Businesses Miss Steps in Big Data Transition
June 28, 2013
The transition to Big Data is a winding path with many forks, and even technologically experienced businesses are at risk of getting lost, according to Ron Bodkin, the founder and CEO of Think Big. Mr. Bodkin says he has devised a map.
“The move to Big Data is not a small, incremental step. It’s a major transition, like the adoption of the Internet in the 1990s,” says Mr. Bodkin, the former vice president of engineering at digital advertising company QuantCast , which makes extensive use of Big Data to analyze audiences. It’s not uncommon for companies to invest money in the transition, and fail to extract value from their efforts, he told CIO Journal. Often, that situation is attributed to a failure to define business objectives for the technology. Mr. Bodkin doesn’t argue with that assessment. “Certainly, IT needs a good business sponsor,” and if the technology isn’t integrated into the business from the start, “nothing will come of it,” he says.
Beyond that common misstep, companies often fail to comprehend how the implementation of Big Data logically ought to evolve—and they get stuck in early stages of deployment, according to Mr. Bodkin. “They often underestimate how difficult the transition will be,” he says.
That transition, done properly, includes five stages, according to Mr. Bodkin:
Traditional business intelligence. This is the bottom rung on the ladder of Big Data. It is no Big Data at all, just traditional business analytics, which often means the use of costly machines and software crunching predetermined queries, according to Mr. Bodkin. It’s an important first step because some of those skills are bound to continue to be valuable.
Implementation of cost-saving technologies. The first foray into Big Data typically involves making a transition to scalable clusters of cheaper commodity servers that run powerful but lower cost software such as the Hadoop framework, which breaks applications queries into chunks that can be reused by various distributed clusters. A lot of the cost savings are realized at this stage, “but companies often get stuck here,” because they use the new technology simply to run the same queries and solve for the same business problems that they did before, Mr. Bodkin says.
More compelling questions. At the next stage of development, companies start to unlock the native capabilities of Big Data. They learn to analyze unstructured data such as social media, which unlocks the power to conduct sentiment analysis, capturing attitudes and emotion. And companies learn to conduct more free-form searches, or “to explore,” Mr. Bodkin says.
The ability to conduct predictive analytics. At a more advanced stage of analysis, companies employ advanced algorithms that are capable of machine learning, a form of artificial intelligence. Instead of analyzing the past, or the here-and-now, they learn to make predictions about how people, objects or markets will behave in the future.
The ability to create and transform. The final step of creating value from Big Data is using it to create new products, much as Google used Big Data to create Google News, Mr. Bodkin says. And companies also use Big Data to transform their organization, bringing in new people with new skills, altering their processes and supply chains and overhauling their logistics.
Mr. Bodkin says he’s advised companies from Facebook Inc. to the Nasdaq OMX Group Inc. and Johnson & Johnson on such matters.
How do you use Big Data? Have those efforts paid off? Drop us a line or leave a comment and let us know.