The forecast for predictive analytics: hot and getting hotter


With analytics becoming critical to the day-to-day operations of more and more organizations, we are seeing the status of problem solvers rise. At the executive level there are new titles, such as chief analytics officer, chief data officer, and chief problem-solving officer. Last month, the Associated Press reported that jobs in the area of analytics could grow as much as 20% by 2018.

The so-called Information Revolution is affecting every industry sector. In the 2010 Information Week cover story which reported on just how much data some companies are using, many of the companies they mention are SAS customers. One of them, Catalina Marketing, has 2.5 petabytes of customer data. According to Information Week, “the company's single largest database contains 3 years' worth of purchase history for 195 million U.S. customer loyalty program members at supermarkets, pharmacies, and other retailers. With 600 billion rows of data in a single table, it's the largest loyalty database in the world.”

Using analytics, Catalina leverages all this data to provide customers with a personalized experience: “At the cash registers of Catalina's retail customers, real-time analysis of that data triggers printouts of coupons that shoppers are handed with their receipt at checkout. Each coupon is unique; two shoppers checking out one after the other, with identical items in their carts, will get different coupons based on their buying histories.”

Analytics can also be used to fight crime. The Los Angeles County Department of Social Services was being hit by fraud, waste, and abuse across their public child welfare assistance program. We worked with them to pilot a project to demonstrate that advanced data analytics and visualization can be deployed proactively to detect fraud. Using 6 years of data from government claims, payments, applications, third-party, and fraud case data, social networks were constructed based on linking common attributes for providers and participants.

With analytics, the County was able to identify unusual behaviors and then prioritize suspicious cases. The pilot program was a success. They had an 83% correct hit rate on provider fraud, and a 40% correct hit rate on participant fraud. In just 1 year, the pilot program uncovered $31 million in fraud.

We have been working with the financial services sector from the beginning. Banks need computations and techniques to happen faster. They can no longer afford to wait up to 24 h or longer to calculate risk. In fact, just over a year ago, a risk manager from a bank in Singapore was telling me that it took over 18 h to compute the value at risk for his bank. He needed to use 40 000 risk factors in the computation.

Using conventional hardware and software techniques, the computations can easily take 18 h to solve. However, we have been able to take advantage of new grid computing technologies to speed up the process. Instead of 18 h, this complex risk calculation is now solved in about 15 min.

Let me share an example in the healthcare sector. At Karolinska University Hospital in Stockholm, the challenge was to automate a manual system—called the global trigger tool method—to improve patient safety. As part of this method, medical staff had to search through patient records to identify adverse events that might have led to extended hospital stays, injuries, or deaths. These searches took about 20 min per record, and the staff only had time to review about 20 records per month.

By applying analytics to the unstructured text from the patient journal system, the hospital could identify potential triggers, such as specific treatments, or changes in condition during a patient's stay at the hospital. For example: Had the patient received an unplanned dialysis? A blood transfusion, or did he or she have a stroke when staying at the hospital? Today, using the automated system with its core analytical methods, the medical staff at the hospital has a better, and more structured, overview of the different events occurring during the patient's stay. There are many other examples. Our customers are using today's technologies to literally transform their industries.

I am concerned, however, that innovation in the United States is on the decline. An article published last year by McKinsey states that there is a shift in technology innovation from the United States to Asia, specifically to China, Taiwan, Singapore, South Korea, Japan, and India. Innovation is moving to where the skills are, and 21st century employers are looking for workers who are proficient in science, technology, engineering, and mathematics. In American schools, those are not areas of interest.

In 2005, a report to Congress called “Rising Above the Gathering Storm” urged the federal government to dramatically improve science and math education in elementary and secondary schools. On September 23, 2010, the authors of the report updated Congress. They said that in the past 5 years “The outlook for America to compete for quality jobs has further deteriorated,” and “the nation's ability to provide financially and personally rewarding jobs for its own citizens can be expected to decline at an accelerating pace.” The researchers also reported that “Only 4 of the top 10 companies receiving U.S. patents in 2009 were American companies, and 77 percent of global firms surveyed said they will build new research and development facilities in China and India.”

We must address this alarming trend. At SAS, we are working with government agencies, foundations, and other companies to bring high schools into the 21st century. One of the programs that we have been working on is the 1:1 Learning Initiative, where the goal is for each student to have his or her own laptop, and learn how to function in a 21st century world. SAS has also been developing online curricula for about 15 years, which we provide free-of-charge to every educator and student in the United States. In the past 2 years, more than 40 000 teachers have signed on. We feel that, by bringing the classroom up to date, students are less likely to be bored. Providing curriculum software that encourages critical thinking and problem-solving will create a more engaging learning environment.

Finally, I would like to share a few thoughts on attracting and retaining knowledge workers. While I was in graduate school, working on my PhD, I took a year off to work on the Apollo Space Program in Florida. I worked side-by-side with statisticians, engineers, and other knowledge workers. We worked in noisy cubicles, our time was monitored, and there were few perks. We had to pay for a cup of coffee. We were treated like workers on an assembly line, but the product we were manufacturing was coming from our minds.

When we started SAS a few years later and began hiring knowledge workers, we decided to create a different working environment. We gave people their own offices. We stocked the break room with free snacks, coffee, and soft drinks. We even provide aspirin and antacids. When one of our first employees had a child, we started an in-house day care. Today, we have on-site cafeterias and on-site healthcare facilities.

It is not difficult to provide benefits when you have a low headcount and healthy margins. We started with five people in 1976. Today, we have more than 11 000 employees working in 54 countries. And we have worked to make SAS a great place to work in each of those places. In just 2010, SAS was named a top employer in Brazil, Australia, India, Canada, The Netherlands, the United Kingdom, Portugal, Italy, Sweden, Mexico, Norway, France, and China. And SAS has been on Fortune's “Best Companies to Work For” list since they started it in 1998. We were very pleased to be named #1 for 2010.

There are some executives who feel that becoming an employer of choice is just not worth the investment. If you look at the short-term cost-benefit, they are probably right. But we have built SAS for the long term. We have had 34 consecutive years of growth and profitability, and we intend to continue that trend.

Let me give you a few examples. Our voluntary turnover, or attrition, is around 4%. Depending on the economy, our industry average ranges from 17 to 25%. Dr Jeffrey Pfeffer, who teaches at Stanford's School of Business, did a study on SAS, and reported that our low turnover saves us about $100 million a year in recruitment, training, and other costs.

Our on-site healthcare is staffed with 52 medical professionals, including 4 physicians and 11 nurse practitioners. Every year, we analyze the cost-benefit and, in 2009, the benefit exceeded the cost by $7 million.

It is difficult to quantify all decisions. Sometimes it just takes the resolve to do the right thing. For example, after the bottom dropped out of the market in late 2008, many companies began laying people off. We were in good shape, but our employees started asking if we were going to be laying off anyone. In early January of 2009, I did a global web cast and made a pact with all 11 000 employees. I told them that we would have no layoffs for the entire year, but I needed them to pitch in and help reduce expenses. Productivity went up, expenses went down, and 2009 was one of our most profitable years ever. There was no way of knowing in January what was going to happen that year, but it turned out to be the right decision. I renewed that pact for 2010, and I am looking forward to this year's results.