Data-Driven vs Dialogical Canvassing

Joseph Buckley
Nov 3 · 4 min read

I first knocked on a door for the Labour Party about 4 and a half years ago, and since I have done far too much canvassing. From general elections in glorious sunshine, to referendums in pouring rain, to local election campaigns in snow. Today I was up in Chingford and Woodford Green (campaigning for the amazing Faiza Shaheen), and the experience has made me quite reflective.

During canvassing sessions I am usually the board runner, the person with the clipboard telling people where to go and recording the data. Running the board can be pretty challenging if you’re with a big team in a complex estate, but for the most part its quite easy, and the time passes quickly enough. One of the upsides of running the board is you get a good sense of what is actually going on, and you can get a good feel of whether things are going to be good or bad. The downside of running the board is that is can get repetitive, and occasionally stressful. For a while, I thought that my favourite thing about running boards was that I didn’t have to talk to voters (definitely a plus after a heavy night out). After today (although admittedly I didn’t knock on a door), I don’t think that’s the case, and my issue with knocking on doors was more to do with the style of canvassing.

I think the ways to canvas roughly be split into two categories, although like everything its more of a spectrum than a binary. The first, which I will call data-driven canvassing, is the type that Labour Party organisers and statistics fans love. The idea is as follows: (a) a lot of labour supporters don’t vote, (b) we need to get them to vote, © we have limited resources and we don’t want to be wasting time asking Tories to vote on election day, (d) we need to have good data on where our Labour voters are. This line of thinking has a lot of merit to it, and I know people in my local party that could tell me exactly which buildings in which estates vote labour. This method of canvassing is also much quicker (“Hi, I’m here from the local labour party. Can I ask who you’re going to vote for? Can I ask who you voted for last time? Thanks, bye!”), allows you to cover more area, generates useful data, and has a much lower bar for entry than any other style — you don’t need any training to ask those two questions.

The second way, I will dub dialogical canvassing. This way looks to have a meaningful conversation with a voter, which will hopefully result in the voter deciding to vote Labour. The idea behind this is obvious, If it could be done effectively, and we could have a meaningful conversation with every Tory voter in the country, there would be a lot less Tory voters. This method however has multiple drawbacks. Firstly, it takes too long; if it takes 10 minutes to have a meaningful conversation a group of 5 canvassers could talk to 48 people in two hours, compared to 480 for the data-driven method. That means 1500 sessions to speak to an entire constituency compared to 150. Secondly, having a meaningful dialogue with a stranger, and then being able to leverage that to help said stranger achieve some momentary critical consciousness in which they decide to vote Labour is not easy, and sadly we do not live in a world of bell hooks’s and Paulo Friere’s. Further to this, the majority of people you talk to on the doorstep don’t want to engage more than a 10 second interaction and if they do, they probably have strong political beliefs already, which will render the whole exercise a waste of everyone’s time. Finally, the data you get from this style of canvassing is generally bad.

Its very easy to convince a non-voter to tell you they’re going to vote, but post-election the Labour Party gets to see who actually voted and who didn’t. There are so many people who start off being recorded as non-voters, will get canvassed, turned into Labour voters, then on election day don’t vote. While we have some evidence of the efficacy of data-driven canvassing, there is nothing whatsoever to indicate dialogical canvassing actually works. I have never met someone who has told me “I used to vote Tory, but someone knocked on my door once and changed my mind”.

So, am I arguing that data-driven canvassing is the way forward? No, not for general elections. For local elections I think data-driven canvassing is the only realistic option — you are lucky to have more than 5 people out for a local election, voters are less likely to be aware its happening, and people generally care less. While I think a squeeze can be useful (“you know its between Labour and the Tories, there’s no point voting Lib Dem”) and maybe a good fact can help (“did you know the old councillor from the opposition is in jail now”), I don’t think you can go further than that.

But for general elections, people are far more willing to engage, and to have conversations. People genuinely do care about the NHS, about education, about social security, about housing, as well as caring about their bins being

collected on time. People want to be won over, and I think its about finding the right balance. When we’ve got the whole establishment against us, we need to do what we can to reach people. While we’re probably not going to make someone experience conscientização on the doorstep, engaging people a bit more can’t hurt, and it certainly makes knocking on doors more fun.

Maybe I’ll start talking to voters this election instead of barking “number 48 is Barbara” at canvassers, if I’m not too hungover at least…


Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade