Wednesday 5 November 2008

Model Value Analysis

Today I read an excerpt of an upcoming book titled "Business Modeling: A Practical Guide to Realizing Business Value" written by David Bridgeland and Ron Zahavi, available on OMG web site. It discusses a very simple technique to ensure that business modeling does not run into trouble due to Model Value Destruction. Many of us must have heard of failed business modeling exercises in our own or our customer organizations. There are various reasons why business modeling projects fail. These include reasons such as model value destruction, scope failures, straight through modeling, creeping complexity, ugly models, and incompetent modelers. Though the book addresses how each of these dangers can be avoided, the available excerpt provides how model value destruction can be avoided with model value analysis.

Value destruction happens when the company takes an action that has a smaller economic
benefit than the cost of the action. Similarly, building some models incur more cost than the benefits that can be generated out of them, thus resulting into model value destruction. It can be avoided by using Model Value Analysis.

Simply put, model value analysis is a summary of the expected costs and the expected benefits, and a comparison of the two, as described in a spreadsheet. One sheet of the spreadsheet should list all the possible benefits of business modeling against which both estimated one-time and annual time savings should be recorded. There are eight benefits of business modeling as given below.
  1. Communication
  2. Training
  3. Persuasion
  4. Analysis
  5. Compliance Checking
  6. Requirements for software development
  7. Direct Execution
  8. Knowledge Management and reuse
It is not necessary that each modeling exercise should save time (and hence cost) against each one of the benefits listed above though one can discover additional benefits during model value analysis. In the second sheet, one must record all time spent (both one-time and annual) for business modeling against following elements:
  1. Constructing the model
  2. Socializing the model
  3. Maintenance
While hours can be the unit to record both costs and benefits, perhaps dollar value is the better unit to use for comparison by taking into account personnel cost.

While this technique looks so simple, it may consume considerable amount of time and efforts if one employs it for making decision regarding complex business modeling exercise. A good rule of thumb is to spend 1% of the total anticipated modeling time on the model value analysis, to decide whether the other 99% makes economic sense.

BTW, after reading this excerpt, I am now awaiting the release of this book as it seems to have useful content on business modeling. Since it is being published by OMG itself, the quality has to be very good!

Monday 3 November 2008

IBM's Smart SOA

IBM is spending millions of advertising dollars on promoting what they call as smart SOA. After ignoring for so many months, I finally decided to look into it. Following are my notes that are trying to capture the essence of smart SOA without any IBM advertising!

Smart SOA is a set of guiding principles based on IBM's thousands of real-life deployments and experiences with SOA. IBM defines four approaches for adoption of SOA: Fundamental, Extend End-to-End, Transform and Adapt Dynamically. Fundamental approach involves departmental, very focused projects to meet an individual business need. IBM has defined five entry points as the means to pursue this approach viz. reuse, connectivity, people, process and information. Projects associated with Reuse entry point typically involve creating reusable service-based business components out of existing IT systems. The connectivity entry point links people, processes and information in the business through the use of an intermediary service gateway or bus. The people entry point improves productivity by aggregating services as views, as well as enabling human interaction in the context of the business process. Common projects associated with process entry point include automating and optimizing manual business processes and human workflows. Finally the information entry point provides access as reusable services to complex, heterogeneous data sources within an enterprise.

The Extend end-to-end approach involves collaboration between business and IT to optimize and bring greater innovation to business processes that span the organization and extend beyond it. BPM enabled with SOA along with maintaining integrity across the processes is a major component of this approach. The Transform approach is a broader approach than the earlier ones as it involves using technology for strategic advantage. It is about innovating the business model itself as opposed to the individual end-to-end business processes through which that business model is implemented. Finally Adapt Dynamically approach involves creating a predictive business that responds to market forces in a semi-automated, intelligent way. It embeds a belief that the emerging ideal of rapid transformation in a semi-automated way will become increasingly possible as a critical mass of service providers in the ecosystem adopt SOA to enable the business to respond with agility to business opportunities.

IMHO, the first approach is the one which will be used by majority of enterprises as of today. Few innovative companies will go for second or third approach. The fourth approach seems to be quite futuristic though it provides a grand vision for SOA adoption.