Guillaume Liegey: ‘Big data is cheaper for electoral analysis than polls’

Big data can help politicians run more effective electoral campaigns. [RachelH_/Flickr]

This article is part of our special report Big data meets politics.

Political campaigns can be refined by the analysis of “big data” to analyse and target voters. The challenge is knowing how best to use the data, Guillaume Liegey told EURACTIV France.

Guillaume Liegey is the founder of LiegeyMullerPons, a start-up specialised in electoral strategy that helps candidates and their activists target their activities, with the help of databases.

Access to electoral data is vital if big data is to be used in electoral campaigns. Is it as easy to access this information in Europe as it is in the United States?

Open data has greatly expanded in recent years: administrations have made lots of administrative data available. There is a certain convergence between the data available in Germany, France, Italy and Spain, for example. This can be sociological, economic, demographic information, etc.

But electoral data is different. We do not have any individual data on the electorate because it is prohibited, except in the United Kingdom.

Does this fact affect how political parties deal with big data?

We do not have the same analytical finesse. We act at the level of the voting booth: we know who voted which way in a given population.

But this kind of precision is not useful for making targeted phone calls or sending letters, which are not necessarily the most effective techniques. We do a good deal of our work by going door-to-door, and in this case we do not necessarily need to know people’s previous choices.

Has the use of big data been important in the Brexit campaign?

Yes, data has been important. For example, the Labour Party used databases crated during last year’s general election.

By making electoral campaigns more technological, doesn’t the use of big data risk dividing the rich and the poor parties?

Actually, data analysis is not very expensive. It costs around 20 times less than conducting a poll. And the data is quite reliable, whereas people can lie in polls, particularly over abstention and support for extreme parties.

But then it is also true that a well-filled database is not enough to win an election. The focus should be on quality: it is better to have 200,000 good quality contacts than a million less relevant entries.

The 2017 French presidential campaign is coming up. Will new technologies play an important role in this election?

It all depends on technology. Effective technologies are those that enable the mobilisation of activists, and for that we need to ally data and technology with human contact. Email campaigns have absolutely no effect for collecting votes: at best people just wonder how their email address ended up in the database. If we had the email addresses of everyone in France, it would have no effect.

On the other hand, email works for fundraising by communicating with the relevant population. The same is basically true of social networks: they are good for mobilising activists and communicating with certain target groups, like journalists and politicians.

We are seeing Europe’s traditional parties losing influence and citizens becoming depoliticised, with falling numbers of activists. Can technology revive people’s interest in politics?

On its own, not necessarily. Technology can help activists in their work. But that is all. To revitalise politics we need new people with new ideas. We would need dozens of Obamas… But there is no magic software to gain votes.

The traditional parties have deeply-rooted campaign habits and practices. How do party activists view this new approach with big data?

Political activists need to do interesting things. If you ask them to go campaigning on public transport at 6am, they will not be very motivated. And it is utterly pointless. But if we develop a technique to make contact direct and efficient, the activists will be very interested.

What is the benefit of door-to-door campaigning? Isn’t it rather an old-fashioned technique?

Direct contact was reintroduced in campaigns in the United States, and then in Europe. We need software that helps us target populations. This is something we used a lot during Hollande’s 2012 campaign and which has developed in recent years, since Obama’s election in 2008, in which I was an activist.

You are currently working for the French Minister for Economy Emmanuel Macron’s “En Marche” campaign. What is your strategy?

Firstly, we are organising a big door-to-door campaign to reach 100,000 people. We identify sample areas of 1,500 people, for example in the 11th district of Paris we have three such areas. Then we share the addresses with our activists, who are not chosen at random but based on socio-demographic criteria, with the aim of representing the population of the district. The activists then talk to the citizens and enter their responses into their smartphones with a specially-created application. We will start analysing the data at the end of August.

Is this a comparable strategy to the one used by Ségolène Royal in 2008?

Yes, that is an appropriate comparison, with a couple of differences: we do not ask people to contribute to our manifesto and we talk to everyone. Royal’s campaign targeted only party activists and sympathisers.

During the French regional elections, you quite accurately predicted the National Front’s progression to the second round. How did you do this? What data did you use?

We created a model to extrapolate the carry-over of votes between the two rounds in the regional elections. This is econometrics. We can identify a constituency, we understand the political evolution of the voters depending on the socio-demographic criteria.

You have worked all over Europe, are the political practices very different or can your electoral marketing models be adapted everywhere with no problem?

All countries think they are special. Of course there are some differences, but they are mostly in the legal detail. It depends on the local data protection authorities. But otherwise, our analysis techniques are the same and work in the same way.

Is it possible to recycle data from one campaign to another, and so to keep data for electoral purposes?

The French data protection authority is very vigilant: if we want to re-use data, the people concerned must explicitly agree. France is without doubt the country with the strictest data protection rules in the world, and we regularly work with our European partners on this subject.

Subscribe to our newsletters