Who should I vote for? There’s (not) an app for that.

Why voter advice apps like VoteMatch and Vote for Policies mislead voters.

Ever been on one of those sites? The ones that tell you who to vote for? They’re called voter advice apps. They have helpful names like Votematch and Vote for Policies and are sold as people power cutting through politicians’ bullshit.

The problem is: they aren’t actually honest about how politics works. And that’s a problem if you sell your app as a tool for disengaged or first-time voters.

Here’s what I get when I try the 2010 VoteMatch app (the 2015 one isn’t out yet): “You agree 66% with the Labour party, 64% with the Green party, 49% with Plaid Cymru and 22% with the BNP.”

Now, I’m not a floating voter, and seven years on a council with Greens has inoculated me against their brand of uncosted left-lite preachy chicanery for life. So maybe you could dismiss me as just being irritated that I score high agreement with the Greens. But Plaid Cymru? And 22% agreement with the BNP? What’s that about? Shurely some mistake.

Usually parties supply a set number of statements on a list of issues put together by a well-meaning campaign group who think that neutral info stripped of politics will encourage the disengaged to engage.

It’s an interesting theory of change, as Sir Humphrey might say, if he worked for a modern NGO.


  1. What the parties don’t say is as important as what they do say

Since 1992’s tax bombshell campaign, there has been a huge electoral downside to revealing too much before the election. So parties only reveal part of their plans — rightly or wrongly. Perhaps the Conservatives’ £12bn unspecified welfare cuts might interest VoteMatchers, but since it’s an absence of policy rather than a policy, and since the Conservatives won’t submit it for inclusion, no-one gets to use it to decide who to vote for. And lists of policies can’t cover issues that haven’t arisen yet: for obvious reasons the 2001 manifestos are silent on whether or not to invade Iraq, the defining issue of the 2001–05 parliament.

2. The weighting that parties give issues matters — but that’s completely unclear in the apps

Presenting lists of questions — now the housing question, now the NHS question, now immigration — makes it appear that all parties concentrate equally on all issues. But that isn’t the truth at all: take housing, which is likely to be a few sentences in the Conservative manifesto but several paragraphs in the Labour one. The VoteMatcher who cares desperately about housing won’t see that it’s an afterthought for the Tories but central for Labour — they’ll just see two policies of similar length, in black and white.

And this is where the irrelevance of telling me my agreement with Plaid Cymru and the BNP comes in. As I don’t agree with the key issues for those parties — independence for Wales and racism, respectively — my agreement with them the NHS or welfare or education is completely irrelevant. The voter apps assume all policies are equal, and proceed blithely from there — with no thought as to the priorities and narrative of the parties in question.

And more than that: presented blandly, without context, a voter gets no sense of the place of a policy within the party’s offer. Did it represent a hard-fought compromise between chancellor and prime minister? Is it a u-turn on a previous position? Is it counter-intuitive and designed to make a point — like protecting NHS spending for Cameron in 2010? All these might influence how a VoteMatcher might view a policy — unless they don’t know.

3. Politics is about trade-offs — but voter advice apps pretend it’s not

Politics is about the art of the possible. Any aspiring government needs to make choices — most importantly about spending priorities. What you have to give up in order to get a policy is rarely clear in voter apps — and there appears to be no penalties for promising the moon on a stick. “We will do X” is an entirely different policy from “we will do X paid for by Y” because politics involves trade-offs, and pretending it doesn’t, and everything is possible and affordable, misleads voters.

4. It’s not just policies that matter

Voting on policies alone, divorced from who proposes them, relates to a world that doesn’t exist, where no party has a brand, record, core constituency or people who represent it.

To give an example: the Conservatives’ record on lesbian and gay rights before 2010 was pretty poor. LGBT people may view same-sex marriage as evidence that the Conservatives have changed, or the Conservative leader’s inability to deliver a majority in his party for it as evidence that they haven’t. But still, a policy promise from the Tories in 2015 on LGBT rights is meaningless without the context and the history of the party in question. How this then influences a voter’s decision is up to them.

Policies should be set within the context of party brand, values and history — otherwise it’s hard to recognise why a cut in inheritance tax is “same old Tories, only for the rich” or a promise to repeal the bedroom tax is “same old Labour, the welfare party”. If voter apps seek to inform, why should voters be denied that information about party brand, so crucial to understanding modern politics? Not least as both parties spend considerable time trying to confound their brands — whilst being unable to escape them, as they contain a core truth.

(Slide from this fascinating poll by ComRes, showing that party brands do mean something.)

5. It’s really not just policies that matter

Government isn’t just about policies: it’s about the ability to implement a programme and cope with the inevitable crises. And that’s why caring about who leads and seeks to represent that party isn’t silly or superficial. Nor is finding out about the views and background of candidates: commentators argue that in 2010 the pre-election beliefs of Conservative PPCs were more useful in predicting Conservative policy in government — small-state, socially liberal, eurosceptic, not wedded to the NHS — than the 2010 manifesto.

6. Policies should be tested in battle

Policies aren’t really until they have been fought over — on the Today programme, at a party conference, on twitter. When I read a policy, I want to know why the other parties think it’s nonsense, where the holes are, and what the party proposing it is avoiding saying. I don’t get that with bald prose on a webpage. Policies without rebuttal is like a premier league match with no fans, no pundits and no commentary.

Conclusion — so who should I vote for, then?

Voter apps are written by people who don’t understand politics. They presume that voting choice can be reduced to a list of policies, disconnected from one another — that an election is just a million shopping lists, aggregated. For the voter app, politics has no ideology, no values, no brand, no compromise, no arguments — and no participants.

That’s a pretty big way to mislead those disengaged voters whilst telling them you’re empowering them.

The world according to the voter apps doesn’t exist. There is no perfect, scientific method to pick the party or candidate you should vote for. Policies don’t exist in a vacuum, and not all “facts” can be “checked”.

So instead, try this: talk to people you know. Seek out people that you don’t think you’ll agree with. Ask them how they’ll vote, and why. Listen to the leaders’ debates. Read the newspapers. Argue with people. Form your own opinion. Take anything into account that matters to you. Ignore stuff you don’t care about.

There isn’t a right answer to the question. Who you should vote for is your call.

(Of course, if you’re my friend and you ask me, I’ll tell you to vote Labour.)