The Rise of Machine Learning

Cost cutting isn’t enough
By now, using Big Data and analytics to cut costs, boost efficiency and grow existing business streams isn’t enough; they’re table stakes. The CEOs seeing the greatest opportunity to drive new revenue streams from big data are those confronting the highest level of disruption from it, CapGemini reports. Many of these companies are actually finding ways to monetize the data itself to create new lines of business.

Silos can be broken
Companies that best harness Big Data break down barriers to its use within and outside. Internally, that means people coming together from various business functions to work with IT to engage the Big Data agenda. It can also mean collaborating with other companies.

Hershey CEO J.P. Bilbrey, for instance, spearheaded creation of a new consortium for data analytics by partnering with Big Data pioneer Palantir, which is getting a group of CPG rivals to swap data to jump in front of emerging trends that affect the industry as a whole.

Garbage in, garbage out
Machine-learning systems can’t correct wrong data and can only complement, not completely replace, human intelligence. “It’s very good at pattern recognition and finding correlations in massive databases, but not so good at reflecting the causal structure of what’s going on,” says Jim Guczcza, chief data scientist for Deloitte Consulting in the U.S. “That’s where domain knowledge comes in handy.”

Humans still count
CEOs must combine Big Data operations with the domain expertise of actual human beings to get the best results. “You can’t just turn it all over to a machine that will do your thinking for you,” warns CEO Avi Steinlauf. “Tools can help, but you need smart people thinking about things in curious and intuitive ways.”

Extra bandwidth becomes available
Effective machine learning reduces the need for human labor, but that doesn’t necessarily mean it should eliminate jobs. “You can certainly reduce staff and cut costs, but we see many clients looking to leverage this newly freed capacity to enable workers to provide additional value to the business,” says Chip Wagner, CEO of Alsbridge, a technology-consulting firm.

Democracy energizes data
The greatest potential for data can be unlocked by giving front-line workers essentially unlimited access to information that used to be difficult to obtain or required more senior managers to interpret. Tesla, for instance, finds that this approach improves manufacturing and quality because test engineers and others can see things in the data that their superiors have missed.

Knowing is not doing
Judging by one important criterion, few companies so far demonstrate actual Big Data skills: Even though the demand for data scientists has tripled in three years, only 6% of large companies employ as much as one of them. What’s more, because data scientists with experience remain few and far between, the talent chase will get only worse. This is one reason that only 27% of executives surveyed by CapGemini described their big data initiatives so far as successful.

Privacy issues must be addressed
Companies’ use of Big Data creates an unwelcome companion: distrust of that use because of privacy concerns. The more data that B2C companies have compiled about them, their purchases, their behavior and their preferences, the more consumers can be cautious about sharing such information and sensitive about how the data is used. One way CEOs can counter this obstacle is to openly offer benefits to customers in exchange for their private data.

For a retail chain that wants to use data it is gathering about in-store shopper behavior, for instance, this could involve offering free Wi-Fi in the store or enhanced service from a “personal shopper.”

Make it a personal tool
CEOs, including Carlos Rodriguez, of Roseland, New Jersey-based ADP, now are using machine learning to prepare more easily and effectively for board meetings and analyst updates.  “Programs can go out and dig up all the information that the CEO used to have to reach for individually, package it up, apply Big Data analytics and present it to him,” says Stuart Sackman,
ADP’s CIO. “For each subsequent meeting the system gets smarter and smarter in condensing and refining data.”

Patience is required
Machine learning, like human learning, remains a work in progress. “It’s never quite done and it’s never quite right,” says GeoDigital CEO Chris Warrington. “But the longer you live with it, the better it becomes. So as a CEO you need to be patient with it.”