I was re-reading a nice piece Larry Goldberg wrote last month called Seven Deadly Sins of Business Rules and I thought I would follow-up with a similar list for decision management.
- Decisions are not identified and managed during the development lifecycle
Unless the decisions that drive your operational systems can be identified, understood and ultimately managed you will make little progress. This means making decision identification and description part of your whole requirements/process elicitation and development approach. - Business rules are buried in other models
One of the critical issues is to ensure that your business rules - the policies and regulations that you must follow - are not buried in other models. Decisions (see #1) are the way to bring rules into processes and systems but it is critical to manage business rules separately so that they can ultimately be updated independently of the rest of your systems and processes. - Traceability between requirements, use cases, business rules, decisions and processes
While separation is important, so is traceability. When a policy changes, you must be able to see which processes are impacted by the decisions that use rules based on that policy. - Data is considered as a fixed asset
The data you have in a system is often considered as though it is fixed or as though it can only be extended by capturing more data explicitly. In fact there are at least two ways to extend your data asset. Firstly you can consider external data that can be bought and integrated with your own. Examples include location data, consumer data, government data. Secondly you can mine and analyze your data to infer new data items about customers, products, suppliers. For instance your historical data might enable you to predict customer loyalty and you can then add that data item to the data available to your applications. - Performance metrics are not linked to decisions
Improving performance means making better decisions and better decisions are those that drive your performance metrics! Unless you link them you can make little progress, especially when the decisions that influence your performance metrics are high-volume, automated ones. - No ability to evolve and change
The "right" decision is a moving target - the logic you use to make a decision must evolve to meet new customer and market demands, new regulations etc. Failing to build in adaptive control is a big mistake. - Business user/IT collaboration is not established
In the end the decisions can only be kept up to date if they can evolve (see #6) and if the business can participate directly in the management of those decisions. This means IT/business collaboration must be improved. Wile technology can help with this, organizational change is critical also.
Technorati Tags: business agility, business rules, collaboration, data mining, decision management, performance metric, policy, predictive analytics, adaptive control, traceability
Subscribe to this blog by RSS
You capture some critical elements of a lengthy decision process. I've learned that the unseen but equally debilitating variable in decision management is unvoiced disagreement about the problem a specific solution is trying to solve. Unless there is agreement about the definition and cause of a problem, passive-aggressive resistance and repeated delays will add unnecessary complexity to the decision and implementation process.