Smart Choices

Chris Boebel, the director of information technology at Delta Sonic Car Wash Systems, had a big sales job on his hands.

Delta Sonic executives knew they could run the chain of car wash and convenience stores more efficiently. They also wanted to boost sales. But they had few tools with which to pinpoint their best-selling products, most-popular car washes, top marketing promotions and the impact of those promotions on sales in other parts of the business. For Delta Sonic to keep growing, such insight into the business and its customers was critical.

Boebel believed that a business intelligence (BI) application could help unlock future growth for the Buffalo-based company. A BI system could feed data on product sales and customer response to discounts back to the finance and marketing departments so that executives could zero in on what worked and what didn't.

But Boebel (pronounced BAY'-bill) knew winning approval from the business for a BI application would require a deft negotiating touch. Some business unit managers were likely to view BI as too expensive and elaborate a solution for a midsize company. Many Delta Sonic executives were wedded to decades-old, mostly paper-based reporting systems and would need to buy in to the promise of the new technology. Boebel would also have to overcome the sentiment that BI was only for the big boys — megacorporations with nation-wide or global operations that manage tera- or petabytes' worth of information.

Page Break

But after months of attending BI conferences, talking with numerous vendors and creating a presentation that showed how a simple BI Web tool could help the car washes to staff more efficiently, Boebel was ready. His careful research and early test results helped him line up support with the CFO, operations manager and some key business unit leaders to argue for a bigger investment in BI. In 2004, he got an OK from Delta Sonic's family owners to invest $US250,000 in Cognos's 8 BI application; this was a large investment for a company with annual revenue of $US200 million, as reported by Hoover's. In early 2006, after deploying some simple analysis of convenience store sales — showing which promotions were increasing beer sales and why, and which brands of cigarettes were the biggest sellers — Boebel and his team expanded the system to other business units such as the car wash and oil change business. The information gleaned from the convenience stores and car washes — based on knowing which promotions work best and generate sales of other products and services plus the ability to better track cashier statistics to prevent losses — have returned enough to recover half of the initial investment. "It's really paid off," Boebel says.

Delta Sonic is part of a growing trend. An increasing number of small- and mid-market CIOs are justifying the cost of BI applications to the business by showing how the insight into customer behaviour these tools provide can be harnessed to drive incremental revenue to the bottom line. Also driving the adoption of BI by this market are less costly applications. BI no longer requires an expensive and complicated set of solutions to access and organize the necessary data, database and storage applications (see "Smart Tools and How to Pick Them", end).

"That adds up to a significant IT investment," says Jim Baum, president and COO of data warehouse appliance provider Netezza. Instead, the new BI tools can provide an appliance-based approach that works within an existing IT infrastructure, which is easier to operate and maintain — and cheaper, he says.

In short, BI isn't just a technology for megacorporations any more.

"BI is showing signs that it is picking up for the midlevel market," says Greg Belkin, a research analyst at Aberdeen Group. "The midlevel market is under a lot of pressure to think out of the box so it can compete against the big guys."

For the midlevel market making the argument for BI can be harder than at larger companies. BI can be a relatively large investment (the investments for the small- and mid-market companies CIO spoke to were in the low six figures). It demands a serious time commitment from a smaller IT department that is likely overtaxed by user requests. But the business case for BI can be compelling, say small- and mid-market CIOs. While there is no research on BI's returns in the mid-market, anecdotal evidence from IT executives suggests that careful investments can pay off. They say a successful implementation relies on four ingredients: negotiating up front with business colleagues about how to invest in BI; establishing quick wins to build trust; giving users a stake; and making sure data is clean before launching an application. In that way, these IT leaders have established BI applications that are beginning to generate more ways to drive sales and curb costs.

"It's taken us time to get people used to it, but it has been worth it," Boebel says. "It's really beginning to show up in sales and cost savings. More will come over time."

Page Break

How BI Helps You Grow

One of the most difficult transitions for mid-market companies is that of making the leap to become a large business. Investing in bigger staffs to support that level of revenue, managing over an increasingly large geographic area and competing against businesses that have even larger budgets and better brand name recognition are just some of the challenges. So some mid-market CIOs have turned to BI as part of a strategy to grow and better compete as their companies bump up against bigger and better-financed operations.

At Delta Sonic, for example, business executives wanted a better idea of what was — or wasn't — driving sales so that strategic business decisions were no longer a case of trial and error. Executives at Clif Bar, a $US480 million manufacturer of nutrition and snack bars, began considering BI after its two largest competitors were acquired by multibillion-dollar global corporations. Blue Mountain Ski Resort got into BI after an investment company bought it and planned to more than double its size by investing nearly $US1 billion into the business over five years. And the midsize financial services firm JBHanauer saw BI as a key to its survival as it watched other mid-market financial institutions go out of business or get bought by the large financial houses. "We're one of the last surviving middle players," says Charlene Barnes, CIO and executive vice president. BI, she says, "is strategic to keeping us alive".

Given the competitive environment in which today's companies operate, BI and data warehousing consistently top the CIO's list of planned spending priorities, according to a July 2006 survey of CIOs by Merrill Lynch. "Organizations of all sizes are struggling with very similar problems," says Aberdeen's Belkin.

Companies are also creating more and more data. Netezza's Baum says his clients' data doubles every year. And with more customer data coming in from e-commerce activities, and technologies such as RFID providing more specific data on individual consumers, companies have information mountains to mine for what drives sales.

For mid-market organizations, using an Excel spreadsheet to inform strategic business decisions just isn't cutting it any more.

Page Break

Getting Business Buy-In

BI can be complicated and expensive. So for mid-market companies, with smaller IT budgets and staffs, convincing the business to invest in a BI application can be daunting.

That's why John Gowers, director of IT for Blue Mountain Ski Resort, says one of the most important steps in convincing the business to embrace BI is to start off with a relatively small investment, show a return and then expand the program from there. It's the plan he followed when 50 percent of the ski resort was purchased by Intrawest in 1999. (The Fortress Investment Group purchased the resort in 2006.)

Blue Mountain has multiple lines of business, from hospitality services to retail to equipment rental. Intrawest's ambitious plan for Blue Mountain included raising the number of skiers visiting the resort on an annual basis from 300,000 to 700,000, increasing lodging units from 200 to 1400 and taking the number of restaurants and boutiques from a handful to as many as 70. To manage that kind of growth and complexity, Gowers knew he needed BI applications to collect some basic information to manage the company's revenue and investments. He wanted to help the business determine which investments were yielding acceptable rates of return, how sales of certain products and services affected the sales of others at the resort and how to manage an expanding inventory of capital, including equipment and accommodations.

In February 2002, Gowers began working with Datavision to construct a simple BI tool to track property management and point-of-sale transactions. He kept costs down by using an older, free version of Datavision's Applix TM1. Using Applix as the back end, Gowers paid Datavision $US40,000 to build an ETL (extract, transform and learn) interface tool to manipulate and display the data. The application allowed the resort to track sales, such as ski rental revenue, beverage purchases and the number of lodging nights paid, on a per-skier basis. Sales could also be tracked by day and time. In the past, the resort had no way to measure sales other than in the aggregate, and those sales were stored in disparate business unit databases. This meant it took days, sometimes weeks, to collect, collate and distribute the data to executives. The resort thus had less time to develop special marketing campaigns to drive business during slow periods or to respond to unexpected up-ticks in business.

Within months of implementing the BI tool, Gowers was able to show the restaurant, ski rental and lodging businesses when their sales dipped or increased. Working with marketing, these businesses now could increase sales by knowing when to push incentives to customers. Sales for the resort's hospitality business have increased 67 percent in the past four years, an increase Gowers attributes in part to better business intelligence. In 2005, he used this earlier success to persuade business unit leaders to invest $US50,000 in a more up-to-date version of Applix to make it easier to incorporate new products and services.

The updated system soon paid off. For example, the ski shop typically would run out of ski boots on weekends, limiting revenue from lift ticket sales and ski rentals. Simply put, the resort had too many big boots and not enough midsize boots, which resulted in fewer rentals. The shop needed to shift its boot buying patterns to meet demand. But how best to do that?

Gowers turned to the updated BI application. He ran an analysis of the rental data to determine how to pay for more purchases of the midsize footwear by reducing purchases of large-size boots. The project started last northern winter, and by the end of the season, the ski shop had increased rental revenue by as much as 15 percent without raising its boot budget. By not turning away skiers, Gowers says, the resort had more satisfied customers, which, while hard to measure, certainly means more repeat business.

"In that particular case we showed a quick return," he says, "generating more revenue by making more intelligent decisions on how to run the business. As you do that, everybody begins coming to you with ideas, which only builds more support."

Page Break

Empowering the User

A quick success can showcase the advantages of BI. But an essential ingredient to getting buy-in is making the results accessible to business unit executives, midlevel managers and front-line salespeople so they feel empowered by the new technology.

For Clif Bar, that meant finding an application that sales representatives could easily use in the field, says Richard Boragno, VP of finance, accounting and IT. It also had to provide them with a simple sales story to help the company build a more intimate relationship with clients.

Working with Clif Bar's sales reps, Boragno searched for a Web-based BI application they could take to grocery store clients to show how each Clif Bar product was selling, how individual sales tracked over time and how sales per customer broke down in real time. The application had to have a simple interface that had a few easily understood icons. Sales reps couldn't be fumbling with the application in the field, where they typically had little time in front of clients and no easily accessible tech support. "They had to be able to fully understand it in about a two-hour training session," Boragno says.

He also asked two sales reps to help evaluate the vendors competing for the contract. "You have to look at it through the lens of a sales rep," Boragno says. "You're trying to solve their pain. To know what you are solving, you need to analyze their symptoms, understand what they need. If it is too complicated to use, it will lead to confusion and distrust" of the IT department's ability to deliver.

In 2002, on the recommendation of the sales force, for $US85,000 Boragno chose to buy the enterprise search application FAST Radar, supplied by FAST. The application provided sales reps with historical and real-time sales for each grocery store outlet Clif Bar supplied. The data allowed reps to show clients how many Clif Bars they were selling over specific time intervals and which were selling the most, allowing Clif Bar to keep those bars on shelves to increase sales for Clif Bar and the grocery store. As a result, Clif Bar reduced its sales forecasting variance from 30 percent to 15 percent. After a decline in sales in 2004, the company rebounded in 2005 with a 25 percent increase in sales, which Boragno attributes partly to the new BI system.

That success led Boragno last year to expand the FAST application to include the ability to input expense forecasts to support the quarterly sales forecast process. Doing so ended the use of Excel templates, saved time and eliminated errors.

By combining the reps' sales forecasts with historical sales figures that are matched with promotions and adjustments for seasonal sales, the finance department cut the amount of inventory on hand from 77 days to 35 days, saving the company $US2 million a year in reduced spoilage and interest costs.

What did Boragno learn from the experience? "Don't get all the bells and whistles right away," he says. "Keep it simple. Choose [an application] that can grow with you."

Page Break

Keep the Garbage Out

A BI application is only as good as the data it is using. If a company's data has errors or differing nomenclature and standards for products collected over decades, then be prepared to spend a lot of time cleaning up the database.

Last year JBHanauer's Barnes began implementing a BI application to track financial adviser and firm performance. It also allowed branch managers to track how different investments were selling at their office. Barnes knew clean data was critical to the effort and that there were inconsistencies in JBHanauer's database. For example, the firm had classified the various categories of municipal bonds, corporate bonds, preferred securities and other investments in different ways based on who conducted the transaction. Also, trades that had been cancelled were being counted as having gone through, exaggerating the number of buy orders. While cleansing data, Barnes suggests CIOs work with business unit heads to determine how much data should be included in analyses. Is it of value to include two years', five years' or 10 or more years' worth of the data? "Those kinds of things turn into significant challenges," she says. "It makes you analyze just how far back to go when including historical data. Is it helpful to go back a million years?"

Barnes recognized that the inconsistent data could undermine the BI initiative. After a thorough manual analysis of any common discrepancies, the data was cleaned up automatically. Barnes says the process took two programmers two months, a significant amount of resources for a company JBHanauer's size. But it was a task that had to be done to make the application useful.

In November 2005, Barnes purchased the Cognos 8 BI package for an undisclosed sum, and in October 2006 the first phase of the application launched across the company. Now, every morning when branch managers turn on their computers, the first application they see is a report on the previous day's activity in their office. Managers receive five separate reports, including a 45-day moving average on sales, which categories of investments are selling the best, how each financial adviser is performing and how the office's performance compares with other branches'. Branch managers use the reports to understand their product mix and identify selling problems or opportunities. "They are asking questions that they've never thought of asking before," Barnes says.

Just how much the system will improve JBHanauer's bottom line and keep it competitive with the multibillion-dollar investment houses remains to be seen. But Barnes notes that in October, when the application was launched, JBHanauer had its best month in 18 years. Was it a result of a good marketing, a hot market or a direct result of the BI application? It's too soon to be sure, but Barnes does know this: The BI application "did allow us to know this simple fact [immediately]".

Page Break

Innovation Time

Success generates enthusiasm and, therefore, support for BI throughout the company. Once that happens, it's time to start thinking out of the box.

At Delta Sonic, Boebel is looking to incorporate outside sources of data into the BI application to help the company manage staffing more efficiently. He hopes to include National Weather Service forecasts as one of the factors store managers use to plan staffing levels. For example, by knowing when and how much snow is forecast, as well as how long temperatures will remain below freezing (which means salt stays on the roads longer and cars remain dirty longer), Boebel can correlate that data with past statistics on car wash sales and snowfall to determine the number of cars that will likely come through Delta Sonic's car washes, avoiding the prospect of having too many workers or not enough.

Boebel has many other ideas for using BI. The only thing holding him back, he says, is convincing managers to use the application more. "But that will come in time as we show what can be done," Boebel says. "This is just the start."

SIDEBAR: Smart Tools and How to Find Them

BEFORE INVITING the first BI vendor in to give a pitch, CIOs at a small- to mid-market company should ask themselves one question: How will my organization use the data a BI application will provide?

Organizations answer that question in one of two ways, says Leslie Ament, practice leader and director of customer intelligence research at Aberdeen Group. And the way you answer that question will direct you toward the type of BI investment you may want to make.

Some organizations plan to use business intelligence to support strategic decisions, such as for sales, marketing, customer service, procurement or production. Little more than half of all organizations that use BI do so to support strategic decisions, according to Aberdeen. The two most widely supported business activities for which organizations use BI are sales and business development, and customer support.

Whether you use BI for strategic decision support or to streamline operations, either approach is valid. Recent vendor product offerings mean it doesn't have to be expensive, with some applications costing in the mid-five figures. "Small- and mid-market players believe that BI has a high cost point for entry and that you need a lot of horsepower," Ament says. "That's not necessarily true.

BI applications are becoming less expensive because vendors are introducing scaled-back versions that don't require so much computing power. Cognos, for example, last year introduced its Cognos 8 BI application, which provides organizations a chance to choose different levels of offerings.

— A HOLMES