19th Century Decision Management

Copyright © 2009 James Taylor. Visit the original article at 19th Century Decision Management.Syndicated from ebizQ
John Reynolds over on the Thoughtful Programmer had a great post a little while back – 19th Century BPMS. In it he said
I sometime find it useful to describe a BPMS in terms of things and people that you probably would […]

Copyright © 2009 James Taylor. Visit the original article at 19th Century Decision Management.

Syndicated from ebizQ

John Reynolds over on the Thoughtful Programmer had a great post a little while back – 19th Century BPMS. In it he said

I sometime find it useful to describe a BPMS in terms of things and people that you probably would have found in an office or factory in the 1890s

This struck me as interesting and I wondered what analogy I might use for 19th century decision management I decided that Decision Services are like having some speaking tubes that let you talk to an Answerer. You can ask questions down the tube, typically a question about a customer, and an answer comes back. Anyone who is connected to the right tube can ask the question (with their specific data) and get the right answer (or potentially a set of options from which to choose).

At the other end of this tube is the Answerer who:

• Knows all the company policies
• Understands all the appropriate regulations
• Remembers what has worked or not worked in the past
• Knows everything there is to know about the company’s customers too (just like the corner shop of old discussed in this post on customer loyalty)
• Applies this customer knowledge to make the most appropriate allowed decision for each customer when asked about them

This is actually a pretty good analogy as nothing changes when you ask your question – you are still responsible for recording what you did – but this means you can ask questions whenever you like without worrying about side effects just like you can invoke a stateless Decision Service without worrying.

Because the Answerer know all the customers and what they did (and what was done to them or for them) he can predict what is likely to work or not work for other customers who are similar to those they have worked with in the past. When the answerer is not certain what might work best, he can conduct an experiment by trying slightly different approaches with different customers. He uses what works to update the way he behaves. Any time the regulations or policies change they get sent to the Answerer and he updates his behavior.

Substitute a standard web services interface for a speaking tube, a business rules management system for his encyclopedic knowledge of policies and regulations, data mining or predictive analytics for his customer knowledge and adaptive control for his experimentation and you have Decision Management. The Answerer but on an industrial scale.

Link to original post