Why even the richest behavioural data cannot explain itself.
In last week’s Superfan Formula, I made what should be a fairly uncontroversial observation:
Data tells us what happened. Insight helps us understand why.
This week, I want to explain, in the simplest and most relatable way, why that distinction matters, using one of the world’s most data-rich businesses as an example…Spotify.
Our lovely friends from Stockholm - although, in the interests of accuracy, it was more often than not their colleagues in New York with whom we worked - knew an extraordinary amount about how people used the service.
They could tell you who listened, what they listened to, when and where they listened, what device they used, how often they returned, which features they explored, which tracks they skipped, who moved from Spotify Free to Premium and, that favourite metric of many a lazy marketer, their age and gender.
If any business should have been able to find the answer within its behavioural data, surely it was Spotify.
Surely.
Except the answer wasn’t there.
WHY WEREN’T PEOPLE UPGRADING?
During our time at EMI and Warner Music, we worked closely with Spotify to help them understand motivations, perceptions and emotions of different tribes of music fan across ten countries.
It seemed perfectly logical to me and the team. I mean at that time, the music business was on its arse. We were well beyond the last days of disco; the glitterball was already a dim sparkle in the rear-view mirror.
Not everyone agreed. I still find it puzzling why an individual on the board criticised me for helping “build Spotify’s business model”. This seems particularly curious now, considering that streaming has helped the music business generate more money than ever, well, for the record companies at least, and, ironically, that individual too.
Nowt as strange as folks, as they say.
Anyway, I digress.
One of the many questions and behaviours that we explored was why so many users of Spotify’s free service were reluctant to upgrade to Premium, despite the apparently obvious benefits. The predictable assumption would have been price. People were using something for free. Spotify was asking them to pay for it. Ergo, they didn’t want to upgrade because they did not consider the additional benefits worth the money.
Perfectly logical. And hey, for some people, perfectly true. But it wasn’t the whole explanation.
When we spoke properly to music fans, observed them, we found that a significant barrier was not the actual cost or capability of Spotify Premium. It was what people incorrectly believed about it.
In our research at the time, nearly half of Spotify Free users thought Premium had limitations that simply didn’t exist. They were judging, and rejecting, the proposition based upon an inaccurate understanding of what they would receive.
Perhaps even more remarkably, around a quarter of existing Premium subscribers didn’t know that features they wanted were already available to them. Spotify could see that people weren’t converting. They could see which features users…er…used. It could see when they disengaged.
Bu what the behavioural data could not reveal, by itself, was that some people were making completely rational decisions based upon completely incorrect information.
The behaviour was visible. The barrier was not.
THE PEOPLE OUTSIDE THE DATA
There was another challenge.
Spotify had achieved enormous name recognition. Its advertising had very successfully established the brand, and music lovers generally knew that Spotify had something to do with music.
Sorry, another quick aside. Having worked in brand communication for many years, nothing an advertising agency loves more than a good old brand-awareness and name-recognition campaign.
Come to think of it, nothing strokes a founder’s ego quite like seeing their brand, pumping out of a prime-time advertising slot.
Right, sorry…I digress yet again.
So where were we, oh yes…many did not properly understand what Spotify actually was, what it offered or why it might be relevant to them. The misconceptions were significant. And those misconceptions were materially affecting the speed of its growth.
And here is the rather obvious problem: you cannot examine the behaviour of people who have never entered your system. Spotify’s existing-user data could reveal how registered users behaved once they arrived. It couldn’t fully explain why millions of knowledgeable music consumers had not arrived at all.
Those people had not subscribed, registered or started listening. They had generated no meaningful customer journey to analyse. They were effectively invisible. To understand them, we had to leave the dashboard and enter their world. We had to speak to them.
BEHAVIOUR DOESN’T COME WITH AN EXPLANATION
This is not a criticism of behavioural data. It is essential evidence. But the mistake many of us make in the music business, and, for that matter, across sport and entertainment, is believing that behaviour explains itself. It doesn’t.
At Sound Effects, we can show you, subject to an NDA of course, multiple examples, from Brazil to Beijing, of two people exhibiting precisely the same behaviour for entirely different reasons. Two more may do nothing, one because they are uninterested and the other because they profoundly misunderstand what is being offered.
On a dashboard, both appear as non-converters.
Commercially and strategically, they represent completely different challenges. One may never be persuaded. The other may require nothing more than the right explanation. If we treat them as identical, we will probably make the wrong decision very efficiently.
SPORT HAS THE SAME BLIND SPOT
A football, rugby or cricket club may know that thousands of local families attend only once or twice each season. The ticketing data can identify which fixture they chose, how much they spent, where they sat and whether they returned. It cannot automatically tell the club why they did not return more frequently.
There could be a multitude of reasons, some of them surprising. Spotify was certainly surprised by what we showed them.
Was the ticket too expensive? Was travelling with children too difficult? Did the fixture time disrupt family routines? Did the children become bored? Did the parents feel uncomfortable in the stadium? Or did the experience simply fail to create any emotional connection?
Each explanation demands a completely different response.
A discount will not solve a transport problem. More children’s content will not overcome a parent’s concern about atmosphere. And a loyalty programme will not help somebody who never felt they belonged.
Yet without understanding the motivation or barrier, organisations frequently respond with the tools they already possess: another offer, another email, another notification or another piece of content.
MOVE FROM OPTIMISATION TO GROWTH
Behavioural data is exceptionally good at helping organisations optimise what already exists. It can show which of two emails generated more opens, which ticket offer converted better or which piece of content held attention for longer. But the most successful option is only the best of the options somebody thought to test.
It cannot reveal the stronger proposition nobody considered, the misconception nobody knew existed or the future fan who has never entered the database. That requires a different kind of enquiry, my friend.
It requires us to explore people’s motivations, identities, emotions, assumptions and wider lives, not simply their interactions with us that show up on a pretty dashboard.
Because the greatest growth opportunity may not be hidden within what fans are doing. It may be found within why they are doing it, or why somebody else is doing nothing at all.
NEXT WEEK:
Owning fan data is essential. But owning the record of someone’s behaviour is not the same as understanding the person, and your CRM cannot introduce you to the people who have never entered it.