CAIRO: “‘Big data' is less about the size of the data and more about the diversity,” said Stephen Brobst, chief technology officer for Teradata Corporation. In and of itself, he added, big data is nothing new and “is not a meaningful term,” since it is relative to the data used by each company and also compared to what it was before. Having said that, data is indeed “getting bigger as we evolve from analyzing transactions to analyzing interactions,” and this is where businesses can really understand customer behavior and experience. This kind of data from clickstream and search stream is, well — huge, he added. Clickstreams are useful in analyzing web activity, market research as well as employee productivity. At a conference in Barcelona this past April, Teradata — which offers database technology, analytic solutions and consulting — brought together executives, partners and clients from around the world to showcase how data is used to drive growth. Citing e-Bay, one of Teradata's clients, as an example, Brobst said the size of its interaction data is 20 times that of its transactional data, and this will continue growing exponentially. “It's not just the dotcoms. It's also the telecos, the evolution from analyzing call detail records to analyzing the interactions with the network is an example of going from transaction [to interaction data],” he explained. Access to information When speaking of the evolution and growth of data, usually it's a reference to “new media sources, such as social media data,” Brobst said. Asked by Daily News Egypt about the privacy concerns of social network users, and how businesses draw the line between mining information and being intrusive, he advised users to be careful what they share. “My view is you shouldn't put anything on social networks that you wouldn't put on the largest billboard right outside your house,” he said. “It's the consumers' choice what they make available on social networks or public forums, and there's certainly awareness, and I think sometimes people don't understand that what they put on some of the social networks — how accessible it really is. “…At some level it's easier not to have a Facebook account if it's not something that you want to put [out there]…so be aware. On the other hand, it's a very powerful tool to do something like brand monitoring. “If I introduce a new product or service, I can get immediate feedback that would've taken months through focus groups and research, I can readjust my pricing, market positioning immediately with that data.” Panelists in a session at the conference noted how some companies view setting strict privacy policies when it comes to user information as a competitive advantage. “Most commercial organizations are not using the blogsphere to target individuals with marketing [promos], [still] there's some of that,” he said. He cited CocaCola as an example. The company's Facebook page, which was not even created by them, has tens of millions of fans. “They do not intervene or use it to deliver offers, they let the fans run it. There's lots of good analytics that can be done with that, but you have to be careful when you start targeting customers,” he said. Too much? “It's the diversity of data that's more interesting, as a data scientist, than the size,” Brobst said, describing it as a paradigm shift that must be analyzed using new techniques. While stressing that there's no such thing as useless data, he explained that data has value density. “You have to compare the value [of data] compared to the cost of storing it. …[That is] the value density, and if there's no value, then economically it's useless,” he said. Most recently, it's the economics of high performance data warehousing that has seen a dramatic change. “Using the e-Bay example, the clickstream data is not stored using the same form factor of technology as their mainstream transactional data. They use the extreme data appliance, which is less than 1/10 the cost per terabyte than the mainstream appliance. That changes the whole economics.” Once you rummage through the huge amounts of data available, “then you have to discover the diamonds in the coal mine, and move that into the mainstream to use more heavily.” The kind of data that tends to be less useful is that which is not customer initiated, or what's called operational data. “Deposits and withdrawals by bank customers, that's very interesting — but automatic interest payments, not so interesting,” he said. Egyptian caliber Teradata's operations in a specific market relies more on the level of competition in individual sectors as opposed to the level of technological sophistication in the country. While retailers in North America capture everything in data, it's quite a different case in Egypt. “Retailing in Egypt…is mostly small shops on corners. They might have a POS but probably not…[and probably] without the economies of scale to have data warehousing make sense,” he said. The telecom sector, on the other hand, is very competitive. “It's not a cultural thing, it's an industry thing. Telecom in Egypt is very competitive and therefore there's more sophistication in using the data,” he added, citing Etisalat as investing heavily in data warehousing. Mobinil also uses Teradata solutions. “The skills sets we get in Egypt are quite high. Teradata opened a very large customer services center in Egypt because we could get the skills set, high educational level, good language skills across French, English, Arabic and so on.”
