In online political speed-dating, your worst match can easily be your best.

Co-authored by Ruben Treurniet

For many, 2016 felt like the pinnacle of Internet-led polarisation. The words ‘filterbubble’, ‘echochamber’, and, for some for the first time since high school, ‘algorithm’, were all proudly added to our collective vocabulary. It turns out that many people get their news from social media sources such as Facebook — an advertising platform that now functions as a means to provide carefully curated, personalised news stories that fits the users’ profile, consisting of both traditional media outlets trying to compensate for decreasing subscribers, flashy online media platforms that mostly copy others’ content, lots and lots of opinions, and of course ad-driven fake news articles. Mediabubbles have always existed, but the personalisation led by advanced algorithms and the in-your-face attitudes of social apps tend to make things a bit messier.

We experienced that it is difficult, if not near-impossible to have a political discussion with strangers — let alone with those who have a vastly different opinion. As mentioned, these barriers, both geographical or based on demographics, have always existed but it seems as if ‘the Internet’ neglects to seize the opportunity when it comes to inclusive politics.

People on Facebook flock together with like-minded people. Sure, there is a little room for discussion, but a well-intended comment underneath a shared article often invites a misinformed reply, a racist slur, an anti-semitic remark, a sexist ‘joke’, and so on. There is no discussion when no one expects you to listen — you don’t get points (likes) for listening.

So what if we create an environment that actively stimulates discussion? Force it upon people to engage with others? A few months before the Dutch elections of March 2017, we decided to create a platform with the intention to break the filterbubble, and to bring the political debate closer to citizens by allowing them to discuss important subjects one-on-one.

Our goal became clear. We needed to create an online platform that connects random voters with different political tastes together, all in less than a minute, with the opportunity to easily switch topics and discussion partners. The concept for, roughly translated ‘’, was almost ready.

So, yes, online speed-dating. For politics.

In order for this to work we had to consider a few important assumptions. First, political taste is a sensitive, personal subject. As seen with traditional pollsters, so-called ‘social desirability bias’, where citizens give answers to surveys based on what they deem socially acceptable yet when alone in the ballot booth they will vote whatever they want, is very present. Some people will not admit to voting far right (or far left, for that matter), out of fear for being called for example a racist or a terrorist sympathiser. We therefore decided that for the platform to work, it had to be anonymous.

Second, in order to create an intimate, open environment, the conversations needed to be strictly one-on-one. Group chats can be interesting but it makes it that much easier for people to succumb to peer pressure, leaving their honesty behind to score points.

The anonymous chat interface of

And third, partly as a result of our commitment to anonymity, we needed to consider the trolls: the ferocious trolls that roam the world wide web, hungry for virtual attention easily acquired by tormenting helpless users on social media. Not that difficult — simply remove their sources of nutrition (as if you would with a real troll), namely provoked responses and an audience to give them gratification. The live ‘roulette’ functionality that puts the ‘speed’ in ‘speed-dating’ allows users to simply leave abusive partners behind. Furthermore, the absence of an audience sort of takes the fun out of trolling. Not quite worth it when no one gets to see your ‘accomplishments’.

Last, we knew from the start that content would be key. The political topics and the statements needed to be curated in a way so it would appeal to a wide spectrum of voters. Moreover, the statements needed to be formulated in a way that would invite, or even provoke, discussion. Basically we were trying to make speed-dating less shallow.

This resulted in an intimate, low-threshold platform that allowed users to anonymously discuss political topics in a one-on-one environment, with the possibility to switch between voters with different party preferences. Users simply chose ‘their’ party (or said they didn’t know), and were then automatically matched with a political ‘opponent’.

Dutch voters could select their preferred political party on the homepage.

A couple of weeks of developing and two test panels later, the platform went live on the 15th of February, exactly one month before the elections. In that period, we managed to facilitate over 35.000 online discussions, in which more than 70.000 statements and topics were discussed. The platform quickly reached the national media, with one newspaper calling it ‘Political anti-Tinder’, while another dubbed it ‘Chatroulette for politics’. For privacy reasons we deleted all the transcripts, making the qualitative success a matter of testimonials.

Luckily, we received many of those in our inbox and on social media.

The testimonials varied from stories about a far-right voter that agreed with a Muslim on immigration policies, to a Greenparty voter who concluded on the basis of a discussion with a liberal that they, in true democratic spirit, agreed to disagree on practically everything. And that was precisely the point of the platform: bringing people together by facilitating communication, not (necessarily) by harmonizing their opinions. Mutual understanding and dialogue among citizens contribute to a healthy democracy, and according to the many positive reactions, the people actually liked doing it.

Not saving the transcripts presented another problem. One that was quickly turned into an opportunity. Since we kept no records of the many discussions held on the platform, the political discussion became something very abstract. Encouraged by the positive feedback we decided to visualise these encounters by inviting users to two physical locations, in order for them to do real-life political speed-dating: we made an offline version of an online platform that was based on offline political discussions.

On two live events, in Amsterdam and Rotterdam, we built a set and filmed all encounters between Dutch voters that responded to the call. In cooperation with a large national newspaper, NRC Handelsblad, we published the best clips in the seven days leading up to the elections.

It turns out that, given the proper platform, people that hold different opinions on politics can actually have a decent discussion. If you leave out the judgment and the audience, and create an intimate environment, people are more likely to open up and tell their peers how they actually feel. Besides the merits for its users, the method also generates valuable data. If you think about it, the method used is something of an interactive survey where users respond to questions while discussing these topics with their peers. This could potentially be a win-win situation, where citizens get more involved in politics while simultaneously providing policy makers with useful (and obviously anonymised) insights into what’s on people’s minds. It is about time that we employ online technology to strengthen the fibers of our democracy, one speed-date at a time.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.