Business intelligence tools that once produced only simple reports of transaction data can now perform real-time analysis of types of data produced both within the organization and in outside domains such as social networks. Data analysis is also moving from the “what has happened?” of traditional data mining to situational awareness that can predict “what will happen?”
Enterprise analytics experts at a panel discussion on “Analytics for Rich Customer Experience” at a recent TCS Innovation Forum in San Francisco described how such “big data” analysis is already providing business value, and some of the opportunities – and challenges – for the future.
Brian Dolan, vice president of product at analytics vendor Discovix described how a “Hypertargetting” team at a previous employer used advanced machine learning and artificial intelligence technology to predict with what he calls “scary accuracy” if a person is about to get married or buy a car.
Gautam Shroff, head of TCS Innovation Labs in Delhi, said that while blogs often trail newscasts in discussing major events, blogs or Twitter often report regional or smaller-scale events first. For example, a plant affected by a flood in the remote part of the world would not get covered on CNN but would be of interest if that plant was a supplier to an organization. If people in chat rooms are talking about switching to a competitor’s product that could signal an important to trend you could react to before seeing a sales dip. We have set up engines which listen to Twitter, running large clusters, and all the time.
Jim Fitzpatrick, global chief information officer of financial information services provider Experian, said social media tools such as Facebook and digital advertising are linked to more traditional marketing services businesses such as email marketing, SMS marketing, list processing and direct mail. The ability to understand who knows who, what their propensity to buy might be, how to track them down in real time and to understand how effectively an ad campaign worked “is pretty mind blowing stuff,” he said.
On the data end, said Shroff, it’s possible to drive business value by combining POS Data, Basket Data, Channels Data, ERP Data (Inventory), Market Data and Social Media Data (Promotions / Launches) and making them available to sales staffs. This type of insight is most effective, he says, when combined when trials of different promotions to gather new data.
One area of interest, particularly in North America a significant increase in B2B sort of analytics: people really want to understand “Who can I trust to do business with?” to enable them take on the next contract, said Fitzpatrick. “In the SME segment, we believe we are seeing an increased volume as people are trying to look for a trusted partner with whom they can go out and do business together…That is a big market, particularly on credit side of it.”
Among the challenges in meeting these “Big Data” needs, attendees said, are making this data available quickly without waiting to process and export it to a data warehouse. Big data also requires new skills. As massive data acquisition and storage becomes increasingly affordable, said, Dolan, a wide variety of enterprises are employing statisticians to engage in sophisticated data analysis. He pointed to an emerging practice of “Magnetic, Agile, Deep (MAD)” data analysis that is a radical departure from traditional enterprise data warehouses and business intelligence.
For more details on the promise and challenge of analytics, read the full white paper here.