The technology industry loves a buzzword, and big data is a favourite. But what does it mean, is it just marketing hype, and what services can it help retailers provide?


If you ask people to define the term ‘big data’, you’ll get different answers. This might be because the term is still relatively new. Big data has “gained widespread use in the last three or four years”, says Darren Vengroff, chief scientist at ecommerce personalisation company RichRelevance.

“It’s a buzzword rather than a well-defined concept,” says Michael Ross, director of ecommerce adviser eCommera. However, it boils down to retailers now being able to access more data than ever before, and it is about using that information to improve different areas of the business.

The emergence of big data is partly a consequence of the growth of online retail. Online shoppers generate large amounts of information about themselves and how they shop. Graham Cooke, founder and chief executive of technology solutions company QuBit, says big data has also resulted from cloud computing, which has made it cheaper to store huge volumes of data. “Cloud computing has given access to thousands of servers to process and store data. The computing power and cheaper storage means we now collect lots of data.” He adds: “Over time as it becomes cheaper and cheaper to do it, retailers have power to access information.”

So when the term big data is used, what type of information is being referred to? Jason Gordon, a customer analytics partner at Deloitte, says big data includes retailers’ traditional, internal data such as transaction details, as well as data from social media, data which has been proactively sourced to shed light on a certain area, and open data, which are data sets released by the Government.

While all this information could leave retailers feeling overwhelmed, Gordon says viewing big data in terms of volume is a mistake. It’s not a case of analysing as much data as possible. He warns against “analysis for the sake of analysis”, which can be “risky, costly, and often fruitless”.

“If people take it to mean ‘very large’ that can be quite costly,” says Gordon. “When we talk about big data we mean enhanced data not just large data.

Big data doesn’t automatically deliver big value. This is about building a unique data set, rather than a big data set. It needs to be high quality rather than big.”

Big data pioneers

“Google and Amazon were doing big data projects back in 2005/6,” says Cooke. “A lot of big data techniques used today were produced at Google.”

Today more retailers are getting in on the act. “Historically it’s been the preserve of the grocers, they have the most mature loyalty cards,” says Gordon.

“But some of the biggest changes are coming from the apparel side of retail. They’re starting to reap some really strong rewards from it.”

QuBit has worked with companies such as lingerie retailer Bravissimo using big data. By analysing data on sales transactions made via different devices it was able to draw some interesting conclusions. For example, customers accessing Bravissimo’s website on a smaller screen were less likely to make purchases. Cooke explains: “They were missing the size guide tool on a smaller screen. We looked at mouse movement, conversion rate, and found there was a relationship between the screen size and conversion rate. We made the size guide more prominent. It increased their sales by £2m a year.”

Qubit has also worked with, enabling it to gain greater insight into the customer purchase journey. “One of the biggest challenges in marketing is where you can attribute value,” says Cooke. While retailers often focus on the site visit where the customer makes a purchase, they might have visited four more times before that and these visits are worth analysing as well. Data analytics were applied to every visit and consequently the company increased its budget on generic key words after finding that certain terms had high returns on investment.

Personalising the experience

Big data can also be used to make targeted marketing a lot more effective. Gordon explains: “The question is not ‘I’ve got a voucher, who should I send it to?’ It’s ‘I’ve got a customer, what can I put together that is most compelling for them?’ It’s about having enough understanding of a customer for that different way of targeting. Use other sets of data to fill in gaps in your knowledge. Nothing turns a customer off like receiving irrelevant information.”

It can also give a clearer picture of how a business is operating across all its channels. For example, it can be used to evaluate whether an online marketing campaign has been successful in increasing store sales, says Vengroff, who previously worked at Amazon. “You might promote it [the campaign] online but people want to see the product in the shop before they purchase. It’s bringing all the data together in one place. When you’re deciding which of 10 campaigns you should run next you can make an informed decision.”

This ability to make more informed decisions is at the heart of big data.

Ross says: “For retailers there are lots of decisions to make which require data, it’s about making better decisions. There are also more micro decisions to make.” Big data offers retailers greater insight into their customers’ shopping behaviour, and enables them to measure and evaluate their business in ways they couldn’t previously. In the introduction to McKinsey’s report, Big Data: The Next Frontier for Innovation, Competition and Productivity, published last year, researchers state: “We estimate that a retailer using big data to the full has the potential to increase its operating margin by more than 60%.”

Tempting as it might be to delve straight in, Gordon has a word of warning: “For those businesses that aren’t already doing this, my advice is start with something sensible and contained where value can be proven.” Retailers using big data need to have a clear idea of what they want from it - the bestapproach is to start with a commercial issue and move forward from that point.

Multichannel makes the most of big data

Crew Clothing

Crew Clothing

Frank Sendler, online marketing manager at Crew Clothing, said the retailer started using big data techniques about a year and a half ago. “As a multichannel business we want as much insight as possible into how customers behave for our marketing strategy,” says Sendler.

Crew Clothing began by using a tool built in-house and Google Analytics. Later, it started working with QuBit.

“When customers order over the telephone, or via the website, the challenge is to put them together, to match them up and make the most sense out of that data,” he says.

“Customer data includes what people are buying, when, from what promotion, in what context. Collecting data is a long process.”

Sendler says Crew wanted to find out how shoppers who received its catalogues were purchasing products. Was it over the phone, in store, or via the website?

The retailer analysed the data and was surprised by the results.

“Many of our customers simply take a catalogue into the store and buy the product in the store,” says Sendler. “It’s useful because we know with the catalogues we can send them out and they will still convert. We have a much better understanding of the context between stores and the catalogue.

The customer is really just swapping between the channels.”

Sendler says big data techniques are particularly useful for multichannel retailers: “It gives clarity and insight for strategy,” he explains.