Wednesday, September 4, 2013

How Big Data Is Transforming Mobile

Five quintillion bytes of data are created every two days; that’s more information than has been generated by every conversation humans have ever had. The question is, what do we do with it: how to get context and value from the three “v”s of Big Data: volume, variety and velocity?

The data

Big Data is nothing new, in fact one of the first uses of the World Wide Web was to share the enormous volume of data being produced in CERN’s LEP particle detectors so it could be analysed, back in 1989. We are all producing data, all the time when we use computers: every keystroke, mouse click, every app used, webpage visited and so on. However once our PCs became connected to the internet this data could be collected and analysed, and as we used the internet for more of our computing needs collecting this data became far easier.

Mobile is a game-changer in that if Big Data is fast, big, and varied, mobile data is faster, bigger and more granular. Why is this? We create data every time we use a device, or an app, or a service.

Bigger makes sense: we have more mobile devices than PCs, and we are more likely to use multiple apps as each app represents a cut-down process. Mobile devices are also permanently connected, and are more transactional: they make greater use of services to send data between apps, or to refresh an app from the server, and the extra location data is a stream that isn’t so heavily used on PCs.

Faster is where mobile data starts to get interesting: the small screen real-estate on a mobile device means that everything which is displayed must be as relevant to the viewer as possible. This means that screens must be able to respond to vast amounts of changes, such as past activity and preference, location, up to date information from the server and so on in order to maximise the relevance of what is on the screen. Add in the shorter attention span (or rather, frustration span) on mobile: the amount of time a user will put up with a screen before saying, “forget it, I’ll look later on my PC”. If the user is a customer browsing retail websites in a nearby cafe, that less-relevant screen could cost you a sale.

Finally, and extending from bigger and faster, mobile data is more granular because the bigger data, transmitter faster, between more services, allows you to build up a better picture of your user.

What can it do?

The possibilities raised by mobile data are incredibly varied, but they essentially stem from two directions: optimise, or advertise. Those in turn are merely two sides of one coin, which is the ability to personalise content.

A perfect example of personalised data is the “quantified self”, a movement rapidly gaining adherents which uses mobile data to provide metrics that help users improve whatever they care about. This may sound like the sort of thing only of interest to especially geeky people, but consider if the app measures time taken to commute or get somewhere you want to go. Imagine the app takes into consideration not only historical data about your movements (which would allow it to estimate your travel speed if walking, or your preference for taking a longer, sedate route over a faster but more stressful one, for example), but also live and historical data from other people’s devices indicating traffic (road or pedestrian), and tie in mapping to let you know when you have to leave… the result would be that you get less stressed and can spend more time with your friends and family.

Equally, Big Data services are getting so good at identifying people that a shopping website might be able to take a good guess at who you are even when you’re looking at it from a different device: or rather, it may not know who you are but your current searches and preferences (even context-based ones like what angle you hold the device at) would allow the site to narrow down its users and display things you are interested in.

This gets even more interesting when you consider that the second point of mobile data is “faster”. Services are becoming able to respond to your inputs and inferred preferences in real time, to provide you with predictive search to reduce the amount of input you need to provide on a small screen, possibly in a moving vehicle. Reducing the amount of input and switching of devices can only increase customer satisfaction.

The challenges

While there are many technical challenges, like ensuring we have enough bandwidth to actually deliver this future, or training business leaders and consumers alike to set effective metrics and make smart decisions based on their data; there is one challenge so huge that it is the Big Data elephant in the room.How do we avoid this personalisation becoming creepy?

This is actually two problems in one, which sit along a spectrum. On one end are targeted ads like inMinority Report, which show differently depending on the person viewing them; and at the other end of the spectrum sits the “targeted” ad that gets it so wrong that the brand-connected experience that the ad is trying to create vanishes instantly.

The latter is probably the easier to overcome, as it is a symptom of poorly created metrics or data that isn’t quite big enough to provide a personalised experience that is up to scratch. While this isn’t a problem now, such ads will become increasingly obvious in the near future.

The other end of the spectrum is more troubling, because while an insufficiently targeted ad might make someone laugh and feel vaguely superior for being hard to predict (who doesn’t love that feeling now and then?), one that is targeted beyond our comfort zone can make the viewer look over their shoulder, worried about what else someone knows about them. There’s a reason this kind of advertising crops up in dystopian science-fiction.

Will Big Data transform mobile, or will mobile transform Big Data?

There is no doubt that the world is about to enter a new computing revolution in which we are going to find far more personalised, interactive experiences. But there are three key human fears that will need to be carefully negotiated in order for this to realise its potential: the “uncanny valley”, Big Brother (Orwell, not the TV show), and our own, personal expectations.

The “uncanny valley” is a concept which holds that humans feel more attracted to things as they approach human likeness, but that there is a very sharp drop at around the “90% human-like” mark, which is the valley. It’s essentially a part of the brain saying that there is something very wrong, and it is believed to be an evolutionary trait which protected us from danger and illnesses, and which was reinforced by social living which promoted a strong set of norms. In terms of Big Data, if your phone starts acting too much like a human intelligence, it’s going to get creepy. In terms of what to aim for, teddy bears and the robots from Star Wars are generally held to have hit the pre-valley peak, and the attraction-revulsion reaction is greatly amplified when the object is moving. Big Brother needs very little expectation: the control device in 1984 has become the seminal example of a scary future. Finally,
The Uncanny Valley, from Wikipedia
expectations: what level of personalisation will be deemed acceptable will vary greatly between people, based not only on demographics but also experience of such interaction, familiarity with the service and their level of connection.

Essentially, services, apps and brands will have to follow the real-life process of “becoming friends”and not overstep the boundaries if they want to succeed. Will Big Data transform mobile? Mobile data is currently the biggest around, and that will transform how we think of data. As the author Terry Pratchett said, “We are proud that we live in the Information Age. We do, and that’s the trouble. If we ever get to the Meaning Age, we’ll finally understand where we went wrong”.

David Akka is the Managing Director of Magic Software (UK) Ltd and have worked with Magic Software for over 14 years.
Visit his blog at: