Big data – loosely defined as high volume, high velocity, complex data that usually requires highly specialised skills and technology to process – has come along way since it came to prominence as a concept in 2012.
Almost three-quarters of European organisations expect it to produce a return on investment (ROI) within 12 months, according to research from Xerox in May 2015. On the flip side of that coin, Gartner claimed last August that on its Hype Cycle for Emerging Technologies, big data has now passed the ‘peak of inflated expectation’ and entered the ‘trough of disillusionment’.
So now that big data has matured does enthusiasm or disenchantment prevail, and is a certain cynicism around big data healthier than unbounded optimism?
“People can throw a lot of money at something they feel is cool or new, and that they need to be involved in, without really being clear on the business problems they are trying to solve,” says Andy Day, business intelligence director at News UK. In the recent past, big data has been one such area, a term bandied around with excitement yet which often lacks a clear universal definition.
Day says part of the problem is the term ‘big data’ itself: “I am concerned with data that I can use to solve business problems, whether it is digital or analogue or people filling in surveys or clicking on a button on an app – I don’t really care, it is all data – and I think the words ‘big data’ only fuel hype.”
He is not alone in being dismissive of the buzzwords. Nic Wenn, chief marketing officer at cashback site Quidco, which uses shopper behavioural data to personalise the customer experience and tailor offers, says it is an amorphous term adopted by companies as they try to utilise more data.
“It doesn’t specifically determine how data is being used,” he argues. Instead, Wenn says Quidco refers to “fast data”.
“We identify key data points that we are able to utilise in real time that will improve our customer experience and user experience. Our strategy is based on how we can use all the data that we have available.”
Orlando Machado, head of customer science at price comparison site MoneySuperMarket, which uses Oracle’s BlueKai data management platform as part of its technology stack, is not a fan of the term either. “We don’t differentiate between big or small data. We just concentrate on trying to get the most impactful insight, and it doesn’t matter whether we are using a few thousand records or millions of records.”
Input not output
High volumes of complex data are being generated by almost every business as customer journeys have become increasingly complex – travelling across multiple devices, often combining online and offline touchpoints. But it is how brands use this data that divides them into either advocates or sceptics.
Day at News UK says that his approach is to identify the business questions that need answering first. “That is absolutely critical. Otherwise you can spend £10m or £100m on a big data set of capabilities such as people and technology, yet deliver absolutely bugger all because you haven’t worked out exactly what problem you are trying to solve.”
Royal Mail is another business that generates a lot of data, with one of the largest tracked delivery networks in the UK and millions of online shoppers. In the last 12 months the organisation has rolled out a CRM programme to its business customer base, applying big data to enable personalisation based on factors such as company size, sector and product usage.
Head of marketing Ben Rhodes agrees with Day’s assertion that working out what outcomes you need first is key to success. “Data is incredibly important but it is an input – not an output in itself. We work through our overall business strategy before we make technology investments.”
But despite the caveats, investments are now paying off for some brands. MoneySupermarket uses data chiefly for CRM and performance marketing purposes, as well as to underpin products for its commercial partners, but big data has enabled the company to take its CRM from standard customer database marketing to the next level.
“For example, some people read emails during weekday evenings, others in the mornings at weekends, so we can send out emails that reflect those consumption patterns,” says Machado. “We wouldn’t have had the data or execution capabilities to personalise like that before.”
He says that revenue driven by big data investments is now a “significant contributor” to the Group’s overall revenue.
Quidco is also using customer data to enable greater personalisation, making predictions about what types of offers will encourage individual users to make further purchases based on preferences and habits.
“We work with close to 4,300 retailers and have over 20,000 offers on the site at any given time, and our task is to personalise this offering to over 5 million customers, all with differing behaviour, frequency, usage and user habits,” says Wenn.
Testament to Quidco’s belief in big data is the fact that it invests over 25% of its net profits annually back into its data and insight capability, including data analysts and technology. Wenn estimates that the company will see a “significant return” on its investment in the area as a result of higher customer engagement and higher gross advertising commissions resulting from more relevant, personalised deal content.
Rhodes is confident of a return on Royal Mail’s investment in big data, which has included appointing a team of data scientists during the last 24 months, but he says the organisation looks at a three- to five-year window. “Judging something on a one year payback doesn’t really fit with an organisation of our size and scale.”
News UK is also moving in the right direction. “I won’t pretend that we have gone from a standing start to paying back all the investments we have made in technology and people, but we are well on the way,” says Day.
The company is applying big data across numerous areas of the business, and some of the biggest progress the company has made in utilising big data has been in simply reporting findings in a more digestible format, “for example, findings such as how many people opened the app yesterday, and which story drove the most click-throughs to our subscription sales site”, he adds.
“We also have dashboards running on the editorial floor which have metrics such as ‘who was the most popular journalist yesterday’. While that data comes from a big data stream, it is stuff that is relatively simple as long as you know what business problem you’re trying to answer.”
The need for geeks
It was this belief that prompted Day to hire data journalists into his analytics team, tasked with translating data into meaningful information. “The people that code stuff, or analysts, are generally good at speaking to computers but [not good] at speaking to people,” says Day.
“The journalists I have hired don’t have an analytics background but they are good at assimilating facts and figures and turning them into stories.” It is this team that suggested using dashboarding to inspire the editorial staff. “It is about using complex analytics to tell the story in a very simple way.”
His point highlights one of the biggest challenges facing brands trying to harness big data – finding skilled practitioners. Machado faces the same problem at MoneySupermarket: “It is a rare skillset – you want people who have technical skills combined with commercial awareness, and also the communication skills to enable them to work with other people across the business. It is an unusual combination.”
As Day adds: “People in roles like mine have to be half data geek, half businessperson.”
Big data is often taken to mean the absorption of unstructured data, but I tend to see it simply as a large and ever growing volume of information that, while having increasing power and usefulness, simultaneously becomes more complex to manage.
Insurance is a very data-heavy sector and given its actuarial underpinning, it has, in effect, only ever existed in a world of big data. Consequently, having created the direct insurance market in 1985 [by offering policies without going through brokers], we have a wealth of data available to us. This is a key source of competitive advantage as it allows us to understand our customers’ needs today, and to anticipate customers’ future expectations.
Our motor telematics proposition provides us with an incredible amount of customer journey insight. The challenge and opportunity, however, is to use it in a way that improves our offering and ensures that our customers continue to be front and centre within our business.
We have recently started to build our data science team to explore how we can acquire, augment and utilise data. The reason we are investing in our data infrastructure is because we believe it will enable us to secure an edge over our competitors in our mission to make insurance easier and better value for customers.
There has been a lot of hype [around big data] but, like anything in life, you get out what you put in. Ambition needs to be in equilibrium with commitment; otherwise gaining a competitive advantage using big data is unrealistic and merely a hope rather than a strategy. Ultimately, you have to put the correct investment in infrastructure and technology in place to ensure that the data you have is materially better than your competitors, and that it is a cost-effective exercise.
How big data powers sports betting
In 2014, BoyleSports, the largest independent bookmaker in Ireland, was looking to modernise its digital customer experience. The company has over 200 shops but knew that there was also potential to grow its small online presence by applying greater understanding of user behaviour, and by using real-time data to inform decisions.
“At BoyleSports, personalisation has always been the long term goal,” says Tom Baker, head of acquisition and conversion optimisation. “As a sportsbook, we’re blessed with a huge amount of data. Whether that’s customer-specific or the array of betting choices – the events and markets that people can bet on – that we offer to our customers. That opens up almost endless opportunities to customise.”
The company partnered with digital marketing technology specialist Qubit with a view to improving data collection and analytics. Work began with the implementation of the ‘Universal Variable’ (UV) approach, which enables more granular data collection than alternatives such as ‘page scraping’ or cookie methods.
“We’ve been able to use the UV to take advantage of the vast array of betting events and data points that are available at any one time to deliver tests that get to the heart of what really drives business performance; namely event-based personalisation, cross-sell into other product verticals, and increasing margin on our sportsbook.”
The data revealed an opportunity to persuade users betting on multiple single bets to bet on accumulators instead, where punters can win bigger payouts if all the bets are successful. When users selected two or more singles bets, they were told the additional amount they could win, which resulted in a 6% uplift in accumulator bets. In addition, BoyleSports was able to segment users by session and location, targeting them with a location-specific welcome message offering £50 for signing up.
This year, the company has seen a 5% uplift in new account creation, and new visits converting to registrations have increased by over 6%. The results of the work have led BoyleSports to develop a dedicated optimisation team to drive digital customer experience.