By: Rajesh Dontula, Consultant, HP Applications Services
Have you ever seen a developer rating the complexity of applications that he/she supports as anything less than "Highly Complex"? Of course not! Every developer considers the application that he/she supports to be highly complicated if not the most complex. In fact, they might even boast and say the other guys' applications are downright simple or easy. Why is that the case? The answer boils down to two words "Subjective Bias", essentially a qualitative assessment of an application.
Let’s assume you’re the CIO of a company. You are trying to figure out the complexity of your applications portfolio. This qualitative approach of asking the developers to rate their applications is just not going to cut it. You need a quantitative approach - an ability to assign a numeric score to an application so that you can objectively compare one application, or one set of applications, to another. This quantitative approach provides the clear view and new level of understanding into your portfolio that you've been looking for.
This is where HP can help. We analyze each application and assign a rating for several factors, including the following key dimensions:
- Size - measures the size of an application from being very small to large
- Usage - measures the usage level of an application from low intermittent usage to very high usage
- Volatility - measures the amount of change that the application is undergoing
- Criticality - measures the criticality of an application in support of your business processes, whether a mission-critical application or a mundane application
- Complexity - measures the technical architectural complexity of an application
We input these high level dimensions or, depending on data availability, input more detailed component metrics into our proprietary classification model to arrive at an application classification score. You can then use the classification score to objectively compare and analyze your entire applications portfolio. And, you now have the ability to track these metrics and the application scores over a period of time to perform a trend analysis.
While no method is perfect for all applications, such as applications that defy such quantitative classification, our approach is valid, applicable and provides useful insight for most of your applications portfolio.
In my next blog post I’ll suggest some ways to use the classification scores in developing business cases for improving specific applications with a view to reducing the overall complexity of your application portfolio. Stay tuned!