10 years, 4 months ago

Back to the Future with The Theory of Constraints

In so many situations today I find business people are much more savvy with IT than they used to be only 10 years ago. And while this is a fantastic advance, the result is they are MUCH more likely to dictate the solution right from the outset. I marvel at how very senior business executives are now so conversant with the specifics of application architecture, particular packages they wish to use and Cloud deployment architecture. But of course this level of direction frequently facilitates rapid action, but without full and thorough understanding of the business issues.

We know that business people should be focused on the inherent business architecture, surfacing the opportunities for common concepts and business services, business platforms, product lines and channels; identifying where standardization and differentiation is appropriate, and where partitions are relevant, all in context with the business and market  model. Get this level of architecture right and you have the chance of delivering an agile business. Get it wrong and you are in instant legacy territory!

With this interesting problem in mind I browsed my bookshelves and came across Eli Goldratt and the Theory of Constraints (ToC). There was an AH HA moment! I first came across Goldratt nearly 30 years ago; I went to one of his lectures in London and have read many of his books, but I haven’t used the ideas for a good while.

The starting point is to develop a Current Reality Tree, initially a list and then a dependency hierarchy of Undesirable Effects (UDEs) (see redacted example below). This is used to determine the root problem and then to develop a Future Reality Tree with Desired Effects (DEs). And the techniques naturally guide the user to focus on the HOW, and separate out the WHAT. In the process I insert an Ishikawa (Fishbone) diagram between the CRT and the FRT. It’s really useful in separating Domain based issues into (people, process, technology . . . ) clusters and also teasing out the root problem.

I would be interested to hear from others whether they are using ToC, or indeed if there are other techniques that do a similar job.