Big data is big business. While it falls foul of being seen as the 'buzzword' of the moment - we all want to claim to know it, sell it, gather it, develop it or profit from it - it is in fact one of the top priorities for businesses right now. Data analysts are in demand like never before and we have to develop an understanding, quickly, of what we actually mean when we say big data and most importantly, where the opportunities lie.
To find out a bit more about this growing business application, we held a roundtable in London with representatives across the industry, including finance, gaming (I am Playr/We R Interactive), music (EMI), platforms (Yahoo!), agencies (Vizeum) and analysts.
Big Data Begins At Home
While we're all attracted to the allure of social data (bits of information originating from our personal social network profiles), an interesting thread to the discussion was the use of data internally. Big data in this respect is nothing new. Established data collection activities such as surveys or email questionnaires are providing businesses with data every day and social doesn't necessarily affect this in any way. When we talk about big data, it's easy to forget that this isn't actually a new concept, but that there are now new opportunities with layering social data on top of data that is generated, which you own.
The music industry is collecting a huge amount of data from surveys. Over the past three and a half years, EMI has interviewed over one million people and at any one time, there are twelve people taking part in a survey for EMI. Now when you combine this with the rise of digital music downloads and streaming, you suddenly have a very exciting data set to play with, but it is easy to overlook consumer research.
Access to music online is only going up. While sales in the music market overall are going down (at -7.5% in 2012), the inverse is true of digital sales which grew 11% in the same year. With this, of course, comes a rise in data that we can access about people's music interests: genres they like, average listen times, singles vs albums, etc. The challenge for labels now is how (if at all) this public, unprompted data can be combined with data that's gathered by the labels, and use it to spot trends and make key business decisions. However, there is a risk that big data does not tell the whole story and focuses on a subset of activity.
Man And Machine
The proliferation of data has led to a trend in personalisation, which is very different to customisation. Whereas with various news sites you can select your interests to filter the information you're shown, with personalisation this work is done for you. This is fueled by the data that you've 'shared' (though not always wittingly), which contributes to your online DNA. Land on a site and you will see that information which has been deemed most relevant to you, based on your social profiles or browsing activity or other signals.
With this comes a risk of your access to information becoming siloed; you will potentially begin to see more of the information that an algorithm has determined is relevant to you and less 'random' information that falls outside your preferences. The process of scanning through a paper and coming across random articles that are outside of your immediate interests is at risk of becoming obsolete. So what companies like Yahoo! is doing, is applying is a mix of man plus machine to create personalised experiences online. There is still a large team of writers and editors that determine which articles should be shown on news sites and sometimes what you see may fall out of your immediate interests.
The effects of this personlisation trend can be seen in Yahoo's data visualization tool, where it publicly makes available details of the 13,000,000 combinations of 400,000 news stories published on Yahoo's networks daily, worldwide. Here you can filter down to see which stories are being viewed by certain demographics:
Or by interest:
What we need to understand as media entities become more efficient is when to let the 'man' or 'machine' lead. Up until now, the man has led, and there have been a few players in the media world that have determined what information we should access and what is the most important (e.g: we all read the same paper when we buy it from a shop).
That concept no longer applies as personalization is taking over, but we have yet to understand the long-term repercussions of this. What does it mean when I read one version of a story while a friend reads another version (where different aspects will feature in the same story), or when the top headlines I flick through are completely different to the person sitting next to me? This is all possible through the rise of big data, but we are too early in the process to understand what this really means. One thing is for sure: It has fundamentally changed access to information.
A near-perfect example of man versus machine can be seen in We Are Hunted: a service that fundamentally runs on data, but with a layer of editorial control. Music blogs are scanned to find the tracks that most people are talking about, then presents these in a top 100 playlist on Spotify. Everyone sees the same playlist and it is no way customized to you, but it runs on the concept of big data: That which is being shared among the 'man' (in this case music bloggers).
Big Data & B2B
Typically when we talk about big data, we are focused on the impact for B2C. This is where the flow of data is most concentrated, but what opportunities are there for businesses who want to harness big data for targeting other businesses? You have an immediate problem in that there is a smaller data set to play with, and the decisions made off the back of this data are likely to be higher-impact; the data has to be trusted and social networks might not necessarily be relevant for this.
Thomas Oriol from Sales Clic talked through how big data can be applied to the B2B process for SMEs and made the prominent point that the analysis of big data shouldn't 'make you think'. And that more importantly, 'bigger' data shouldn't mean bigger and more extensive reports, but rather the opposite. The key for big data in the B2B process is to know what sources of data you have available, whether these are public or generated through tools you use.
This is a fundamental learning for any application of big data. Rather than trying to gather data from every available source, think about those that are actually relevant to you, and while these may be more niche for SMEs, the same data analysis process should apply.
Is It Creepy?
A common thread in the discussions was a concern or awareness of the potential for big data to become creepy. This depends on what is done with it, rather than how it is gathered, as most people don't even realise the data they're freely giving away. But there is a very real risk that we personalize too much, or 'accidentally' let it slip to the consumer just how much we can - or do - know about them. Many people are frustrated with seeing singles ads on Facebook, or ads for slimming down, yet they genuinely don't realize that they're seeing these particular ads because of demographic or interest data they've given up to Facebook.
Regardless of how little truth it bears to the reality, we like the feeling that we are using these social networks for free without having to give over anything, or that we can read news without paying for it and it's suddenly become a free commodity. We sort of don't think about the economical impact of this and that now media players or social platforms are profiting of us in a different way: With our data. If companies play on this too much, the realization will hit the wider consumer and we will likely adapt our behavior or become suspicious and stop using certain sites.
How you apply big data is ultimately down to you, whether you have gathered it yourself or are aware of how to access data from other services such as Facebook or Twitter. But it is a process where there are likely mistakes to be made.
Big Data For Good
Where it gets really, really interesting is when you look at the impact of big data for the good of wider society. The U.N. are showing this in action with its data exercise Global Pulse. Here it's gathering public data through social tools as well as calling on businesses to make data available so they can examine key trends which until now would have been largely hidden. Robert Kirkpatrick talks through the virtues of the experiment here and how they are beginning to use the analysis of big data to impact people's everyday lives.
comments powered by