Big data may be the big buzzword these days for big business, but it is also applicable to small businesses—especially those trying to wade through volumes of data generated by e-commerce.
“There are a lot of things that are improperly called big data, like a real estate agent with a lot of pictures,” said Curt Monash, head of Monash Research. “Your need for it will probably be more correlated to the size of your Internet presence than to the size of your firm.”
Small businesses that adopt big data are generally seeking to add analytics to e-commerce websites, explained Mike Olson, co-founder and chairman of Cloudera, a leading provider of big-data management software for the enterprise market. (Cloudera is a major supporter of the Apache Hadoop open-source software framework for large-scale data storage and processing.)
With the right analytical tools, “content is assembled dynamically and the system decides what to put on the page based on the behavior of the customer so far,” Olson explained. “The best way to interact with the customer is decided by big data.”
But the process involves tracking the activities of visitors to the site, in terms of what pages they go to, how long they spend on each page, what offers they consider, and what purchases they do make, etc., he added.
“There are many interactions between your browser and the server. As you scroll down and new content is fetched the delay shows how long you spent looking at the content. The Web log can get voluminous very quickly—it can easily generate terabytes of data,” Olson noted.
His experience is that customers turn to big data analytics if they have 10 terabytes of data or more, or they have smaller amounts of data but very complex analysis. A complete system can run $30,000 to $50,000 for software and training, he said.
Those willing to set up their own data centers can run it on industry-standard rack-mounted servers, or they can rent cloud-based facilities to do the same thing, he explained. The software can also reside at the user’s web hosting company, or they can get it through a software-as-a-service arrangement with a third party, he explained.
Elsewhere, “human-generated data usually involves transaction records, the volume is very correlated to the size of your business, and it grows slowly as machines to handle it get cheaper,” noted Monash. “Machine-generated data is the new flood, as machines get cheaper and spit out more and more data.”
“Big data is a heavily-hyped trend. If you need it you will know it,” Monash said. “It’s not something you adopt because someone you know at the country club is using it.”Tags: Business,Data Center,Productivity