Welcome Guillaume Liegey, one of the co-founders of LMP. Can you give us a brief overview on yourself and your company?
Sure. Thanks for having me. I am one of the co-founders and CEO of LMP. LMP is a campaign technology company, who developed a software to help people basically to win votes during an election. We are based in Paris and now Berlin, we recently started a new office in Berlin and we are very happy to hopefully soon start working on German campaigns. At the core of what we do there is the combination of three ideas, data, technology and human. If you are able to combine data, technology and humans, that is the recipe for the best campaign today.
You are basically developing an app for door to door campaigning. What is it about?
The idea of door to door campaigning is, that if you want to change someone’s mind, the best way to do it, is to engage with people face to face. That is what science tells us. There have been scientific experiments in the US and in Europe, that show that direct contact, door to door, is the most effective way to gain votes. That is the first thing. The problem is, you have to know, at which door to knock at, because it takes time to knock at doors. You have to choose what are the right doors, what are the right neighborhoods. Therefore you have to use data and predictive algorithms, that make sure, you use the right doors and talk to the right people. The second thing is, you need technology to scale up your campaign. Technology makes it possible to be more efficient, to gain and save a lot of time when organizing a door to door campaign. Our app, that we developed, does this, it helps to choose the right neighborhood and helps organizing field actions like door to door.
This kind of campaigning was developed in the US, where data protection is less important than in Europe, where do you actually get your data from?
That’s a very good point. Of course it started in the US. We were very lucky, my partners and I, to be part of the Obama campaign in 2008. We were volunteers for Obama and that’s where we discovered, how American campaigns were using technology, data and human to be more efficient in their campaigns. The data question is very important, in most European countries those privacy laws are very strict. All across Europe, if you leave aside the UK, but in France, Germany, Italy, Spain, Belgium and Switzerland, all the countries, where we have worked, data privacy is very important and the laws are very strict. But it is not the problem, when you are organizing a campaign, because we use aggregated data. We don’t target individual people, we target neighborhoods. For example, in Berlin, I’m going to tell you in which neighborhood you should campaign and in which neighborhood, there is no use in spending time. For that we use electoral results, which we get very often from public sources, it is open data. We also combine electoral results with sensors data, what is the age of the people living in that neighborhood, what type of job do the have, what is their level of education, how much money do they make, what is their level of employment, what is the profile of the households, etc. But all that data is at an aggregate level, so that we comply with privacy laws. We put all this data in an predictive algorithm, and that algorithm will give us a priority score to a neighborhood and that’s what you use when it comes to choosing, where you want to send your volunteers.
Are you also collecting new data during conversations with potential voters?
Certainly, of course, when you have the opportunity to have a conversation with them, they give you a lot of information. That’s what we did with Emmanuel Macron, when we started his new political movement, „En Marche”, last year. He organized a massive door to door campaign outside of any election campaign, because he wanted to listen to people, he wanted to be able to understand, how normal people see France, see the country, what they think the key problems are and wanted to use that information, to be able to build a more concrete, more pragmatic program. So, yes, in a way you can use those informations, that people share with you, you have to mention, that you are using it, they have to give you an explicit consent, etc. But if they do all that, than you can use, what they tell you, for explicit purposes. One possibility is, what Macron did, to diagnose the country, but you can also use it, to recruit new supporters for your party. You can also use it, to test some of your key messages and test, how people react. You can also use it to test your candidates popularity. There are many question, people can ask with our apps, it depends on the type of campaign you are running. If you are running a get all the votes campaign, a few weeks before an election, I think the key message is, „get out and vote, here are the key messages you need to know”. If you did what Macron did, a listening campaign outside of an election, you need to listen and learn from voters.
You were in the US during the presidential elections in 2008, during the first Obama campaign, how come and what did you learn?
Initially we don’t come from politics, my co-founder and I. I worked in China, I worked with McKinsey, the consultant company, as did one of the other co-founders. The third, Vincent Pons did a Ph.D. in Economics at MIT, now he is a professor in Harvard Business School. We don’t come from politics, but we were somehow captured by the Obama Campaign. It was difficult to resist and not join the campaign, because there was so much enthusiasm. We were in the US, because we went back to school. After a few years in professional live, I went back to school, to the Kennedy School of Government at Harvard and I arrived in August 2008. Literally 80% of my classmates were volunteering for Obama. At some point, I went to a guy in the class with a „yes, we can” t-shirt. I asked him, „I’m sorry, I am French, but can I do something?”, and he said, „of course, dude, go to the website”. So I went to Obama’s website, I registered, I left my phone number, e-mail, address and my zip code. Literally, four or five hours afterwards, someone called me and asked if I wanted to join a door to door campaign the next Saturday. By the time, I had no idea, what door to door was, but I said, „let’s try” and that was how I discovered it. Next Saturday I met with a group of volunteers on a parking lot and we all drove to New Hampshire, which was a swing state, north of Massachusetts. Arriving there, they gave us a short training session, they gave us a list of addresses and said, „knock at their door and write down what people say”. I was like „wow, how do you know those addresses and not others?”. I started asking questions after the session and people explained to me, what I’ve told you, there has been a lot of experiments in campaigning and they all have the same conclusion, door to door is the most effective way to gain votes. That’s the first thing, and the second thing is, we can use data to choose the people who you want to talk to, or the neighborhood, in which you want to campaign. I thought this was interesting and had to no idea, that this was possible and with Arthur Muller and Vincent Pons, the two other founders of the company, we imported this idea in France first and ran Francois Hollands field campaign for the 2012 election. That was somehow the first big test for the methods of the Obama campaign in an European context. It was such a success, that we started LMP afterwards.
There are also critics which say that door to door campaigning is not working, some say it is too expensive regarding its effects…
Good point, when you send an e-mail you just push a button and have literally a million e-mail addresses, reaching a million people. Same thing with social networks. Actually the studies I’ve mentioned, a lot of them do calculate costs, so they value the dollar equivalent of an hour of volunteering. But even if you valuate cost, it is still more effective. For example, we ran an experiment, it is a scientific paper, you can view on my colleagues Vincent Pons Harvard website. We measured the effect of Hollands door to door campaign and the result was, that five million doors knocked by 80,000 volunteers, not paid, they are volunteers, helped Hollande gain 250,000 votes. That’s not what made Hollande win, it’s 20% of the margin victory against Sarkozy, he probably would have won without it, but it is still a very significant gain of votes. I would still say that a well organized campaign using technology and data, etc. can help win 2-4%, not 10 or 20%. You have to be measured, to be reasonable, if someone tells you, he can help you win 10% or 20%, that person is a liar. We never tell that to our clients. We tell them, if you run a larger campaign, that it will be more efficient and reach out to more people, reach out to the right people, which minds can be changed by a campaign. I always tell them, „eventually, you will win 2, 3 or 4%”. That matters for a lot of elections, if you look at the French presidential elections, that takes place right now, it is very, very close, 2% can make a huge difference. Also, in a campaign there are many things you can not control, like a scandal, but you should focus on what you can control. That is our promise, that you can better control your volunteer base and talk to the right voters.
As you already said, door to door campaigning relies on volunteers, it is therefore easier for bigger parties, with a lot of members. How about smaller parties, like the Green Party in Germany?
We worked for very big parties and big campaigns, we worked for very small campaigns, people running for mayor in a city with ten thousand inhabitants, we actually did work for the Green Party in Bavaria in 2013, for the regional election. The greens are a good example, because they are now running a national door to door campaign, I think. Even they don’t have the same base as the CDU or SPD, they do have volunteers and they are making the most of what their volunteers are able to do and they’re right to do it. They might not have the same numbers as the SPD, but it still has much more effect than tweeting or posting things on Facebook. Social networks are powerful for a lot of things, for example, to mobilize your supporters, to share information, etc., but from a scientific perspective, there is no evidence, that you can change people’s mind using a social network. It’s like the same people talking to the same people, it is like a reinforcement of believes, but you don’t expand the base. So if you are a campaign manager from a small party, you have to use your volunteers’ time in the most effective way. I really believe, that field work is the most effective way to use it and than you adapt your scale on how many people you have. „En Marche” is a good example, they had 0 volunteers when they started, after working with us, they had 6,000 volunteers, who knocked at 250,000 doors in two months, outside of any election period, so there was no urgency and yet it worked. One of the reasons was that people, the volunteers, enjoyed it. They were able to get in touch with people outside their work space and expand their view on what France is. It is really a very positive experience for volunteers and this matters a lot when you want to work with volunteers.
Back to the US elections, how much does the candidate matter for door to door elections? If you compare Clinton to Obama…
The first thing you have to know is, I would love to have a software able to clone Barrack Obama, but technology does not make this possible. But you’re right, the character of the candidate matters when it comes to mobilizing volunteers and there is so much technology and data can do alone, but in the end you need the third pillow, people. If you don’t have them, you can have the best technology in the world, it won’t do anything. The type of the candidate matters a lot. There were less people joining Clinton’s campaign than it was the case for Obama, so she did get less people to knock doors and get less votes, what might have made the difference in some of the key states. But you don’t choose which candidate will be running for election, if you are the campaign manager, it is a data that you start with, it is a given, you have to optimize it. That is where innovations in campaigning are useful, because even if you have a candidate that nobody knows, you can start growing by organizing a field campaign. Today, technology makes it possible to run a campaign, almost as professional as Obama, even if you are someone running for mayor of a small town in Germany or France. That is the beauty of innovation
How did your work for Hollande running for the Parti Socialiste, an old and big party, and Macron, with his new movement „En Marche” differ?
They differ in two things, one positive, one negative. The effect that you have a lot of members to start with, is a tremendous advantage, you have a pool of potential volunteers that is very large. That’s good, „En Marche” did not have that and the Macron people wanted to keep the project secret until the launch of „En Marche”, so we started literally with no volunteers. The positive side was, that it was like a start-up, so it had to move very quickly, I could test many, many things to recruit new volunteers. „En Marche” was more reactive than an established party. But my experience with a lot of parties all across Europe is that old parties can also innovate, can also change. There still a lot of things old parties can do in order to transform themselves. People were doing amazing things within the Parti Socialiste or the SPD in Germany. The future in politics does not only rely on people starting new parties. The technology we bring benefits them all. We worked for 300 campaigns and we think there is hope from many different directions.
Since you recently started an office in Berlin, do you consider working for a German party in the upcoming federal elections?
The timing may sound like we want to be involved, but we opened an office in Berlin for two reasons. The first one is, that it is an important market when it comes to political campaigns but also for corporate clients. Germany also was a natural market after France, there are many elections, federal elections, local elections, there are elections all the time in Germany. The second reason is, that we already worked in Germany, I mentioned the Green Party in Bavaria, and we have a decent network that helped us meeting people when we opened the office. There is not much time until the federal elections this year, so our goal is more in the long term. We want the German company to be as successful as the French company. We are recruiting German speaking people right now, because as you noticed, I’m speaking English and not German, but you need to have a local team to be successful in the long term. Maybe we’ll work with candidates now, but our goal are the local elections and the federal elections after those in September.
Who will make it to the second round in the French elections on Sunday?
If you purely look at the data, there is a straight answer: it will be Macron and Le Pen. The reasonable answer is, that it is too close to call. It’s literally a four peoples’ race, Le Pen, Macron, Mélanchon and Fillon. They are one or two percentage points away, we never seen that in politics. Making a guess now is like throwing a coin. The level of undecided voters is higher than ever, that is why the campaigns try to converge them. My best guess is Macron and Le Pen, but I want to say that with a huge disclaimer, because this situation is unprecedented in French elections.
Thank you very much for the interview