Six Sigma and Decision Management — Introduction
This month is the first of two articles on Six Sigma and Decision Management. I will introduce the topic this month and describe the specifics of applying Decision Management in a typical Six Sigma project next month.
Pioneered by Motorola in the 1980s, Six Sigma (6σ) is an official quality management program designed to reduce defects below 3.4 per million opportunities. More typically, Six Sigma is considered a management approach to delivering customer requirements consistently and on time. Implemented correctly, a Six Sigma program is data-driven — "management by fact" if you will — in order to implement consistent, repeatable business processes. Six Sigma measures processes and understands where potential defects are — it identifies inefficiencies. To some degree, that's also the philosophy of Decision Management. Decision Management uses data, and the analysis of this data, to help enterprises maximize revenues, minimize costs, and improve processes.
But there are important differences. Six Sigma is geared to correctness — the reduction of the number of defects per million opportunities (DPMO) — and to processes. The main goal is to look at a process from a different angle: to break the process down and examine which steps are necessary for the final result while determining the time, materials, and money that are used at each step. In addition, Six Sigma looks for ways to improve the process and reduce defects. Decision Management, by contrast, is focused on making accurate and efficient decisions, particularly operational decisions. This requires a focus on the decision making process itself and on the decision as a distinct event independent of the process context. Yet these concepts do overlap and Decision Management can be integrated with Six Sigma. Indeed, some of the disciplines and approaches of Six Sigma can be used in Decision Management.
There are a number of ways to approach Six Sigma, but the most common kind of projects go by the acronym DMAIC (Design, Measure, Analyze, Improve, and Control) projects. These projects aim to identify the cause of a problem in a process, and apply the most appropriate solution. Each of the elements — design, measure, analyze, etc. — are phases in the methodology. For a typical project you will likely identify suppliers, inputs, process, outputs, and customers (SIPOC) to get an overall view of a process. This will help identify any sources of variations, critical measures, and provide a definition of success.
Decision Management can be an important part of such a project. For example, consider a process that produces, or consumes, information products. Some processes are entirely focused on information products — credit origination, for instance — while others might produce an information product in addition to a physical one, e.g., fulfilling a book order with both a book (actual product) and a loyalty offer (information product). To apply the DMAIC approach to such a process, you must consider the parts of the process that create or transform these information products. These transformations are almost always business decisions, for example:
- Given the customer and product ordered, decide what loyalty offer to make.
- Given the application information, decide what credit line to offer.
- Given this claim, decide how much to pay.
To improve this process, to reduce defects, you must consider these decision steps and Decision Management is ideal for this. You will need a more sophisticated, or at least a different, measure of "defect." When making marketing offers, for instance, a "defect" might mean an offer not accepted. It might also include an accepted offer that lost money or a product priced lower than the optimal price for a given customer. To determine these kinds of "defects" in decisions you need to consider the precision, consistency, agility, speed, and cost of your decisions and compare these to industry norms or our own targets. Any decision that fails to meet these criteria would be "defective." Similarly, you could use the ongoing measurement of these values to track how you were doing over time as part of ongoing monitoring and improvement.
Actually improving these decisions is another challenge. If you are analyzing a decision that is being made manually then you can only analyze, improve, and control the decision by closely managing the person who makes the decision. This is not, generally, an effective approach for decisions made in high-volume, operational processes. The use of a business rules management system to automate decisions within a process has a lot to offer within a Six Sigma program. Rules-based decisions are explicit about how they were taken (which rules were fired). As such, the rules being fired can be easily measured, considered as part of root cause analysis (another important technique in Six Sigma), and systematically improved in a way a manual decision cannot be. Logging the rules fired for each transaction enables true analysis and the ongoing improvement of these decisions.
When considering information products, the use of a Champion/Challenger or Adaptive Control approach (discussed in my previous column on this topic) to compare current and possible future approaches to decision-making allows for defect comparison in advance of making a change. This enables you to see exactly how a change can impact my decision; how many defects it might eliminate; what other impacts there might be; and so on.
On a more general note, the use of analytics to compare defects has a lot to offer Six Sigma projects. While simple analysis might identify trends and other problems, more sophisticated techniques are required in decision-intensive processes. For instance, a marketing "defect" may result in a successful offer (one that was accepted) but that was unaffordable so the customer ended up in collections. Tracking this and modeling alternatives require more sophisticated analytic approaches like decision modeling and simulation.
Next month, an example of applying Decision Management in a Six Sigma project.
Don't forget that there is a lot more information on this in previous articles in this column as well as in my book Smart (Enough) Systems. You can also use RSS or e-mail to subscribe to my blog (note new RSS and blog URLs).
 James Taylor, "Adaptive Control, Champion/Challenger, and Decision Management," Business Rules Journal, Vol. 9, No. 11 (November 2008), URL: http://www.BRCommunity.com/a2008/b450.html
# # #
About our Contributor:
February 6-8, 2018
April 17-19, 2018