We recently attended Innovation Enterprise’s “Big Data and Predictive Analytics” summits in London. While the various conferences remained instructive, much time was given to discussing the concept of “Big Data”; how can it be defined? Is it just a buzz word? Is it actually beneficial or are companies feeling the need to jump on the bandwagon? It seems the jury is still out on the first two questions and although critics describe it as more, messy and muddled, the overriding opinion is that if used correctly, big data can be extremely useful for your business’s bottom line and for your customers.
The definition of big data is something that we have discussed before but to keep it simple, Sophie Crosby, Vice President of Insight at Ticketmaster, used the four Vs- big data involves volume, variety, velocity and veracity. During the talks it was clear that Hadoop, a framework for storage and processing was the favoured provider, but what really matters is using the right data and finding a constructive means to analyse it. If you don’t know which data is relevant, collect it all in “the cloud” until you decide the right questions to ask. This tricky but more interesting part involves finding the key indicators to analyse and using the results to competitive advantage; it’s great having all that data but without analysis it does nothing except anger privacy rights activists. This is where predictive analytics comes in.
Predictive analytics essentially allows people to do their job faster, cheaper and with greater impact. With Hadoop setting a record in sorting 1.65 terabytes in one minute (I certainly left the summit a Hadoop fan) there are few limitations on what can be done. Chris Gobby from EE discussed how the company was able to collect huge amounts of data from multiple sources and successfully harness and monetise it. As an example he showed how EE could track the activity of commuters in the morning around Waterloo station. They record who uses the Internet on their phone, broken down by age and gender, which sites they visit and for how long. This is legal because EE use it for improving their own service but they are also able to sell it to companies looking to use data to help influence their KPIs, find out who their customer base is or for targeted marketing. For instance, firms have been advertising online shopping near to where they know lots of people shop on their phone. This throws up all sorts of privacy issues but it does show how big data is proving useful.
Another of the more interesting talks, Matthew Eric Bassett, Director of Data Science at NBC Universal described how a “certain” motion pictures company was using analytics to reduce risk, by predicting the most profitable dates to release movies. This company made a model focused around dates and the likelihood of filmgoers to choose the cinema over anything else, regardless of the quality of the film. It looks at releasing a movie on X date with Y demographic target audience and the possible results. You might consider this capitalist exploitation, when an individual can almost be tricked into watching a terrible film that they would not have otherwise paid to see, simply because of the day it was released. Regardless, the model more often than not is proven to work, and the majority of consumers remain oblivious or don’t care.
While analytics may have a few teething problems and raise privacy concerns, it’s clear that it won’t be going away any time soon and those not yet on the bandwagon should feel increasing pressure to “jump on”.