The perils of polls… and blackout Friday

After political debates, the spinning begins. Everyone wants to make sure that the story is how their candidate dominated (and won) the debate.

And in these circumstances, snapshot polls – such as the one we organised last week after the Eurovision debate – can become political footballs, as parties try to rally their supporters to back their man or woman. And then into this mix comes our reliance on technology.

On Friday, the Europe Decides website crashed due to volume of traffic. We are flattered by – but did not fully expect – this level of interest in the website. The timing of the blackout was not great: our post-debate poll was running at the time, and the advertised link to the poll was not functioning – a source of frustration to many (not least us).

Despite the technical difficulties, we decided to close the poll at 13:00 CET on Friday, as previously advertised, with the following results:



We took the decision to close the poll based on the fact that:

  • we had made efforts – from just after the debate, when access to the site was slow – to advertise via Twitter an alternative link to the external site of the poll provider, thereby allowing people to cast their vote. This link was included in five tweets that gave an ‘@’ mention to the respective candidates and parties; the new link was also posted on Friday morning;
  • with 3,300 votes cast – around twice the number case in our poll after the debate in Maastricht – the increase was in line with the overall increase in discussion on social media about the debate. We therefore had no great reason to think that lots of people who wanted to vote could not do so, or had not done so, via this other link.

We understand that not everyone who wanted to vote uses Twitter, and that our email newsletter on Thursday gave only the original link. It was only later on Friday – after the closure of the poll at the previously advertised time of 13:00 CET – that we saw tweets by at least a couple of political parties calling on their supporters to back their candidates via our poll, using the original link.

Regarding the poll itself, we had a lot of votes from Greece, for Alexis Tsipras – approximately a third of all votes. We had put in place measures to prevent double-voting but, as ever, self-selecting polls can be manipulated by parties mobilising their supporters (although it should be noted that Tsipras came first or second in many other countries, too – it was not purely a Greek phenomenon).

We have made clear throughout (for both this poll and the post-Maastricht poll) that the results represent a snapshot, not a scientific poll – and that point stands.

Everyone has their own story from the debate and our poll is only a tiny part of the commentary that surrounded the debate. We understand why some people may be surprised or upset by the result, but unfortunately, technology sometimes lets us down at precisely the wrong moment.

We’ve taken steps to ensure this type of failure does not happen again and apologise to anyone who wanted to vote and could not.

But in the end, it is voters who will make their choice next week. And those are the only numbers that really matter.

David O’Leary
Project leader, Europe Decides
David Earnshaw
Chief Executive Officer, Burson-Marsteller Brussels