Technology Data Music Streaming

Making sense of big data: A view from the music streaming business

By Christian Harris, Managing Director UK and Ireland

May 26, 2016 | 4 min read

There’s no shortage of data in marketing today. In some ways, there’s too much of it. But the consequence of a data-rich environment puts greater emphasis on our ability to make use of and exploit data effectively. Capturing data that matters and using it to uncover trends is making our roles in marketing more effective and less complex.

music streaming big data

Humans have always used data collection and data storage. Over 20,000 years ago in the Belgian Congo, ancient humans of the Upper Paleolithic era used a primitive form of blockchain called tally sticks to record information to aid memory and record simple transactions of trade and payments. Primarily, they used the bones of dead animals to etch marks which represented these events. It was a simple, effective and practical method for recording data and making use of it over time. By around 300 BC things had moved considerably and the first formal data centre was conceived in Alexandria – a library, until the Romans destroyed it.

Since prehistoric times, humans have always had a need to record and analyse data in the same way that marketers do today. Yet, unlike our prehistoric ancestors, we have the advantage of utilising technology to compile and analyse complex data at rapid speed.

The same applies to the music industry. Think of the early days of collecting data to better understand the listening likes and dislikes of music lovers, from radio top 40 hits to album sales and tracking top charts, this has continuously evolved over time. We now live in a data-rich present, where real-time info drives a deeper understanding of the listening preferences and habits of our users. For example, at Deezer we offer Flow – our intuitive, personal music recommendation tool that is tailored to each individual listener. It’s a combination of both art and science, as we’re able to immediately and accurately determine a listener’s preferences through harnessing our own data and combining it with the research and hand-picked recommendations our fifty music editors.

Where can you see these two principles being applied in the world of marketing today?

Programmatic advertising relies on real-time bidding for inventory in order to secure the best price whereas brand campaigns require the human touch of copy, design and dedicated creativity to make a campaign work around a specific publisher. Neither is perfect. Programmatic can handle huge volumes assuming the conditions are right and there is both the supply and the demand, however, they tend to be of lower value. On the other hand, campaigns requiring human time and effort clearly cost more and require more planning, but result in a more immersive experience and therefore commands a higher value and can better address a specific marketing need.

The same is true in music streaming, there’s a huge volume of content out there but it only holds value if it’s managed around a person’s actual likes and dislikes. If not, it’s just a mass of content waiting to be solved. The two approaches of algorithmic recommendations and human curation applied in music today is paralleled perfectly by programmatic and bespoke advertising.

So, what does this mean for marketers?

From the two extremes of data use, starting with tally sticks and winding up all the way to blockchain-distributed databases, the use of data has been a constant need among the human race. It’s changed our behaviour and improved our understanding of the world around us. By its application to music and advertising, a new world has arisen in the form of programmatic and algorithmic automation. But the man-in-the-loop remains critical. There can never be a complete substitute for human curation in the same way that there can never been a complete substitute for marketing which requires the creativity and innovation that comes from marketing initiatives exploiting different channels in new and engaging ways.

Through data science, we can embrace better targeting and capitalise on the volumes available, making our trends more accurate and therefore more valid. Man plus machine is necessarily the answer, we just need our own Alan Turing to show us the way.

Christian Harris is managing director of Deezer, UK & Ireland

Technology Data Music Streaming

More from Technology

View all

Trending

Industry insights

View all
Add your own content +