Interview: Exploring Data Science at Brandwatch with Hamish Morgan
By Olivia SwainAug 23
Published May 16th 2019
For almost everyone with an interest in the future of the UK, 23 June 2016 was an odd day.
For me, it involved clambering out of a hot tent at Glastonbury Festival to hear the news that the UK was leaving the EU. I wasn’t best pleased by this news, but also I was at Glastonbury so I was feeling pretty happy about the day ahead.
For Christopher Kahler, CEO and co-founder of Qriously, the day was even more of an emotional roller coaster.
“June 23rd was actually a really fateful day for us. One because of Brexit. Second, it was my birthday. And third, we were waiting on results for a final decision to raise a big VC round. So three things coincided in one crazy-ass day.”
By most accounts, the Brexit result was unexpected. Polls had predicted a win for the Remainers and plenty of pro-Europe voters went to bed the night before confident that the Britain they’d wake up to wouldn’t be much different to the one they’d drifted off in.
Few predicted that the UK electorate would opt to leave the European Union, but Qriously, a market research startup based in London that replaces traditional smartphone display ads with surveys, did.
“Brexit was our moment of truth. To make a good prediction like that and for it to stand out, you have to have a certain alignment of events. Namely, the event has to be of great importance. There needs to be consensus against what your method is predicting. Third, you have to be right in predicting the contrary result.
“Those things aligned for us with Brexit and later with Trump and that’s when we knew the methodology worked.”
Let’s start at the beginning of the Qriously journey and work up to that moment in 2016 when they realized how powerful the technology they’d created truly was.
The Qriously story begins nine years ago when Christopher was working in Shanghai on free consumer-facing apps that were getting thousands of downloads.
“We were monetizing them with ads and we invested a lot of time in making the UX and the design stand out. It was kind of a shame to spend so much time on the design and then see a nasty ad at the bottom,” he says. This was what got him thinking about the nature of ads and what they could be used for.
Essentially, ads allow advertisers to buy people’s time and attention. Instead of thinking of this simply as a selling channel, Christopher and his co-founders Abraham Müller and Geri Mueller began thinking about using ad space in a way that looked great and gave advertisers the opportunity to get feedback from the people they were targeting. They got to work on their first iteration.
“We thought, ‘can the user answer a simple question or a series of questions instead of clicking an ad and have that be the value exchange instead? So they have a free app and in exchange they share their view on something. We thought that was a cool dynamic to explore and that’s how we got started.”
In true start-up fashion, the roles of the founding members were flexible. Christopher spent a lot of time in product design as well as running things on the commercial side. The mindset he learned in his background in mechanical engineering was applied to all manner of problems, despite those problems mainly playing out on screens instead of in machinery.
The team attracted some early investment, but Christopher admits they ran out of cash and things had to go back to basics (in every possible way) for a while.
“We couldn’t afford rent for all of us. We shared one small single bedroom apartment in Shanghai and we all slept in the same bed. We could only afford to take showers every three days so it just compounded matters and made it really terrible. Being poor in China, you have to be talented at being poor – it’s like next level poverty, so that was really humbling.”
Luckily the team’s fortunes turned around. But not at first.
They took their product to continental Europe to share with potential VC investors and were promptly turned down by everyone they pitched to.
In the UK, however, things were very different. Everyone they approached gave them an offer.
“It was a dream scenario fundraising-wise, because we were out to raise a couple of hundred thousand and ended up getting 1.6 million from one of the best VCs on the continent,” Christopher says.
The first version of Qriously sat firmly in the ad space, Christopher says. It didn’t feel quite right – it worked well, but they wanted to expand further.
“We iterated in 2016 to build a research methodology which is still our focus today. For that we built a data science team and a research team, to see if the method produced viable results and if it had any meaningful improvements over traditional methods. It did, and that’s where we are now.”
It was in 2016 that they found Qriously accurately predicted the results of both Brexit and the election in the US.
I quizzed Christopher on the team’s methodology for this research.
“Our innovation is on the sampling side and using an approach where respondents aren’t incentivized. Apart from that we followed standard research methodology – we weighted on age, gender, income, education, that kind of thing. Apart from respondents that weren’t financially incentivised to answer, we also had an advantage over traditional methods through location from the devices, so we could weight the data much more accurately.”
The company grew to 17 employees, including Christopher, Geri and Abraham, for the most part based in the UK. Christopher runs commercial, but he can’t help but spend time with different parts of Qriously throughout the week.
“I don’t like routine. I’m an eternal seeker of novelty, and I’m addicted to progress and things getting better. Routine for me is really boring and somehow stops me thinking about stuff,” he says.
When I ask what the Qriously team was thinking about when they entered into talks with Brandwatch, Christopher takes things back to the nature of the offering he helped to build.
“Research is all about trying to understand what people think and to try, at a very abstract level, to predict meaningful human behaviour. And when you’re using surveys you realise that that method alone has certain shortcomings. People don’t always do what they say.”
Qriously removes many of the biases that might hinder traditional forms of surveying different groups, but he says there are still inherent limitations to the question-answer type of approach.
“We always thought ‘wouldn’t it be cool if we had different sources of data?’ If we were trying to come up with the ultimate research method without any prior knowledge of how it’s done now, what would that method look like? We always believed as the founders that it would be some combination of our data and what people would naturally say – prompted and unprompted.”
And so, in early 2019, Qriously joined the Brandwatch family.
“If we had just continued to fundraise we’d be stuck in the survey space. But if we can combine our method with Brandwatch’s method and our data with Brandwatch’s data then we can build the ultimate tool that we set out to build in the very beginning,” Christopher says.
“Market research is a huge market and it’s hilariously undisrupted. If we can understand what people really think, which is what market research is at its core, and get it that much faster, organizations will be able to make bolder, more informed decisions at a speed that fits business needs as opposed to vendor limitations. Who doesn’t need that?”
Thanks so much to Christopher for chatting with the Brandwatch blog.
Stormy weather is forecast in Chicago this week, but the real thunder is happening at Brandwatch's flagship conference.