How companies can rev up rules engines to drive up profits.
For the rest of the March 2002 issue of CRM magazine please click here
It's not enough today to just be nice to customers. It is now necessary to treat different customers differently--for their benefit and for the benefit of your company. There are many technology tools to help accomplish this, but one of them has recently garnered renewed attention: rules engines. Rules engines are as old as mainframes and about as taken for granted. Software that works with quiet effectiveness in back-office enterprise applications, rules engines are designed to automate business procedures. Yet with the emergence of CRM's strategic importance, rules engines are today evolving from team oxen to show ponies as business executives finally understand how dramatically they can contribute to a CRM application's ROI by delivering significantly strategic improvements in sales, marketing, and customer service.
"CIOs are keenly aware of what value these rules engines can bring," says Steven Jeffes, a CRM practice director at PricewaterhouseCoopers (PwC). Apparently so are a growing number of senior line managers expected to deliver rapid investment pay back on any IT project as well as other strategic financial benefits where they may exist. The reason? Rules engine technology is a concept easy to understand, and though the technology is invisible, its performance potential in CRM software is quite visible.
A simple example of this technology exists in such disparate industries as telecommunications, financial services, and consumer products. Companies use a CRM call center strategy called skills-based routing to improve customer satisfaction by better aligning the customers calling in with the most appropriate customer service rep. Although this is known as skills-based routing, it is rules engine technology that make this possible. Upon a customer's first point of contact, she is identified by phone number or is tagged by problem type as revealed in the choices she makes wading through the layers of prompts. Companies might match phone numbers against a customer profile then, based on rules engine routing, send the call to the proper customer service rep. There are two reasons for matching reps and customers. The first is to cut down customer service costs, because the rep with the proper expertise captures the call. The second is to increase customer satisfaction among a company's best patrons by linking them with the best reps in the organization.
"Most of the focus [today] is on segmenting the level of service based on the value of the customer," says Alan Crowther, managing officer of Adjoined Technologies, a Miami-based CRM consulting and integration firm whose clients include AT&T Corp. and American Express Corp. "This is an increasingly common tactic," he says.
Companies are looking to spend more on their better customers and less on lower value customers, according to Crowther. As a result these companies save on the cost of customer service, because they are only delivering a high level of service to the customers who provide most of the profits. Differentiating service based on customer profitability also is an effective way to define which customers are most valuable, which have the most potential for growth, and which might best be politely ushered to a competitor.
This strategy is demonstrated in many call center and customer service CRM applications. Embedded in these applications are a series of if-then programmatic statements, which are the foundation of any rules engine. One rule might be, if the caller is one of our best customers--defined by lifetime value or contribution margin--then send her to Hotshot Hank, our best customer service rep. The company rule or strategy is to differentiate the customer service experience. The rules engine houses all the data points or profile information of both the rep and the customer, which will be parsed at the time of call routing so the rule is obeyed.
Beat the Clock
Rules engines can also drive automated customer service, allowing support personnel to resolve a problem before a customer even knows she needs to call.
PwC's Jeffes tells the story of a commercial air conditioner manufacturer that uses a customer service CRM implementation to support service contracts with customers. Commercial air conditioners contain sophisticated self-diagnostic tools to monitor unit operation and detect impending system failure. The manufacturer created data links for remote system monitoring and when an operational problem arises--water or refrigerant chemical spills or a fan or compressor unit is in need of attention--the CRM application will alert building facilities management or a service company of a pending failure. Rules engines are used to route alerts to the appropriate party. A simple problem might be sent to the building's facilities management department, a more complex problem to a service organization. This alert can appear, for instance, as either Web-based information accessed by a customer service rep based in a call center or sent via a wireless device to a field representative--accelerating response times or intervention before a complete system failure.
It is possible to further refine the rules determining who receives the alert and subsequently what action should be taken based upon customer value--nature of the customer service contract, credit history, level of business transacted with the manufacturer, active warranty, or extended service agreement, Jeffes says.
"Some of these rules are not very sophisticated at this point, some are perhaps several conditions deep with a couple of people messages might be sent to," Jeffes says. "But as these CRM platforms become more robust, the decision making capabilities behind them is going to grow dramatically." The manufacturer improves performance on two levels. Mirroring the call center example, the air conditioner manufacturer aligns customers with the appropriate service levels, saving it customer service expense. Second, quick routing of diagnostic information raises the manufacturer's service performance and customer satisfaction, because costly air conditioning system failures are avoided. Jeffes says the company measured a 15 percent reduction in system failure of certain units by virtue of earlier intervention into machine performance problems. The consequent increase in customer satisfaction is invaluable.
Improving Channel Performance
SolidWorks, a $90 million software developer of CAD applications directed at companies designing mechanical products, is also realizing this kind of business process improvement. The company, which and sells through value added resellers (VARS), began shopping in the spring of 2000 for a CRM system to help better manage its sales lead distribution system. Within weeks it chose MarketSoft's eLeads lead management product, DemandMore Leads, because it was drawn to the application's rules engine capabilities. "eLeads is used for lead distribution and management, but we're using it a little differently," says John McEleney, chief operating officer of Concord, Mass.-based SolidWorks. "We're trying to control and influence behavior. Part of the earliest discussions with MarketSoft was asking what kinds of rules we could put into place, what kinds of behavior modifications [with VARs] could we force."
The behavior modification McEleney sought was simple. He wanted to improve communications with VARs about lead status. Before eLeads the company would gather sales leads generated from the Web or direct response mail in a spreadsheet and send them weekly to resellers. Often there would be a long lag time between lead capture by SolidWorks and when a VAR would actually follow up with a potential customer. And SolidWorks had no way to track the status of any individual lead through the sales cycle. "There was never any confirmation that in fact the VAR received these leads, so there was no handshake if you will," McEleney says. "It was difficult to understand the effectiveness of [our] marketing programs."
To correct this SolidWorks created a set of rules that would compel VARs to follow up on leads and communicate their progress back to SolidWorks. The first rule is the most basic: Anybody requesting product information from the company's Web site who is already a licensed user of its software becomes a lead that goes to the original VAR, a referring process handled in SolidWork's Onyx CRM database. One of the worst sales mistakes a company can make is to introduce an existing customer to a new VAR, McEleney says. The customer is confused and the VAR of record is furious.
The rest of the rules pertain to new leads. A new lead engages a collection of nested rules across multiple decision points to determine where that lead should go. These decision points include whether the contact is an existing customer that has not yet been assigned a VAR; where he is based; and with which VAR does SolidWorks want to solidify a sales channel relationship.
The lead is then sent to the chosen VAR, who either accepts or rejects it. If two days pass and the VAR has not confirmed that it is possession of the new lead, then the lead is kicked to a higher level of the decision tree supporting the software rules engine in its choice of the next VAR in line to get the lead. Eliminating dead time between lead routing and capture confirmation improves sales channel communication, one of the most important elements of operational performance to SolidWorks, McEleney says.
SolidWorks has improved channel communication and management in other ways, as well. The company selects a quantity of leads to send per VAR, determined by a number of variables including the amount of past business the reseller has done with the company, as well as whether SolidWorks wants to intensify the level of the existing relationship. If the VAR wants the next lead--whether it be the sixth or eleventh, etc.--they must update the company on the status of the other leads by logging into eLeads applications from Web browsers and entering lead information, even if there is nothing new to report. The rule embedded in a rules engine is, if the next step in the sales cycle for any given lead has not been updated, do not send that nth lead.
"The more we can increase the frequency of feedback, the more we can refine the system," McEleney says. With such a high degree of communication with VARs, SolidWorks will be able to track the effectiveness of marketing campaigns. "If we found a marketing campaign directed at medical device manufacturers has moved the potential customer from initial contact to demo much quicker versus a campaign directed at mechanical machinery makers, that might tell us that the buying proclivity of medical guys shows they have a greater need for our product," he says.
SolidWorks' rollout of its rules engine is confined today in the Northeast, but McEleney hopes to replicate this CRM strategy with other sales offices when time permits. "Our IT resources are massively strained because of our continued growth," he says. McEleney estimates at least a 5 percent increase in the velocity of the sales cycle due to the new VAR incentives that the rules engine enable.
Take a Load Off the Sales Force
Fidelity Investments also sought performance improvements using rules engines. The company, which sells financial products, uses rules engines in its eLeads implementation to more effectively manage lead routing out of several call centers employing more than 5,000 people. These leads are sent to the 80 regional, retail branches, where employees develop face-to-face relationships with customers. Branch managers do not know how many leads will come their way at any given time, so if the branch office is experiencing heavy foot traffic, eLeads software will set a kind of circuit breaker. If leads exceed a predetermined limit for any branch, they are rerouted back to the home office or to one of Fidelity's call centers.
Juggling lead distribution in this way allows branch employees to balance their time between customer and potential-customer traffic walking through the door and leads brought in over the network, says Garrett Sickles, vice president of program management at Fidelity.
But efficient routing is not the only way in which Fidelity uses eLeads. Rules engines might also kick in a particular cross-sell opportunity for a call center rep. A rule might be, if the data repository detects that the caller has a new zip code, then find out if this move is because of a new job. If yes, then suggest an IRA rollover. "This presents the opportunity to have a highly relevant cross-sell discussion, or at least make sure we follow through with customer service," Sickles says.
Fidelity implemented last November what Sickles believes is one of the more ingenious features of the software: the ability for business managers to edit rules on the fly if in-branch business conditions warrant it. Theoretically a branch manager could edit the rules--no programming required--to control lead handling within their branch. "We can centralize some of the core rules, but distribute others so the power rests in the hands of the businesspeople," Sickles says.
Until now a branch manager could not change the centralized rules that determine which branch gets which leads. These rules are based upon several data points, including where the person lives, amount of money the person wants to invest, type of product, and age or life stage. But once those leads are sent to the branch, the manager can decide what to do with them next. "We will give them a base set of rules and they can choose to modify them within that given pool of leads," he says.
Fidelity also leverages rules in a custom marketing application. In a classic example of marketing automation CRM, existing customers logging into its Web site might see customized content generated within their Web browsers communicating a specific offer. The rules engine parses who is eligible for what offer based upon very targeted criteria, such as profile information including transaction history. If the specific visitor falls within the audience segment, a content repository splashes the offer. Sickles says the company has measured increased revenue from these rules-driven market efforts, but declined to provide specific numbers.
Whether improving the customer experience, streamlining company processes, or encouraging communication, rules engines are versatile and effective. The process improvements rules engines can bring often directly correlate to the bottom line.
John Berry is principal of according2jb.com Inc. He can be reached at firstname.lastname@example.org.
Playing by the Rules
Even rules engines have rules. Here are six musts for using rules engines effectively.
Coding rules in rules engines is easy. Defining business rules is hard. "Defining the rules is really the most challenging part," says Garrett Sickles, vice president of program management for Fidelity Investments. "What we're leveraging is the tremendous amount of customer and business knowledge from various people inside the organization." He estimates it took a couple of months to develop consensus among business managers as to what rules should be embedded in its rules engine. Managers trying to define business rules should remember that rules are fundamentally about strategy. Rules engines simply automate strategic actions of the company.
Rules engines are only as good as the rules. "If you're cross-selling and you're trying to pitch the right offer, the tool is only as good as your demographic and your understanding of how to segment customers and how to pitch to each profile need," says Alan Crowther, managing officer of CRM consulting and integration firm Adjoined Technologies. Rules engines are worthless if the customers segments and associated offers are misaligned.
Look for places in the organization where consistency can improve customer relationships. (Read: everywhere.) Consider call center personnel. A company might decide that it wants to treat all customers who do a certain dollar volume of business a certain way every time. Leaving the treatment of the customer up to the ingenuity of the call center rep without any structure is a prescription for inconsistency. Managers need to create that structure.
Implement CRM applications that provide analytic capabilities that tie rules engine usage back to business benefits. "At the executive level people will be focused on the rules engines that they can tie back to measurable and quantitative improvements in performance," Crowther says.
Develop an environment where business rule discovery and mining are performed as part of normal business operations. Companies should get in the habit of reviewing existing rules for their continuous operational contributions and learn and deploy new ones. It is also a good idea to pilot test rules among various consumer segments before moving to full-scale deployment. This will maximize rule effectiveness and minimize negative consumer reaction.
Rule flexibility is important. Since rules and rules engines are about automating business processes that make sense to that business, line managers should have some control over rules that affect them. "You want to implement a package that lets the businessperson define the rules without a technology department getting involved," Crowther says. Code tests and deploy cycles aren't processes business managers should have to concern themselves with. --J.B.