“Business process management” is arguably the hottest buzzword in management theory today. But does it mean anything in particular? After all, everything in business is related to “business,” everything is part of a “process” and everything is “managed.”
So business process management seems to boil down to “pay attention to what you’re doing” – sound advice, but advice your mother long ago provided for free. And Mom even threw in a home-cooked meal. Just try getting a good pot roast from a management consultant.
Still, as Mom also pointed out, the devil is in the details. Though the importance of managing business processes is obvious, doing it well is hard. In concrete terms, business process management requires defining and executing rules so the processes are run efficiently, consistently and effectively. The critical underlying technology is therefore a rules engine.
Any computer program is a set of instructions, and thus a type of rules engine. But in a conventional program, the instructions for business policies are mixed with instructions for everything else. This means even a simple business rule change may require digging through the entire system. Worse yet, major business processes often span multiple systems, making changes still harder.
Specialized rules engines address these problems by isolating policy-level rules, making them accessible for easy modification, and letting them execute across multiple systems. By simplifying rule changes, they make it much easier for organizations to actively manage their business processes.
These goals imply several requirements. One is easy integration with the underlying rule-execution systems. This requires supporting many types of connections because the underlying systems will vary. Once connections are made, the rules engine should hide the details from end-users to simplify rule development.
A second set of requirements relates to managing the rules themselves. One goal is to let business managers construct complex rules without help from technicians. Another is to ensure that rules function as intended and are properly authorized before taking effect. There are also common administrative requirements for role-based security, backup, version control, audit trails, use reporting and such. These goals imply a sophisticated user interface that allows great flexibility within a structured environment.
A third set of requirements involves scalability. This has many dimensions including number of rules, rule complexity, number of simultaneous interactions, response time for real-time systems and records per hour for batch processes. Different dimensions will be important depending on the situation.
Generic rules engines can be adapted to any application. But users with specific needs sometimes can save time by purchasing a system tailored to their requirements. For marketers, an example is software to make product recommendations during call center or Web interactions. These systems come with prebuilt frameworks for tasks and data types peculiar to product selection. Connecting to and customizing these frameworks is much easier than setting up the same processes from scratch in a generic rules engine.
MarketSoft DemandMore (MarketSoft, 781/674-0000, www.marketsoft.com) began life in 1999 as a system for lead management. It has expanded to cover the entire demand-generation process, starting with lead creation and ending with results tracking.
Extensions include using the core rules engine for referral management (opportunities passed from one person to another) and project management (to help run lead-generation campaigns). Another extension uses licensed event-detection software to identify business opportunities. Results reporting provides multidimensional analysis and management dashboards showing MarketSoft information and external data.
Still, the core of MarketSoft remains lead management, and the core of lead management remains its rules engine. What makes the MarketSoft engine special is its ability to share rule creation throughout an organization. Fine-grained security allows precise control over which users control which rules, and a clever interface lets casual users build rules by combining predefined clauses to form English-language sentences.
These features let MarketSoft combine central control of basic lead management processes with variations based on local needs and preferences. For example, senior managers might define major policies such as lead-allocation rules, response-time standards and disposition-reporting requirements. But local sales managers could specify distribution rules within their region, while individual sales people might choose whether to receive notifications by e-mail, pager or telephone. This flexibility has let MarketSoft function successfully in sales organizations with thousands of users.
Of course, the predefined options have to come from somewhere. MarketSoft provides standard rule components for common lead management tasks such as ranking, routing, notification and disposition recording. During implementation, clients choose which components to make available and connect them to company systems. Rules can read data from such sources, which simplifies implementation and maintenance compared with the traditional approach of storing the data within the rules. If existing components do not meet company needs, new ones can be built and stored in XML files.
Unlike a pure rules engine, MarketSoft also maintains its own database. Basic structures for leads, project tasks and other major pieces of information are provided with the system, and users can add custom attributes as needed. Being able to work against a known data structure allows prebuilt rules to provide much richer out-of-the-box functionality than a generic system.
MarketSoft also provides an end-user lead management interface with functions to view, select, reassign and record dispositions on leads. Clients can use this interface by itself or integrate pieces of it with an existing sales management system.
Predesigned reports also draw on the standard tables, allowing automatic generation of standard lead management analyses such as results by source and by user. The system maintains a small dimensional data mart for detailed data analysis.
MarketSoft uses a Web-based interface on a J2EE platform. It runs on Windows 2000, IBM AIX or Sun Solaris servers and uses Oracle or SQL Server databases. A minimal installation starts around $100,000, with pricing from $300 to $600 per user based on the number of users and modules purchased. The system has more than 50,000 seats installed at 45 clients.