YouGov's final general election campaign poll, published on election day in the Daily Telegraph, predicted each party's share of the vote to within one point:
YouGov's final campaign poll for the Sunday Times (published on 1 May) was exactly right, and its final survey for Sky News (published on 4 May) had an average error of one point. Other polling companies also produced election-day figures close to the result; the difference is that YouGov, alone, consistently published figures throughout the campaign that were close to the result.
Looking back, it is clear - not only from YouGov's surveys but from the average of off-line polls - that there was little net change in the support for any party, beyond a decline of around two points in Conservative support and an equivalent increase in Liberal Democrat support in the first half of April.
This stability allows polls to be judged not only against their final election-day prediction, but against their results throughout the campaign.
Altogether, YouGov published 15 campaign polls. Fourteen of these reported Labour's share within one point of its result on election day; ten were "right" to within one point for the Conservatives, and ten within one point for the Liberal Democrats. Every figure in every YouGov poll conducted in the final three weeks of the campaign, following the initial net swing from Conservative to Lib Dem, was "correct" to within two points. This record for consistency exceeds that achieved by any polling company in any British general election.
YouGov's performance comes less than two weeks after its shares were listed on AIM, the London Stock Exchange's Alternative Investment Market. Ahead of the float, and the election, much attention was paid to the accuracy of the polls. As the Sunday Times recorded (May 8): "Throughout the election campaign a battle raged over whether YouGov, using online polling, would be more accurate than those mainly using the telephone."
This controversy was fuelled by two elements: the arrival of YouGov as a high-profile online polling company, and the tendency of most conventional polls, at each of the last three elections, to overstate Labour's support and understate that of the Conservatives. This time, the final predictions of the telephone poll companies were close to the election result. However, their figures during the five weeks leading up to polling day were far less consistent. They ranged from a Conservative lead of five points to a Labour lead of 14 points.
This is not the first general election that YouGov has polled, nor the first we have got right. But never before have we been so much in the media spotlight. We knew this was a test we had to pass. We believe that we did so, not only in terms of our final election-day prediction, but throughout the campaign. Clients of opinion surveys and market research rightly demand not just occasional accuracy but consistent accuracy. Over the past five weeks we have shown that we can deliver this, fast and cost-effectively.
YouGov's performance has begun to attract international attention. Two days before polling day, Angus Reid, the doyen of Canadian pollsters, told Canada's leading political website: "Canadian pollsters will eventually have to switch from telephone polling to online polling, as practised by the UK company YouGov".