An EPM reader from the Carnegie Mellon Software Engineering Institute recently emailed me to tell me about a new paper from the SEI. The paper concerns software architecture, and examines the use of software architecture to achieve three goals:
- Better quality: delivery of product qualities that fulfill customer needs and expectations.
- Faster delivery: reduction of unnecessary rework that delays product introduction.
- Easier maintenance: improved flexibility for changing the product to meet evolving customer needs.
If you’re interested in software architecture and would like to read the white paper you can do so by clicking here. Note that you’ll have to enter a few details first.
One question which has has been on my mind for some time concerns the overall value of software architecture – has anyone measured the monetary value of software architecture to organizations?
Such a study would need to consider the following:
- By embedding an architecture team into the development process we reduced defects by x% saving $a.
- However the architecture team directly cost $b
- They also changed development time by +/- y% which had an associated opportunity cost of $c in the market.
- z% of the architecture was reused in the next project, impacting that project in some financial way ($d) too.
To my mind, these of just a few of the metrics which would be need to be covered by such a study. A good starting place might be to examine this study by the US Department of Commerce, which examines the cost of inadequate infrastructure for software testing.
I don’t know of such an end-to-end study but think that the methodology behind such a study would make an excellent blueprint for organizations to use to determine ROI from their architecture investments. By benchmarking their architecture ROI against other organizations in the same industry, organizations will understand how their architecture efforts compare to their peers and establish plans to address any shortcomings. The work would also help architecture teams themselves understand where their time is best spent to deliver maximum return, along with those areas where concerted efforts lead to just margin gains.
As I said, I do not know of return on investment from software architecture having been calculated in an end-to-end fashion anywhere, but if this is something you have attempted in your organization, or you’re aware of research in this area, then do write in and tell us about it.