A higher form of intelligence
Analytic applications help companies make smart business decisions
By Sidney Hill, Jr., Executive Editor -- Manufacturing Business Technology, 9/1/2001 12:00:00 AM
Put away the shovel and pickaxe; the days of mining for data are over. We are now in the age of business analytics.
Software vendors are offering tools that automatically convert raw data into valuable nuggets of information, and literally put that information at users' fingertips. As a result, business managers can not only see how their companies are performing; they often are able to take immediate action if some part of the business is not performing up to par.
"What we are seeing with the use of business analytics is companies taking a more holistic view of their businesses," says Carol Ptak, an executive with IBM, Armonk, N.Y., who helps small- and-medium-sized enterprises select and implement e-Business solutions. "This new generation of business analytics technology is being used to identify those business rules that are constraining a company's ability to continually enhance its profitability."
In many cases, this is causing companies to act in ways that defy conventional wisdom. "Traditional business software was designed to accommodate existing business rules," Ptak says. "With the advent of business analytics, software is becoming a tool for evaluating whether we are using the right business rules.
"For instance, in distribution, we've always said you should deliver products from the distribution center that is closest to the customer," Ptak continues. "That's just common sense, right? But when you step back and analyze the situation, you may find that you can minimize costs, and still provide the same level of customer service by violating that rule. We never had the technology to do that type of analysis before."
Not only is such analysis now possible, but the technology to do it is available, in some form or other, to companies of all sizes. This technology also is coming from a variety of sources.
Sources of intelligence
Bosma Machine and Tool, a $10-million-a-year manufacturer of custom metals products based in Tipp City, Ohio, boosted its on-time delivery rate by 20 percent and cut more than $260,000 off its annual labor costs after it began using the analytic capabilities embedded in its enterprise resources planning (ERP) system. Bosma purchased that ERP package from Made2Manage Systems, Indianapolis.
KLA-Tencor, a $2-billion-a-year supplier of semiconductor manufacturing equipment based in San Jose, Calif., is using the analytic capabilities in its manufacturing execution system (MES) to continually improve the quality of its products. The MES vendor is Datasweep, which also is based in San Jose.
In addition to ERP and MES suppliers, vendors that specialize in supply chain management (SCM) and customer relationship management (CRM) software have embedded analytic capabilities into their products. There also has been a recent explosion of start-up companies offering standalone analytic applications that extract data from these other systems.
The newer startups typically refer to their products as supply chain event management (SCEM) or supply chain performance management (SCPM) software. Companies in this space include SeeCommerce, Palo Alto, Calif.; and Atlanta-based Viewlocity. Silvon Software, Westmont, Ill., has evolved from its roots in data mart solutions to offer analytic applications for business performance management.
To a degree, all of these analytic products—both the standalone packages and those embedded in application suites—evolved from the business intelligence (BI) applications that have been around a decade or more. When they first emerged, BI applications promised to help companies grade their own performance against a set of key performance indicators (KPI). The idea was that consistently hitting certain KPI targets—such as 98-percent on-time delivery or a 1-percent product defect rate—indicated that the business was in good shape. But manufacturers never fully embraced BI applications, largely because they were expensive to install and difficult to use.
"Just five years ago, we were hearing stories about BI projects that cost $5 million to $10 million and yielded very little in the way of results," says Mike Hennel, CEO of Silvon Software. "Today, companies can get a complete set of analytic applications up and running quickly, and achieve real results for less than $500,000."
No more stale data
Deploying BI first required building a data warehouse, or a central repository for all of the data that was generated in a company's transaction systems. Once data was in a warehouse, someone had to figure out how it could be used to first assess—and then improve—the company's performance. Companies that could afford to do so hired data analysts who were trained to dig through these mounds of a raw data for nuggets of information that provided insight to business performance. This is where the term data mining originated. In most cases, the data analysts delivered reports to managers and executives, who often found they were getting the information too late for it to be of any real value.
This was the case at Bosma Machine and Tool, where managing production costs was the key reason for adopting business analytics. "If we find out about potential cost overruns while a job is still in progress, we can make adjustments to deal with those issues," says Matt Sheridan, Bosma's chief financial officer. "With our previous system, we never knew the cost of a job until it was complete. At that point, if we had a cost overrun, all we could do is try to remember to do things differently the next time." But even that was not a very effective policy for Bosma, since many of its orders are for one-of-a-kind products that it will never be asked to make again.
The developers of BI software made it somewhat easier to implement their applications by introducing the concept of data marts, which essentially were smaller data warehouses containing information related to specific business functions such as sales, marketing, or customer service. But data analysts still were needed to extract the right information from a data mart, and the information still tended to be outdated by the time it reached a manager's desk.
Choosing target markets
The current generation of analytic applications addresses nearly all of these issues, although they do it in different ways. The analytic features embedded in ERP or other application suites typically are geared to get at data that is commonly sought by companies in the target market for that particular application. The Made2Manage ERP system, for example, was built specifically for small- to medium-sized manufacturers like Bosma. The system also can accommodate various manufacturing modes, including Bosma's job-shop style.
Sheridan says the system also proved flexible enough to provide the exact information Bosma needs to keep its production costs under control. "We started out using the system's standard costs reports, which summarized the cost of an entire job," Sheridan says. "But we wanted to know how much time was being spent on each individual operation as the job was progressing." That was necessary because Bosma prices its products by estimating the number of hours for each operation required to complete the project, and it wants to constantly compare the actual hours spent on each operation to those initial estimates.
"We were able to expand those summary reports to track the progress of each operation," Sheridan says. "Our department managers get these reports every day, which allows them to change production schedules or processes right away if they spot problems." As a result, he adds, Bosma has been able to both lower its production costs and improve its on-time delivery rate.
The Datasweep Advantage package used by KLA-Tencor was built specifically for controlling processes on the plant floor, thus its analytic features also are geared for improving plant-floor operations. KLA-Tencor uses those analytic capabilities to isolate the reasons for any failures that occur as it tests its products. The system kicks in as soon as KLA-Tencor begins building a product or receives a component.
Each component is assigned a unique serial number that is entered into the Datasweep database. That starts a tracking process that will continue through the life of that product. "We can collect information about everything that happens to each part–—its test history, repair history, who has worked on it," says John Moore, KLA-Tencor's e-quality program manager.
Faster problem-solving
"When something goes wrong, you typically have to fix the manufacturing process, the material used to build the product, or the training you are giving the production workers," Moore says. Before installing the Datasweep system, KLA-Tencor attacked problems with what Moore refers to as gut feel.
"Most of the time we solved the problem that way," he says. "But sometimes it was a long process, because anytime you get more than two people looking at the same problem, without any data to rely on, they will have different ideas about what caused the problem. This system gives us the data we need to find out exactly what happened—and under what conditions—so we are solving problems much faster. This has cut our manufacturing cycle times, and improved the quality of our new products because we can pass on the things we learn to our design teams."
Ptak, the IBM executive, believes the ultimate value from business analytics will be an ability to use them to create special programs for serving individual customers. "Going forward, simply providing good products will not be enough for companies to maintain profitability," she argues. "They will need to develop deeper relationships with customers by providing solutions to their customers' specific business problems. Analytics provides the technology to create those solutions."
Ptak also contends that the age of business analytics began when SCM vendors introduced applications to design optimal configurations for complex distribution networks. Up until that time, she says, most companies treated all customers the same. But those rules started changing once these network design systems started revealing the cost of serving individual customers.
Today, a growing number of manufacturing and distribution companies are turning to vendors like Atlanta-based SynQuest for help in monitoring their distribution networks. With SynQuest's applications, companies can use the data contained in their existing business systems to build virtual models of distribution networks. They can then apply optimization technology to those models to develop processes that allow the company to achieve certain business goals, such as the highest levels of customer service.
A holistic view
"We do a pretty good job of looking at things holistically, analyzing both financial and operational performance," says Chris Jones, a SynQuest executive vice president. "In some cases, we find that a company is losing money by serving a particular customer. That is when we recommend that they change business rules. They may have to institute a surcharge for delivering to a particular area of the country, or not promise quick delivery to that area."
The SynQuest applications primarily comprise a strategic business tool. Jones says companies typically use the applications to conduct monthly or quarterly performance reviews and then decide whether any policies should be changed.
The SCEM vendors say their products can serve both strategic and tactical purposes. On the strategic side, these systems also contain modeling capabilities that allow designing business networks that adhere to a specific set of rules. The tactical components of these systems take over once those networks are running, in the form of alarms or messages that alert designated individuals to potential problems.
Beyond simple BI
On the surface, most SCEM applications appear to have simply followed through on the vision created by the developers of early BI systems. These applications all have their own databases, which are used to aggregate data taken from existing systems operating within an enterprise. Instead of data marts, however, these newer systems typically have a series of prepackaged applications that are designed to analyze the operations within a specific part of the business.
For instance, Silvon's product suite, which is called Stratum, has separate applications for analyzing sales, marketing, purchasing, and manufacturing operations. It also has a series of programs for examining customers' buying patterns, as well for scrutinizing the performance of Web-based business initiatives and tracking the cost of various business processes.
The SCEM and SCPM vendors say there are two major differences between the applications they are selling today and the previous generations of BI applications: the profile of the average user, and the timeliness of the data.
"Very few people in a company actually used BI applications," says Mike Monsbach, a vice president with SeeCommerce. "Typically, a couple of people with degrees in operations research were the only ones who understood those systems. Performance management empowers people across the organization to make decisions on-the-fly. They also give people access to the information they need to do their jobs, in a form that they can understand. A person in the logistics area can view data through a personalized portal that is different than what someone in marketing or sales would see."
Maher Masri, a vice president with Viewlocity, says the ability to work with real-time data is what drives the current generation of business analytics. "With BI applications, you were dealing with data that just sat in a warehouse until someone figured out what to do with it. It was historical data that did not reflect the day-to-day activities of the business."
As Masri points out, trying to steer a business correctly by relying solely on history is just as hazardous as trying to drive a car forward by only looking in the rearview mirror.
| FOR MORE INFO: | ||
| Datasweep www.datasweep.com |
IBM www.ibm.com |
Made2Manage www.made2manage.com |
| SeeCommerce www.seecommerce.com |
Silvon Software www.silvon.com |
SynQuest www.synquest.com |
| Viewlocity www.viewlocity.com |
||
