When David Stephenson takes to the stage at the 5th Big Data Insight Group Forum in two months' time he will explain how eBay uses big data analytics to ‘create new insight and improve customer experience’. As the head of business analytics for eBay’s Classifieds Portfolio (eCG) he is extremely well placed to shed light on just how important data is for the company.
First and foremost, it almost goes without saying that eBay has masses of data at its disposal. “We’re collecting probably 50-100 TB of data every day,” Stephenson says. “And we’re running over 100 PB per day through our data systems.”
As this suggests, the company is not purely concerned with data on each of its individual customers but rather on the wider picture of understanding the preferences and habits of the online shopper. As Stephenson states: “For us it’s not so much about collecting personal user data as it is about collecting general online consumer behavior. We can use that data to make the online shopping experience as efficient and productive as possible for both buyers and sellers.”
Unsurprisingly – despite the fact that, as we’ll see, there is a vast array of benefits to be realised across the board from big data analytics at eBay – it is the way that it enhances the company’s knowledge of customer behaviour that represents the most prominent and valuable use.
“Internet data provides a wealth of information,” Stephenson states. “Ultimately it reveals the decision processes that the customers followed while buying or selling with us: what brought them to us, what questions they asked, what items they compared, and whether or not this was converted into successful buying or selling experiences. This is hugely important and it brings us back full circle to a two-way, personal shopping experience from which everyone benefits.”
However, analytics does not solely play a role in delivering consumer insights; it is “important in pretty much every area of the business.” Beyond ensuring that the exchanging of products is as painless as possible, eBay also uses its data to improve: security, traffic acquisition, onsite search efficiency, sale conversion rate – and that’s just some of the more prevalent ones. Indeed, Stephenson asserts, because eBay is fundamentally a data-driven company, almost all of its innovations are underpinned by analytics. “What’s perhaps most relevant to the Internet industry,” he adds, “is that eBay is also constantly analysing A/B split tests, so that we can very quickly determine which innovations work better than others.”
Of course eBay is used to having to deal with data, and lots of it. But Stephenson explains how advances in technology have made it more of a valuable asset to the company. He says: “The newer systems allow us to store greater amounts of data while also being able to process the semi-structured data in real-time which empowers very detailed analysis. This, in turn, supports very specific segmentation and allows us to mine answers to very pointed questions. What’s more, keeping the multiple petabytes of data rather than discarding them allows us to retrospectively ask new questions of our historic data to help us improve moving forwards.”
Despite its success, Stephenson is keen to stress that everything eBay does is geared towards continuous improvement and big data plays an important part in this. “There is a constant effort,” he says, “to position analytics within the company in a way that makes the most sense, both in terms of business focus and geographic focus. In addition, technologies are changing, and we continue to work these into our analytic processes so that we’re always improving what we do.”
As for the challenges that big data presents to eBay, Stephenson states that they fall into three main categories: consistency, structuring and real-time accessibility. This means that the huge volumes of data the company is collecting all of the time must be structured in some way – no mean feat when you consider the range of forms it comes in, from web metrics monitoring online activity through to consumer search trends or purchase information. This data then needs to be put in the right place so that it can be extracted and analysed, with the insights from it then fed to the relevant people. Crucially, this all has to be done within acceptable time frames.
Stephenson’s advice for anyone starting out on their big data journey is to have a look at Bill Franks' book, ‘Taming the Big Data Tidal Wave’, which will help those new to the technologies get a “foundational understanding of the issues involved”.
For more information about how you can claim your free place at the 5th Big Data Insight Group Forum, where you’ll hear David Stephenson’s case study presentation, revealing far more in-depth details about the great work eBay is doing with big data, please click here.