As companies gather massive amounts of data, making sense of it all and turning it into competitive advantage is all-important – and that’s where data scientists come in
As Albert Einstein famously said: “Information is not knowledge” – and he knew a thing or two. You can have all the information at your disposal, but if you don’t know how to interpret it, it’s useless. And that’s probably never been more true than in this fast-moving technological age.
Smart cards, smart phones, smart cars, smart washing machines and a whole host of sensors, allow us to analyse a vast amount of data in a way we’ve never done before. This stack of information that companies or organisations collect is usually referred to as ‘big data’.
However, deriving meaning from big data – transforming information into knowledge – is what’s crucial. And it’s creating a booming, lucrative industry for skilled ‘data scientists’.
As Aonghus Fraser, Group Chief Technology Officer at C5 Alliance, explains, while the term data science is new, the general concept is not. “In the past, the term was ‘data mining’ or ‘machine learning’ – the principle is the same, but the difference now is the sheer volume of data that companies are working with.”
Transformation potential
The potential for transformation using insights from data analytics spans every industry – from healthcare, retail and banking to transportation, manufacturing, policing and leisure. Analysis of data sets can unearth new correlations that can help identify business trends, prevent diseases and even fight crime.
Early adopters of data science have already been able to reshape their business either through new products or an overhaul of customer delivery and engagement. At a very simple level, rather than bombard a customer with multi-product recommendations, a bank may focus on a home loan recommendation based on increased recent customer transactions at B&Q.
However, not all companies are up to speed with what data science brings to the business equation.
Essentially, a data scientist combines analytic, machine learning, data mining and statistical skills with experience in algorithms and coding. This information is complex and difficult to grasp for the lay person, so the data scientist is tasked with having to explain the significance of specific data in a way that can be easily understood.
As Fraser points out, some companies have looked to third parties to make more sense of collected data. “The Netflix Prize is a good example,” he says. “The company offered $1 million to the firm that could improve the algorithms for the business and ultimately see more clearly what was happening in their consumer base.”
In 2009, the prize was awarded to BellKor's Pragmatic Chaos team, which improved Netflix's own algorithm for predicting film user ratings by just over 10 per cent.
If data science is primarily about ‘big data’, does that mean data scientists are only worth employing at the biggest companies? Not so, according to Claudia Imhoff, President and Founder of Intelligent Solutions.
“Big data to me is not just the volume and variety of the data, but the complexity. Small companies can have very complex sets of data and can ask very complex questions. So, if you’re a small company with complex questions and complex data, you need a data scientist.”
Whatever the size of company, there’s frequently a lack of understanding at boardroom level of what data scientists provide.
Malcolm Williamson, Commercial & Enterprise Director at business support specialist Exemplas, agrees with this view. He suggests that the effective organisation of colossal levels of data into something that’s coherent, focused and directly beneficial to the business is a challenge and one that many firms shy away from. But this is where data scientists prove their value, he insists.
“If you’re dealing with enormous volumes of data, effective analysis is likely to be significantly more valuable than any product or service being made,” he says. “I don’t have a problem with paying people large sums, so long as they demonstrate a return on investment. I suspect the best data scientists repay their costs 10 or 20-fold.”
Wealth of information
Despite the abundance of data, the ‘scientists’ making sense of it all are a relatively new breed. They’re able to demand large salaries because expertise in this area isn’t widespread – but as Graham Coultas, Innovation and Programme Director at Exemplas, points out, their god-like status is unlikely to last long term, as salaries level off once supply increases.
In terms of levels of data available for analysis, Coultas highlights the ever-increasing sophistication of information gathering. He points to a supermarket chain in the US that uses face recognition software to know when someone has visited the store, then logs where they go in the store and how long they spend in each aisle.
Wearable technology has also proved something of a game-changer. Activity bracelets can provide up-to-the-minute data on anything from blood pressure, heart rate and body fat, to diet and exercise. There are also apps enabling people to take an instant ECG on their phone or to scan a meal to calculate and record calorie and nutrition levels. This is of interest to healthcare and insurance companies in terms of improving claims management and reducing costs.
Data from mobiles can also provide crucial data for analysts. Using wireless technology linked to consumers’ smartphones, retailers can now track shopping patterns in the same way that online retailers have done for years. The data can help make informed decisions on staffing levels, what fashion lines to carry more of and how to maximise store displays.
However, finding patterns and significance in recorded data isn’t always as straightforward as this retail sector example. Nor is it always easy to make accurate predictions. As Coultas explains: “Even the best brains can analyse a raft of data, but can they translate that into what future trends may be?”
This is where the skill factor comes in – the nuggets of useful data are pinpointed and strategic decisions are made that, in theory, should benefit the business. “You’re never going to have 100 per cent certainty, but using sophisticated calculations of probability, you’re able to make the best of your data,” says Aonghus Fraser.
The more often data scientists can demonstrate their value to a business, the more they will be viewed as a necessity rather than an option. However, a Digital Masters roundtable earlier this year (organised by digital networking and recruitment specialist The Up Group), suggested data scientists needed to be integrated more into companies to fully utilise their skills and showcase their effectiveness.
There was general consensus at the meeting – which included representatives from Accenture, ASOS, Barclays, Burberry, Camelot Group, eBay, Experian, Facebook, lastminute.com and Prudential – that data scientists working alongside product managers would provide a clearer understanding of each other’s roles. Farming and interpreting meaningful data would prove far less taxing if these two skill sets communicated more closely.
However, it was noted that this would inevitably prove more challenging for less new-age businesses and those with legacy issues as a result of M&A activity.
Recruitment challenge
Undoubtedly, data analysis can help improve operational performance and, where necessary, help transform a business and ensure it remains relevant. But not all businesses have the capabilities or the understanding to fully embrace data science – yet. Recruitment is another challenge too.
Demand for experienced data scientists comfortably outstrips supply. To combat this, one delegate at the Digital Masters roundtable suggested bypassing what can often be a drawn-out search and application process. He proposed that if an impressive CV were received, a direct call to the applicant (not via HR) should be made immediately. And rather than bombard a potential recruit with 10 pages of application forms, the first step should be a test of technical aptitude. Only after this should the standard boilerplate process come into play.
While the benefit of big data analysis are clear, there are also privacy and security issues. As data security specialist Darren Harmer explains: “Data is collected every day on where people go, who they communicate with, what they eat, what they read, what TV programmes they watch, what physical exercise they do and even their sleep patterns. Their lives are analysed in ways inconceivable a decade ago.”
Undoubtedly, big data allows companies to make objective decisions – insurers responding positively to healthy lifestyle statistics, for instance – but what if data files on individuals are inaccurate or the algorithms are questionable? And what if big data in the hands of banks, insurers and employers has an impact on someone getting a job or qualifying for a loan?
Does the individual have any powers of redress with regard to limits on the personal data that companies collect and retain? “Yes,” says Harmer. “You do have a say in who you share information with and for what reason. But this means consumers reading privacy policies and terms of service agreements, which
in reality seldom happens.”
We live in a brave new world, where data is king, and the companies that make most sense of it are arguably the ones that are going to succeed well into the future. Having so much personal information floating around in cyberspace may make some people uncomfortable, but the harsh reality is that it’s nigh on impossible to put the genie back into the bottle.