open source, and has been adapted to create similar websites in New Zealand, Ireland, and Australia. In 2013, working with Kenyan PMO Mzalendo, mySociety released Pombola, a “less-UK centric” codebase that it actively helps PMOs to install in order to deploy parliamentary monitoring websites in their own countries. Pombola is already in use in Ghana, South Africa, Nigeria, and Zimbabwe.Read More
predates open data reforms in the UK by several years and can in fact be said to have contributed to open data policy’s inception. This case study is included in the report because (although not all its data inputs are strictly open data) it is an example of an “output” project that repurposes public information and that has existed for over a decade. As such, it may provide a window through which to view the more long-term impacts of such public data re-use outputs.
Indeed, the increase in political rebellion, attributed by experts in part to the increased visibility of MPs’ voting records that provides, is a positive indicator that is having a long-term impact. If politicians are choosing to serve the people who elect them more, and the party political machinery less, then this is a good outcome for representative democracy.
Of interest too is the amount of time may be saving civil society groups and journalists (and, of course, corporate lobbyists too). The monetised time savings used in this case study should not be relied upon and do not bear repeating, since unlike in the TfL case they are based on speculation about the amount of time such users may have lost in seeking out information elsewhere, and not recorded Lost Passenger Hours. The calculations have rather been included here in order to remind readers that as well as expecting websites with a social/political mission to achieve long-term positive social/political impacts by themselves (in this case better representative democracy) we can also expect them to save time for people trying to achieve those impacts through other means.
Although Matthew Somerville sees no reason why won’t survive another 10 years (“Parliament will certainly still be around”), it should be of some concern that Tom Steinberg describes the site as a “residual” from before mySociety changed its mission to focus on helping international NGOs develop similar websites through projects like Pombola. Somerville sees lots more work for mySociety to do with , and says that even if the official website becomes easier to use, many of ’s most important functions—like vote analysis and email alerts—may never be provided by Parliament:
There are many things the site still doesn’t cover, but it’s not like we have had countless volunteers [clamouring] to add select committees, or what have you.
This observation further calls into question how realistic it is to anticipate sustained activity from civic hackers (or indeed David Cameron’s “whole army of effective armchair auditors”) in response to the release of government data as open data. In fact, that pool of volunteers may be more limited than first imagined. mySociety has done recent, funded, work on TheyWorkForYou, the 60,000 additional pages of vote analysis supported by the Joseph Rowntree Reform Trust. But the fact that the site does not serve its Scottish users as at July 2015 because no one has volunteered to update the tool so that it can scrape the Scottish Parliament’s new website should be of major concern.Read More
For open data advocates
The Deloitte study on the impact of TfL data on transport users, and its analysis that TfL are delivering transport users annual monetised time savings of between £15m and £58m, is one of the most quoted stories of open data impact in policy circles. The Deloitte methodology can be borrowed to discuss impact in other spheres too.
For open data developers
Shiny new projects are fun and inspiring, but established sites like need volunteer developers too. Open data developers should consider whether knowing you’re contributing work to a project that is having long-term positive impacts for representative democracy makes up for some of the hassle of working on someone else’s code.
For fundersRead More
United Kingdom's Open Public Services Network
United Kingdom's OpenCorporates
Battling Ebola in Sierra Leone
New Zealand's Christchurch Earthquake Clusters
Singapore's Dengue Cluster Map
United States' Eightmaps
Kennedy vs. the City of Zanesville, United States
United Kingdom's Coeliac UK and Brandbank
United Kingdom's Ordnance Survey OpenData
United States' New York City Business Atlas
United States' NOAA: Opening Up Global Weather Data in Collaboration with Businesses
United States Opening GPS Data for Civilian Use
United Kingdom's Transport for London
United Kingdom's HM Land Registry Price Paid Data
1 Although in the case of postcode information (for Northern Ireland) provided by the Office for National Statistics, this data does not conform to the open definition, as it is made available for non-commercial use only.
2 . (n.d.). About TheyWorkForYou. Retrieved September 30, 2015, from :
3 mySociety. (n.d.). Homepage. Retrieved September 30, 2015, from mySociety:
4 Hogge, B. (2011, May). Open Data Study: New Technologies. Retrieved from Transparency and Accountability Initiative:
5 Today, the official Parliament website is published under an Open Parliamentary Licence, which encourages re-use. In 2014, the development team behind the official Parliament website began releasing Parliamentary proceedings as open data.
6 Email-based interview, Matthew Somerville, Lead Developer, TheyWorkForYou, mySociety.
7 Email-based interview, Matthew Somerville, Lead Developer, TheyWorkForYou, mySociety.
8 Nixon, M. (2014, February 20). Open Data day resources. Retrieved from mySociety:
9 Hadi, Z. (2014, May 2). Welcome to data.parliament – The Blog . Retrieved from :
10 Tom Steinberg founded mySociety in 2003 and acted as its CEO until August 2015. He was interviewed for this report in June 2015.
11 Interview, Tom Steinberg, Director of mySociety 2004-2015.
12 Email-based interview, Matthew Somerville, Lead Developer, TheyWorkForYou, mySociety.
13 BBC News. (2013, May 16). Three quarters of people ‘cannot name their local MP’. Retrieved from BBC News:
14 Most postcodes in the UK match precisely to constituency boundaries, meaning the OS CodePoint Open product is generally sufficient to serve ’s users. But there are exceptions. A 2013 Twitter conversation between Matthew Somerville and a user reveals one occasion where the data did not match: Although inhabitants of the user’s road all shared a postcode, those living in odd-numbered houses were represented by one MP, and those living in even-numbered houses by another (mySociety 2013) ( Access to the Royal Mail’s Postcode Address File (see page 12) would allow to provide all its users with accurate information about who represents them in the UK Parliament.
15 Escher, T. (2011, May). Analysis of users and usage for UK Citizens’ Online Deomcracy. Retrieved from
16 Interview, Tom Steinberg, Director of mySociety 2004-2015.
17 Hurst, G. (2006, February 27). The MPs who can’t stop talking. Retrieved from Times Online:
18 Prentice, G. (2008, October 23). HC Deb, 23 October 2008, c151WH. Retrieved from :
19 The Economist. (2012, October 27). Unleashed: Why backbench MPs are becoming so much harder to control. Retrieved from The Economist:
20 . (n.d.). Frequently Asked Questions. Retrieved September 30, 2015, from :
21 Private communication, Myf Nixon, mySociety marketing and communications manager
22 Escher, T. (2011, May). Analysis of users and usage for UK Citizens’ Online Deomcracy. Retrieved from
23 Email-based interview, Matthew Somerville, Lead Developer, TheyWorkForYou, mySociety.
24 Email-based interview, Matthew Somerville, Lead Developer, TheyWorkForYou, mySociety.
25 Prime Minister’s Office. (2010, May 29). PM’s podcast on transparency. Retrieved from :