Among the many loud vows of how big data will revolutionise business are some quieter words of caution of what it cannot do. As the number of companies offering big data products and services explodes, some on the receiving end of vendor and start-up pitches are growing wary of the hype.
“When it becomes a cocktail party conversation term, its value is lost,” said Roger Ehrenberg, founder of IA Ventures, which invests in data start-ups.
While many believe in the power of big data, there is also concern that some early movers may be overpromising and under-delivering.
One advertising executive recalled a story of a shoe company that spent several months and lots of resources researching the factors that spurred people to buy certain types of shoes, in the hope of gaining insight into how to target ads better. The result was obvious: people tend to buy boots in the winter time and sandals in the summer.
Ad executives also caution against using bits of data to make business decisions, such as determining which ads are driving sales, without having a full picture.
“People will use data and information like a weapon,” said Quentin George, chief innovation officer at Interpublic’s Mediabrands advertising company. “In the medical industry, there is nothing worse than a little bit of knowledge. It is the same thing with marketing – if you don’t understand what these numbers mean or if you are responding or acting upon them without knowing why they are what they are.”
Healthcare providers have been barraged with pitches from big data companies hoping for a slice of the government subsidies that help hospitals and doctors’ offices upgrade technology.
“It’s almost creating a level of anxiety in the healthcare industry,” said Kathy Scheirman, senior vice-president of corporate services at Kaiser Permanente, a health system that has invested billions of dollars in information technology.
Other providers have become jaded by technology that is so new and unrefined, it is more cumbersome and frustrating than helpful. Nurses and doctors complain that computer scientists do not understand the workflow of a hospital, and the effort of negotiating with engineers to get the analyses and products that are truly useful can often outweigh the benefits.
“What vendors say will happen, doesn’t always happen,” said Bonnie Wakefield, a research professor at the University of Missouri’s school of nursing.
In various industries, some big companies have attempted overly ambitious, long-term overhauls of their systems for big data and been disappointed with the results, said James Manyika, a director at the McKinsey Global Institute, a research arm of the strategy consultancy.
“Some of the frustration we hear is from people who have tried to re-architect [their systems] and clean up all their data,” he said, adding that frequent experimentation, even on imperfect data, has tended to yield better results.