Twitter: the real campaign was elsewhere

What impact did Twitter have on the election in Wales? Overall, probably very little: less than a quarter of the candidates were registered on Twitter, and many of these were either casual users or failed to Tweet at all. However, for individuals the picture is more encouraging: active Tweeters performed significantly better than non-users or casual users. And when we look at each party's overall performance, active Twitter users tended to outperform their party colleagues (although there are individual exceptions, of course).

It would be tempting to look at the crude voting data and decide that Twitter had delivered the vote. In Wales Twitter users were well represented in the results: of the 10 candidates who showed the largest increase in the share of their vote, 5 were Twitter users, more than double what we would expect to see. 13 successfully elected MPs were on Twitter, while Lembit Opik lost his Montgomery seat, having not used his Twitter account at all during the campaign. This analysis would be misleading, though. In the case of Opik, he lost to a well organized Conservative campaign that did not use Twitter either. Eight of the newly elected MPs on Twitter actually performed worse than average for Wales. Furthermore, of the ten candidates who lost the largest share of votes in Wales, 5 were Twitter users.

What we can see is that more engaged Twitter users were significantly better represented in the voting figures than casual users: 14 active Tweeters performed better than average, compared to 8 who performed worse in terms of increasing their vote. However, this overall analysis is skewed by the much wider picture of what was happening to the parties on a national and regional scale. Instead, it may be more helpful to examine the possible impact of Twitter within each party, thereby enabling us to allow for any 'bounce' caused by events such as the Leaders' Debates and the progress of the campaign as a whole.

As we noted in an earlier blog, the Conservatives seemed lukewarm about Twitter, with less than a quarter of their candidates signing up for it. Prior to the formal Election campaign starting, there were indications from Conservative headquarters that the messages going out on Twitter would be tightly controlled, but we saw no real evidence of this. Most of the traffic seemed to be fairly spontaneous and authentic. Rene Kinzett (Swansea West) posted more messages than any other candidate during the campaign, a mixture of views, links, news and debate, and was without a doubt the most active candidate from any party using the medium. As a result he was able to build up an audience of over 800 by voting day, more than any other candidate apart from sitting MPs. In Delyn the Conservative candidate Antoinette Sandbach engaged well with the constituency, as did David Jones who was defending Clwyd West for the party.

In Wales, though, the Conservatives performed less well than nationally, increasing their share of the vote by just two percentage points, against a national rise of 3.8%. Their candidates' more successful Tweeters all performed above average except Matt Wright in the Vale of Clwyd, whose vote only rose by 3.5%. Their star performers on election night (Glyn Davies who took Montgomery from the Liberals and Karen Robson who came close in Cardiff Central) did not use Twitter. Neither did Simon Hart who won Carmarthen West from Labour. Anthony Ridge-Newman, though, increased his vote in Ynys Mon with an active Twitter campaign as part of his election armoury.

The Green Party's national vote held constant, something reflected in Wales where there was no noticeable change. The party fielded 13 candidates, of whom only 3 used Twitter. Their most competent user was Sam Coates in Cardiff Central who performed above average. The two other users were less engaged than other candidates in Wales using the medium, and also performed worse than the Welsh average for the party, actually losing votes.

Just one third of Labour's candidates in the election were signed up to Twitter. The most active users were three sitting Members, Chris Bryant (Rhondda), Julie Morgan (Cardiff North) and Kevin Brennan (Cardiff West). Only Bryant made it into the overall Wales top ten, though, with a mixture of news, view, pictures and chat. However, Bryant lost nearly 13% of his vote which was one of the worst showings across Wales as a whole. Five other Labour MP were represented on Twitter, but their efforts and those of other party candidates were less noteworthy. Several candidates failed to use their Twitter accounts at all, including Merthyr MP Dai Havard, who performed even worse than Bryant, as did token Twitter user Owen Smith in Pontypridd.

Nationally Labour lost 6.2% of its votes, although in Wales the position was slightly better with a decline of 5.8%. One of the party's most active Tweeters (Julie Morgan in Cardiff) lost her seat to the Conservatives, but actually did less badly in share of the vote terms than most of her colleagues. The same is true of Kevin Brennan (who kept his seat in Cardiff West). Indeed, Nick Smith (Blaeneau Gwent), Madelaine Moon (Bridgend) and Richard Boudier (Ynys Mon) who were all active Tweeters (albeit is a modest way), bucked the national trend and increased their share of the vote (in Smith's case by over 20%). And only 2 of the 20 under-performing Labour candidates used Twitter, and one of these in a minor way. Following the election, Labour had 26 MPs in Wales, but only 8 of these were using Twitter actively.

The Welsh Liberal Democrats also had a comparatively strong presence in Twitter with a third of their candidates registered to use it. Three of these featured in our top ten for Wales, including Matt Smith (Blaenau Gwent), John Dixon (Cardiff North) and Amy Kitcher (Merthyr). Once again, though, there were variations: 3 of the Liberals' nominal Tweeters didn't use the medium at all during the campaign – including, surprisingly, MP Lembit Opik who had the advantage of a very large audience already built up. Once again, the Liberal Democrats used Twitter to build an element of urgency and excitement into their campaigns, highlighting local issues and events that were happening on the ground.

It could be said that Twitter for the Liberals was a case of missed opportunities: MP Opik lost his seat, while Jenny Willott in Cardiff Central (not a Twitter user) lost nearly 9% of her votes – these two MPs representing the worst performing Liberal candidates in Wales in 2010. Twitter is by no means all the answer, though: only 3 of the top performing Liberal candidate were active users, while John Dixon's skill with Twitter in Cardiff North merely resulted in a below-par result, a minor drop in his vote. Nationally the Liberal Democrats increased their share by 1%: in Wales, though, they did better, achieving a rise of double that figure.

Of the parties, the most active was Plaid Cymru, where over half of the candidates were using Twitter. Of these we assessed four to be in the top ten users in Wales during the campaign. These were Heledd Fychan (Montgomery), Caryl Wyn Jones (Vale of Clwyd), Ian Johnson (Vale of Glamorgan) and Myfanwy Davies (Llanelli). However, the picture is confusing as Wyn Jones under-performed the party as a whole when it came to getting the vote. The same was true of another active Twitter user, Jonathan Edwards in Carmarthen East, who dropped 11% against a 1% drop for the party as a whole. Similarly, the large number of Plaid users can be misleading: four of their candidates had Twitter accounts but didn't use them at all during the campaign, and a similar number were only occasional users. Fychan and Jones were both very effective communicators, using a mixture of comment, pictures, videos and cross links to other web pages: but while the former added votes to her tally in Montgomery, the latter lost support (in both cases, it has to be said, in the 1-2% range. The active Plaid users engaged with their audience, answering questions and taking part in debate.

Independent candidates in Wales increased their share of the vote by 2%. Only two candidates used the medium: George Burke in Cardiff South was highly engaged, but performed less well than the average for Independents in Wales; Dave Rees in Islwyn, by contrast, spent very little time on Twitter, but increased his share of the vote by over 4%. Similarly, UKIP which contested all Welsh seats increased its share of the vote above the UK national average, when there was virtually no usage of Twitter. Likewise, the Christian Party which fielded 8 candidates in Wales performed exactly in line with other party showings in the UK as a whole – holding their vote. No candidates from the Christians used Twitter in Wales.

Conclusions

Active Twitter users in Wales tended to perform better in terms of the final vote than non-users or light users. This is not to say that the rises were due exclusively to Twitter – as one candidate told us “Twitter was fun: the real campaign was elsewhere”. Voting was affected by macro issues from the national campaign such as the Leaders' Debates, the so-called bigotgate, controversies over off-message comments by various candidates, etc. But in many of these cases Twitter contributed as commentators and ordinary voters took to the medium to air their views. This is still very much the case during the current negotiations about forming a government. In other words Twitter's influence on a range of audiences was probably greater than the crude voting figures give credit for.

Far from being condemned as ineffective and faddish, these results suggest that Twitter is a medium whose potential has yet to be realized. Many candidates (particularly the younger ones) were starting to experiment with ways to use Twitter through their iPhone, getting across with images and text the immediacy of the campaign on the ground. This was something that was not in evidence at all a year ago during the elections to the European Parliament. The seemingly unstoppable rise in the sales of smartphones is already changing the dynamics of communication with social media conversations increasingly moving from fixed machines and laptops to mobile devices. At the same time, social media conversations are evolving from text-based posts to multimedia offerings – almost all of this driven by iPhone ownership. Twitter is already benefiting from this trend, and politicians will need to respond in terms of their ability to react and in choosing the narratives they need to develop.

11 May 2010



The study was carried out in Wales by Robin Croft (Principal Lecturer, University of Glamorgan Business School), and in Yorkshire by Dianne Dean (Senior Lecturer, University of Hull Business School). The findings of this study and the research completed in the European Elections are due to be presented at a conference of the Political Marketing Association in Thessaloniki, Greece, in September.