When Deborah Leff was preparing her presentation for IBM Think, she looked online to find the best antonym to the word “insight.” The top results were “ignorance” and “stupidity.”
While not exactly an example of data science in action, the findings showed Leff, IBM’s global sales leader of business analytics, the stark difference between those who gain value from the way they manage information versus those who don’t.
“It’s not about (using data to make) pretty pictures, it’s about making confident decisions,” Leff told the IBM Think crowd, which was livestreamed from Las Vegas. Big Blue is using IBM Think to showcase a mix of research and technology across areas such as artificial intelligence, the cloud and other elements of what it calls “cognitive computing.”
Of course, the idea of data science and analytics is no longer new, but Leff said IBM has been observing a lack of “analytics governance” — in other words, the structures and policies around the way analytics projects are carried out.
“Data science often starts as research projects — companies are working on data with the assumption that the insight will reveal itself and will be the justification for a project,” she said. “That is a really risky and time-consuming approach at a time when companies have to move fast.”
Instead, Leff said data science teams should begin with an established goal that maps directly to creating business value. She gave the example of a bike-sharing rental service which needed to improve the way it managed inventory and ensured maximum ridership.
This started with descriptive analytics that showed which bike stations were used most during the week versus weekends. Diagnostics analytics were then applied — using weather data as well as historical customer data — to look at how changes in the weather affected ridership. Advanced analytics could then suggest when demand was likely to be high, while prescriptive analytics showed which bikes should be moved to a particular station to make the best use of inventory and adjust maintenance priorities to decrease tire pressure on bikes that tended to get flat tires.
The same kind of thinking can be applied across many other kinds of companies and markets, Leff said. She cited one customer, a $20 billion business, which was was able to drive an incremental $4 billion in revenue over five years with just one data science project.
Beyond the educational value, IBM Think is showing off all the technologies the firm has to offer these various flavours of analytics, including its Cognos business intelligence suite as well as SPSS, a statistical software package.
Leff said the tools need to be matched by the right training to ensure everyone that looks for insights can come away with a uniform look at what it’s showing.
“Each analytics user thinks about data differently. That can be a beautiful thing that leads to creativity and insight — but it can also lead you down the wrong path,” she said, adding that humans tend to look at visualized data as “evidence” versus text-based data, which tends to be viewed as an argument. “Too many activities are leading to metrics mayhem. There are people showing up to meetings with conflicting numbers.”
IBM Think customer speakers included FleetPride, Inc., which has been working on its approach to analytics governance for about four years. The Irving, Tex.-based firm is a B2B distributor of parts to the heavy-duty truck and trailer industry. According to Homarjun Agrahari, the company’s director of advanced analytics, the biggest challenge was getting to understand the current state of its inventory and other key performance indicators (KPIs). As those KPIs were established, the firm has used Cognos and SPSS to set goals and measure its progress on optimizing various aspects of its business.
Agrahari said the biggest success factor was basing his work off real problems brought forward by the business, such as overtime costs going too high, too many dropped calls in its call centre and so on.
“If I have a project where I cannot deliver value . . . I just do not have enough time for that,” he said. “Sometimes the solution is a dashboard. Sometimes it’s a change in business process. Sometimes you have to build new software to solve their problem.”
Seth Dobrin, IBM’s chief data officer, said the company has been doing similar work internally to better manage the way it collects and synthesizes data across its various divisions.
“Governance is critical — usually when you mention the word ‘governance’ people cringe, but governance actually frees you to do these really cool things if it’s done properly.”
IBM Think runs through Thursday.
Latest posts by Shane Schick (see all)
- Why are COOs the forgotten B2B marketing target persona? - November 19, 2019
- ActiveCampaign VP predicts its machine learning features will offer an alternative to A/B testing in e-mail marketing - November 15, 2019
- InsideView product exec says Data Integrity will address B2B firms’ cleansing and visualization needs - November 14, 2019