Companies considering adopting advanced business analytics should:
- Determine precisely what analytical tools the company needs
- Weigh the advantages of buying versus building
- Assess your ability to commit the necessary time and resources
Doug Twiddy isn’t the sort of fellow you would expect to find using advanced business analytics. Based in the Outer Banks of North Carolina, in a small town named Corolla, Twiddy, 66, runs Twiddy & Co., an eponymous real estate company that rents out beach homes to vacationers. It’s a good business, with annual sales of more than $50 million, but it’s hardly high-tech.
It turns out, however, that the family-owned company must handle an astounding amount of data. It manages 906 properties owned by other people and must stay on top of the comings and goings of renters each weekend during the season. It schedules cleaning and repair services involving some 1,100 plumbing, heating, carpeting, pool and spa, and extermination vendors. It tracks how much of the rent it collects goes to the owners of the homes versus how much it keeps. Traditionally, the company has handled all of this information manually, using spreadsheets among other tools.
But after Twiddy’s son Clark, now 36, came home from serving as an intelligence officer in the U.S. Navy, Twiddy & Co. started working with SAS, the software company based in Cary, North Carolina, to capture all the data electronically. It took time and effort to implement a SAS system, but along the way the company reduced errors by 15 percent and also cut labor costs. More than just collecting data, it also can predict how much rent can be charged at a particular time of year—it can differentiate between a six-bedroom home on the ocean versus a four-bedroom home located four blocks inland—and evaluate which vendors are the most cost-effective and reliable. “I’m tickled to death with the system,” says Twiddy Sr.
“One of the errors that a C-level person can quickly make is not giving enough executive support or attention to the process.”
It turns out that small- and medium-sized brick and mortar companies can use analytical tools just as the largest corporations can—or the hottest Web-based social media startups or the biggest intelligence agencies with three-letter names. They can use those tools to eke out real competitive advantages against rivals that haven’t embraced the new capabilities. Even the most advanced tools, such as those IBM developed to such powerful effect with its Watson competitor on the Jeopardy game show, are within reach of companies with $10 million, $50 million or $100 million in annual sales. It’s happening across the landscape in all industries, says Andy Monshaw, general manager of IBM’s Midmarket Business, based in Somers, New York. “The only thing that is a barrier to adoption right now is the knowledge of the mid-market customers themselves,” Monshaw says. “They don’t yet all know that they can leverage this.”
In fact, the business of selling business analytical systems to small and medium-sized enterprises (SMEs) is one of the hottest segments in the information technology (IT) industry. IBM has purchased 25 companies in the area, including Cognos, which specializes in business intelligence systems. Hewlett-Packard’s $11 billion purchase of Autonomy was intended to help it secure a position in the business of analyzing “unstructured data,” one of the hottest niches in the data business. SAS, which argues that it is a leader in the SME market, said in December that its sales of business analytics to those companies were up 42 percent over the prior year. All of the giants—including Google, Microsoft and Oracle—are in the game, as are many smaller, specialized outfits.
As the field has rapidly evolved thanks to increasingly sophisticated algorithms and raw computing power, the vocabulary appears to have shifted. The phrase “business analytics” now encompasses many sub-disciplines—databases, data mining and historical pattern recognition, dashboards, online collaboration and predictive tools. Increasingly, customer relationship management (CRM) overlaps with business analytics because tracking what customers buy and whether they liked what they bought is essential in understanding the customer. More and more companies are combining the management functions that oversee customer relations and IT systems because the two have become so intertwined.
There seem to be at least four stages in adopting an analytical system:
STAGE 1: What Will We Analyze?
It’s important to go through a considered thought process before any decisions are made about what type of systems to purchase or develop, says Paul Magnone, co-author of Drinking From the Fire Hose and a 21-year veteran of IBM. There are specialist companies and there are integrators who bring various specializations together under one roof, “but the step before that is to get a grasp of your business and ask the right questions,” says Magnone. “It’s all about the framework you take going into it and about the culture within the company and within your own self. The tooling comes later.” Those questions include: What is most important to the business? What matters most to your customers?
Most companies embarking on their first push into business analytics have been gathering data on multiple systems that don’t communicate with each other. “Before embarking on data mining or analytics, one of the biggest things the company has to tackle is the variety of data sources,” says Tapan Patel, global product marketing manager for predictive analytics and data mining at SAS. “How can I integrate those data sources?”
The planning stage may require real introspection. Ken LaVan, co-founder of Fort Lauderdale, Florida law firm LaVan & Neidenberg, which represents disabled veterans in dealing with the Veterans Administration and Social Security disability programs, started out by examining every step of its processes. “We documented everything we did in the office,” says LaVan, who had prior experience as a computer consultant. “Anything that anyone did manually, we put that on paper.”
It’s only at that point that a CEO can know precisely what analytical tools he or she needs. “If your needs are for accurate forecasting of car sales or pricing, then the analytics should focus more on forecasting,” says Patel. “If you’re trying to determine which customer segments you should pursue and what offers should you put before the customer, that’s more like data mining.”
STAGE 2: Do We Buy or Build?
One of the debates in the field is whether small- and mid-size enterprise CEOs should try to develop their own business analytics in cooperation with vendors or simply rely on the outsiders to install systems that essentially “plug in” to what they already have. The big vendors argue that they have already built hundreds of industry-specific models and can tweak those systems for a particular SME. They can even deliver the services via the cloud, meaning the customer pays for use as he or she downloads or utilizes software and other services. That raises a corollary issue: do you want to take a big plunge on a major expense or do you want to proceed with a step-by-step implementation with a long-term partner?
The reality on the ground seems to be that most small company CEOs want to have a hand in developing their analytical capabilities gradually, not in a single big-bang moment. LaVan had six different computer systems, which made it difficult for his firm to manage millions of documents associated with winning benefits and disability payments for veterans. Starting three years ago, he examined off-the-shelf software solutions as well as the offerings of large developers. “I didn’t think any of them could be converted to run our business,” he recalls.
Partly because his previous database was based on Lotus Approach, he gravitated in the direction of IBM, which owns Lotus. He ended up working at an IBM partner company, Georgia-based Group Business Software, that was able to bridge the gap between a relatively small law firm and Big Blue to create a custom solution that combined the firm’s email and instant-mail systems, collaboration tools and client data base. “We were extremely hands-on during the process,” LaVan says. “Internally, we spent 3,000 hours on developing the software.” Combining internal and external costs, he figures the total tab was $1.5 million.
Likewise, Oberweis Dairy, a fourth-generation family ice cream and milk company based in North Aurora, Illinois, had a mish-mash of systems. Even though sales are in the $50 million to $100 million range, the company faced surprising complexity. The reason is that it had two main products—ice cream and milk—and three distribution channels: a chain of 46 retail ice cream stores, a home delivery system for milk in glass bottles and wholesale distribution through 1,000 grocery stores in five states.
The company’s different databases didn’t talk to each other, plus it had custom applications to manage its milk routes, as well as other point-of-sale systems in its stores to manage Moola, its loyalty card program for customers. “If we wanted to ask ‘how many bottles of milk did we sell yesterday?’ we would have to go to a couple of different systems to get the answers,” recalls Joe Oberweis, whose great-grandfather started the business in the 1920s. “It was cumbersome and complex.”
So the company took all its databases and recreated them in SAS “tables,” or common formats, so that they could communicate, and created what are called SAS “views” of the data. Oberweis now knows whether a customer who receives milk on a home delivery route also buys ice cream at a store and what sort of special offers can be made to lure that customer into other purchases. It took time and effort to transfer its systems into SAS tools, but it was more practical than simply buying a completely new system. (See “Oberweis: Milking Data” below)
STAGE 3: Are We Ready to Invest?
If a CEO decides to co-develop a business analytics system, odds are that he or she will need internal talent to help. Twiddy, in North Carolina, could not have gone down that path without his son, the former naval intelligence officer, and LaVan was able to rely on his experience as a computing consultant. Similarly, Joe Oberweis ended up hiring Bruce Bedford as his vice president of marketing analytics and consumer insights; Bedford had been an external IT consultant.
Other people also will probably need to be trained. “The process of merging SAS capabilities with my company was more intensive than I expected—it fooled me,” says Doug Twiddy. He had three to five people working on developing “cubes,” or storage mechanisms, for weeks and had to send some of them off for training at SAS headquarters. “I’d go check on the people in the room and ask, ‘How are you doing?’” he recalls. “They’d say, ‘We’re building a cube.’ I was waiting for results and they were laying the foundations of merging our stuff with SAS.”
SAS’s Patel says the challenge is more than just building a piece of software; it’s about knowledge transfer. “One of the errors that a C-level person can quickly make is not giving enough executive support or attention to the process,” he says. “They have to provide leadership. They have to make sure they have the appropriate skills in-house to understand the technology.”
STAGE 4: Do We Understand the Impact?
The reality is that reaching a certain point of sophistication with business analytics changes the way the company is run and challenges the traditional culture. “The sophistication of the platform is scary to a lot of people,” says Oberweis.
He says he can sense in his company that there is a divide between people who understand the business but not the software versus those who understand the software but not the business. He wishes that 30 people would be anxious to access the business insights coming from the data, not just five or six. All of which explains why going down the path of business analytics can be so profound. “It’s not just a software purchase,” Oberweis says. “It’s a strategic decision.”
The ROI of Analytics
Whatever complexities may exist, the payoffs from the successful implementation of an analytical system are clear. LaVan invested $1.5 million, but he says, “The return on investment has been tremendous. We’ve captured five times that much in new business.”
His firm has been able to reduce advertising costs and to quadruple the number of clients over three years. The firm has gone from having 40 employees and 5,000 clients to 120 employees with 20,000 clients, reflecting clear gains in per-employee productivity. “We have a pretty significant competitive advantage over competitors,” he boasts. “What our system does is manage the claims and organizes the millions of documents. It does the analysis and then cues it up to the employee to review.”
Improving a company’s entire decision-making process, and thus giving it a leg-up in the marketplace, is, of course, the ultimate goal of going down the path toward business analytics. Concludes Oberweis, “We wanted to make decisions with really clear information rather than just guessing what reality is.”
The bottom line? Even very small companies can benefit from business analytical tools that may literally transform their businesses.
Oberweis: Milking Data
Oberweis Dairy has come a long way in a very traditional business. Just a few years ago, the company didn’t have a solid understanding of how the buyers of its milk, ice cream and yogurt products were behaving. Some bought milk from traditional milkmen on their routes, some made purchases at Oberweis dairy stores and others shopped at grocery stores that carried Oberweis products. There was no way to get a complete picture of who was buying what and why, because there was no centralized analytical system. “Two years ago,” says CEO Joe Oberweis, “you’d walk into a room with file cabinets with paper stacked all over the place, completely disheveled.”
Unhappy, Oberweis started working with IT consultant Bruce Bedford to improve his analytical systems. After taking a series of relatively modest steps, Oberweis decided to take the plunge by implementing a SAS business analytics system and hiring Bedford to manage it. “Bruce pushed me over the cliff,” recalls Oberweis. “The first major step was getting the data organized in a fashion we could use it.”
As a result of the implementation, the company can now match up the data from the two sales channels it completely controls, traditional delivery and its own dairy stores. “We can look at customers across channels,” says Oberweis.
It also can segment its customers and experiment with different ways of marketing to them. It sends promotional offers to some customers, working with Valpak, a direct marketing company owned by Cox Communications. “We use SAS to analyze those results and modify the solicitations,” says Oberweis. Those tools work both for maximizing the number of home customers and driving traffic into the company’s stores.
The company also works with Groupon, the Internet-based company that promotes special offers from retailers to targeted online audiences, and it has become much more sophisticated in using its Moola loyalty cards. “We wanted to know if loyalty card customers behave differently in our stores than customer who don’t have loyalty cards. SAS helped us understand the difference in behavior,” Oberweis says. “We get more frequent visits from loyalty card holders with Groupon offers than from non-loyalty cardholders. We drove increased traffic and we increased the size of the spend per visit.”
So the power of data and the ability to extract lessons from it has given Oberweis a powerful boost. Says the CEO: “It’s not just more sales, it’s more profitable sales.”