How the New Consumer Giants Do Research: Insights Examples From Spotify And Airbnb

Our leisure time is increasingly shaped by a series of consumer-facing challenger brands: Uber, Spotify, Netflix, Airbnb. They’ve all used the flywheel effects of consumer-generated data to their advantage: they have access to information and customer feedback in staggering quantities and can use it to continuously improve.

When these brands first emerged, some people asked what need they would have of research? Market research in its old sense of generating and structuring information for businesses seemed irrelevant to these new titans of consumer analytics.

But of course, that wasn’t the case. All these brands conduct market research. They just do it in different ways. On the first morning of TMRE 2018, presenters from two of these businesses – Spotify’s Marion Boeri and Airbnb’s Yoni Karpfen – gave their spin on how they used insights.

The two presentations were quite different, but shared one crucial thing in common. For both brands, research existed to help fulfil a core part of the brand’s purpose or mission. For Spotify, that was its claim to understand the world through music. What someone listens to defines them – and by guiding and deepening that listening, Spotify can improve people’s lives. Or that’s the theory. Their research project – a hugely complex multi-modal piece working with agency YPulse – took understanding people through music as its start point, and asked: where’s the proof?

For Airbnb, the role of research is intimately tied to the company’s purpose vis a vis their hosts – to help every host become the best “hospitality entrepreneur” they can be. The possibility of how good that ‘can be’ is left open, and no surprise: 82% of low-performing hosts rate themselves an “A” already.

For any brand whose service generates data, the central research problem is this: how does what I measure about people relate to how they actually live? What is the relationship between data and reality?

For Spotify, the problem was everything the data leaves out. Listening to music is part of almost everyone’s life, but Spotify didn’t know exactly where in the jigsaw of those lives that part fitted. This matters, because Spotify’s product is its audience.

For Airbnb, the problem was that what they know to be true from the data doesn’t match the perceived reality. People think they are good hosts, when they’re not. And when they’re not, it’s Airbnb who gets the backlash, despite having zero actual control over the guest experience. (It’s not Yoni Karpfen who doesn’t clean those sheets.)

What were the solutions? Spotify’s was simply: more data. With YPulse it talked us through a mighty octopus of a project – eight discrete stages, including everything from pop-up online communities, to quant ad recall metrics, to video ethnography and a “teen summit” for idea generation. Oh, and halfway through Spotify got YPulse to go back and recontact everyone, so the brand could fit the insights more closely to its existing language and frameworks.

Integrating and bringing to life this mountain of data sounds like a nightmare brief – I got the strong impression the unsung heroes of the project were software package Tableau, which mapped the vast universe of musical moments identified; and designer Oliver Pangborn, who crafted the wealth of information into story guides which helped Spotify cement its position as thought leaders about millennials and music. It was worth it in the end, but the journey doesn’t sound fun.

As for Airbnb, their approach was different, focused on taking existing consumer feedback – which the firm is swimming in – and using storytelling to provoke internal action and change. Research, Karpfen informed us, is of no value whatsoever…. unless it changes how somebody acts. The firm took what they’d found about their high-performing and low-performing hosts and mapped a journey towards mastery which they could help hosts along. With little direct control, their options for doing this are limited, and the same gentle persuasion Karpfen relies on to move his colleagues was also in evidence in the firm’s philosophy towards its hosts.

Karpfen talked about enforcement and empowerment. Enforcement can take a host from the most rudimentary level – providing basic shelter with no real inkling of customer service – to a point where they can enter the Airbnb ecosystem. Beyond that, empowerment must take over.  A lot of Karpfen’s method for using research was creating simple structures and models which his colleagues could use to explore how they related to hosts. If there were six factors which separated an entirely consensual model from an entirely coercive one, where on each factor did the optimal balance lie: between “Do” and “Don’t”, to take the simplest instance?

The variety of answers to these produce different worlds – or different stories Airbnb can tell its hosts about themselves. Karpfen saw his job as helping his company explore these worlds and tell these stories. “Research creates order from chaos”, he said – a refreshingly clear vision of what insights do.

He left us with a challenge – “who’s driving the bus?” If the insights team is reactive, struggling to fill the gaps in knowledge left by consumer data, then it’s not in control of the journey. But if research can tell useful stories – creating elegant models and ideas that proactively explain people’s behavior – then it’s in the driving seat.