It's getting harder for business managers to delegate. First, the invasion of desktop computers enabled them to type their own correspondence. Soon they'll be programming their own software - the programs that do the thinking for their enterprises - thanks to the critical role that business rules play in e-commerce. This change is real, it's happening now, and it means that executives will need a whole new set of skills and a different way of thinking.
Business rules are the instructions that tell software - or people - how to operate. They are a reusable set of instructions that enable an organization to operate in a consistent yet flexible manner. Ronald Ross, co-founder of the consulting firm Business Rule Solutions, LLC, says, "Business rules are literally the encoded knowledge of your business operations." "Upgrade platinum flyers to first class before gold flyers" is a business rule. So is "Most Valuable Customers are those who order at least $1,500 of merchandise in a year."
"Personalization is the killer app for business rules," says Ross. He argues that since business rules drive most of the leading personalization and e-commerce applications, companies now have a compelling reason to invest the time and effort in documenting the way they want to do business.
Instructions That Live Outside of Programming Languages Business rules are not new; what's new is that today many rules are being created and changed via interfaces accessible to non-technical business managers. Rules used to live deep within programming that was difficult and time-consuming to change. The resulting inflexibility of systems resulted in equally inflexible companies.
By "externalizing" some business rules, companies are seeking to be more adaptable and efficient as they get closer to their customers to deliver personalized service. Not every rule should be accessible to business managers, who lack the technical training-or the inclination-to manage system reliability and security, but far more should be accessible than is generally the case. "Almost every one of our clients is pursuing a strategy of having business ownership of basic rules," says Alan Crowther, CEO of LogicSpan, an e-business consulting and technology integration firm. "They want to allow business units to specify and make changes to rules quickly, without heavy IT involvement or coding."
Dealing with inflexible companies can be maddening for customers. Take a bank, for example, that requires a loan application for every customer, regardless of the customer's net assets or history with the bank. You can almost hear the assistant manager saying, "The system won't let me process your application without a completed application." She's right; that requirement is probably buried deep within the bank's programming, where it would take years to change.
The business rules approach makes it possible for such requirements to be contained in a far more accessible location. Thus, the bank could create a rule that would accommodate special circumstances, such as, "If the customer has a clean credit report and the loan is secured and the customer agrees to keep at least 20 percent of the loan's value in cash in his account, then waive the application requirement."
Transferring Knowledge From Your Head It's easy to understand the value of business rules; it's difficult to transfer the knowledge of each employee-and of the organization as a whole-into a set of effective rules. Think about the best salesperson you have ever met. Wouldn't you like to be able to capture what makes her so talented and imbue an automated system with those qualities? That's one promise of business rules, but the challenge is to translate her approach into a manageable number of repeatable principles. In this example, you might start with rules such as these:
Listen to the customer's needs before you present any products, services or offers
Present offers specific to the customer's stated needs before you show him other special offers
Show the most expensive items first, then less expensive items that enhance their value or usefulness Stew Leonard's is a Connecticut-based "dairy store" that is world-renowned for staying close to its customers and generating immense sales. In a largely nonautomated setting, the store has captured its principles in a set of rules that produce extraordinary service and a highly effective merchandising system.
One of those rules is that if an item is not properly priced, the customer gets it for free. I experienced this firsthand while buying an eight-pound piece of filet mignon for a holiday celebration. When a price didn't come up, the cashier gave it to me at no charge. I began to protest-the meat was worth at least $50-but she explained that because the company did so much volume, it was cheaper to give the item away than to hold up the checkout line. And besides, "The customer is always right."
Stew Leonard's has "The customer is always right" carved in stone at the front of the store, but for most businesses the slogan isn't quite true. Ken Molay, director of product marketing at Blaze Software (now part of Brokat), explains that business rules can also be used to balance the needs of a customer with those of your company. "The goal of personalization isn't just to give customers whatever they need, but rather to keep them happy while also ensuring that the company profits from delivering the service."
ShopIBM, a computer configurator at the IBM Web site, incorporates Blaze's rules engine to recommend products to customers. The system makes recommendations based on a combination of the customer's needs, the products IBM wishes to promote and the current inventory levels. Rich Lloyd, senior manager of online relationship marketing for Dell Home Systems, acknowledges the realities of bottom-line constraints: "It's not just what customers want. Products have different margins, and the brands have different ideas about which products should take precedence."
New Rules, New Skills It's not hard to create a simple business rule, such as "Send a thank-you email to every customer who places an order." It is difficult, however, to create and manage a full complement of rules, especially when the goal is to treat each customer differently.
As rules multiply, the odds increase that they will cancel each other out or cause unexpected results. This is especially problematic when multiple business units target the same customers. What if 10 units create rules that specify what happens when a new prospect visits the firm's Web site? Does the prospect get 10 different offers, all seemingly oblivious to each other, or does one take precedence?
Rules have the potential to accidentally produce outcomes that are undesirable or even illegal. Crowther warns, "Enterprises have to be careful that rules don't result in certain content that is accessible only to certain people based on inappropriate data. For example, a financial services site that uses location to determine which people see loan officers could result in outcomes that could be considered redlining. While no one may pursue this explicitly, algorithms that augment business rules can make this happen inadvertently."
Colleen McClintock, Jrules Product Manager at iLog Software, admits that there is a political aspect to the creation of business rules. "Rules can support the interests of the organization writing the policies," she says.
But many think this is a good thing. Business managers already have profit-and-loss responsibility, and giving them direct access to the instructions that control customer interactions should result in far more customer-focused behavior. Responsibility for rules gives the business manager more control over his or her fate.
The reality is that businesses have incredible new tools, and they have barely begun to build the skills necessary to exploit them. Few business managers have the training, expertise or demeanor necessary to create and manage business rules, even when those rules are written in plain English.
Ryan Martens, director of product management at BEA Systems, believes that critical jobs such as "Web channel manager tactician," which would have responsibility for writing the rules, haven't yet been identified on the business side. "Programmers are used to the thought process necessary to create and manage business rules, but until recently business managers never had to think in similar terms," he says.
Truth is, the language and detail involved in managing rules can make business managers' eyes glaze over. Here's an excerpt from "Creating and Managing Rules" posted at BEA Systems' Web site:
"Content selector rules construct queries on the fly and return content based on the user profile. This type of rule adds time and content components to the basic classifier rule and may use references to classifier rules to define it. It also produces dynamic queries at run-time to select content from a document collection."
This kind of language and approach works for programmers, not business managers. But business managers increasingly are responsible for defining how the firm will interact with individual customers.
Document What? When firms use business rules to automate existing practices, the first step is obvious: document the way the business operates. Members of an IT or consulting team will typically interview the business manager and document current practices. This makes life manageable: IT possesses the project-management and programming skills; the business manager understands the business and the needs of its customers.
But what happens when companies want to reinvent new business models and personalize the way they serve customers? Existing business practices aren't necessarily valid, or even helpful. Suddenly, the process becomes unwieldy, and it's no longer easy to separate tasks into neat little packages.
Business rules expert Ross suggests a three-step process to deal with such a situation:
- Record and manage the current rules of engagement with customers
- Make new or modified rules of engagement operational quickly
- Manage the rules of engagement on the business side, not the IT side. In other words, re-empower the business users to manage the rules directly
It's the second step that confounds most enterprises. Established companies are successful because they have learned to perform the same tasks repeatedly; that's what makes them scalable. But creating new rules of engagement means inventing new approaches and services, and business rules experts aren't natural inventors. They tend to come from a data- or project-management background, and out of necessity learn to focus on semantics, process and precision.
The result is a wide-open field for professionals with initiative who are willing to stretch their boundaries and develop a hybrid skill set. In practical terms, the experts on using business rules are those who roll up their sleeves and test new approaches that leverage Customer Relationship Management (CRM) and personalization applications.
Dell's Lloyd notes that theory is often much different from practice, and personalization rules are no different. "Rules get written through trial and error, by testing and retesting. We start with some sort of hypothesis, and then we work hard to isolate the confusing noise from what's really working."
For example, Lloyd's team might try an approach in which the division sends an email promoting printers to everyone who purchased a laptop but not a printer in the past 10 months. But they don't look at the result of that promotion in isolation; they track results over three months to determine whether the email reduces the response to other promotional messages sent to the same group of customers. The question to ask, says Lloyd, is, "Am I better off with this additional rule in place or not?"
One of the first challenges for these pioneers is to develop rules for interacting with individual customers. While personalization and CRM have become hot topics in recent years, many companies have only recently identified their customers, or begun efforts to do so. Until the rise of the Web, any firm that sold through a distribution channel had long been disconnected from its end users. Even firms that deal directly with customers have often engaged in monologues ("buy this; it works") rather than true dialogues.
Who Is the Customer? The first challenge in creating customer interaction rules is to agree on the definition of a customer. Author and consultant Terry Moriarty says many companies define customers in too limited a manner.
"Most information systems today have a billing view of the customer," she says. They know only the one responsible for paying the bill, but that isn't necessarily the one to whom you market the product." This isn't just a B2B problem; even in a consumer setting, children or roommates can influence when and from whom a person buys. But there's no information about those parties. "The ultimate business rule says: here's what a customer is," Moriarty continues. She believes that systems should have fields such as "involved party" and "role" (i.e., decision maker). Building such flexibility into its database enables a business to capture real-world data about how purchasing decisions are made.
In a household with teenagers, the odds are good that the teenagers have more influence than the parents over which ISP the family chooses or which computer they buy for the family room. In a business, there are always experts whose influence isn't acknowledged on the organizational chart or by the vendors' systems. But these influencers tend to be active participants on Web sites that contain facts and opinions about these services, and thus companies have the capability to identify them and build relationships.
"If the database provides separate building blocks, a firm can derive any relationship they will ever need," Moriarty says. "Each business unit can manage its line of business; the enterprise can manage the total relationship."
Who "Owns" the Customer? In a typical company, Product Manager A wants the Web site to send an offer for his product to every new site visitor. But Product Manager B wants to send an offer for her product. No one can document the correct processes, because no one agrees on them. That's why increasing usage of business rules may accelerate the pressure on firms to shift from product-management to customer-management organizational structures. If one division "owns" a certain group of customers, it becomes clear who is responsible for managing the rules that apply to those individuals.
The process of creating and managing rules highlights the limitations of product-management structures. These organizations were logical when the toughest challenge an enterprise faced was creating and selling high-quality products. But today the toughest challenge is managing relationships, often in real time. If divisions are competing internally, they make decisions that aren't in the customer's interest.
Alan Crowther observes, "Organizational barriers mean that personalization objectives are often compromised to individual business objectives. Even though it may make sense to personalize a site so that certain products or services are not highlighted, a business unit may have the clout to demand it."
Of course, the real loser is the company that remains burdened by an organizational structure that is illogical in today's era of real-time interactions. Most customers are starved for time and increasingly unwilling to tolerate treatment that doesn't make sense to them. The bottom line: Creating rules for e-commerce and personalization is much easier when one person or business unit has clear ownership of a portfolio of customers.
Even with such clarity, large enterprises are still, well, large. So business rules need to be created using a common vocabulary. Ron Ross explains: "You must be able to trace who's using what terms in what context; you need to be able to trace the impact across the business environment." While business rules can theoretically be created using everyday language, each term must be used with more precision and consistency than we use in normal conversation.
Another issue in large businesses is traceability: being able to follow the impact of certain rules across the entire business environment. "You need to know what task each rule relates to," says Ross, "and you also want to be able to trace rules to the parties that have a stake in a particular rule. When it comes time to change the rule, you must know whom to call."
What to Expect The more you study the use of business rules to drive personalization, the more obvious it becomes that the approach has tremendous potential but is still in its infancy. While there are no tried-and-true best practices, it is possible to offer some conclusions about what's likely to work best.
Minimize the number of enterprise-wide rules. Only about 2-3 percent of all business rules are "core rules" that impact the entire enterprise and specify qualities that give the firm its identity and differentiating qualities, Ross says. These rules usually existed long before the advent of automated systems.
Here's an example that used to apply at Domino's Pizza: All pizzas are delivered within 30 minutes or the pizza is free. This rule drove the company and its marketing messages, but it eventually was changed after communities began to question whether the pressure on drivers to deliver so quickly was resulting in serious traffic accidents. Note that rules such as this need to include two portions: the rule itself, and what happens if the rule is violated.
Use the simplest approach possible. The more complex an approach, the more likely it is to fail. The best approach is to ask, "What are the simplest changes that will have the most powerful positive impacts on customer relationships?" One practice to consider is that all databases containing customer information have at least one field with a common, firm-wide format, enabling the firm to link data. This is a much simpler approach than building a data warehouse or requiring all business units to conform to more restrictive data conventions.
Combine business rules with other approaches. Systems that match business rules with specialized data mining, domain-specific algorithms and other approaches achieve better results than through business rules alone, Crowther says.
Black Pearl, a personalization software company that specializes in the financial services industry, is a good example of this multipronged approach. Its Knowledge Broker system combines business rules with a recommendation system that is powered by data-mining technology and an ontology that enables communication between multiple computer systems. Ontologies provide a consistent, logical model for accommodating multiple data sources and goals. Done right, they provide a flexible system for understanding and managing interactions with individual customers.
Maintain a separate rule base. A rule base is essentially a database for rules. By separating business logic from specific applications, rules can be shared by numerous applications. This approach speeds development and increases flexibility.
As the number of rules grows, it becomes increasingly important to be able to test in advance what impact the addition of new rules will have. This task is easier when all the rules are stored in a common location and format.
