The Business Value of (Effective) Architecture – Part 2
In my last blog I used the Standish group Chaos study to show how large projects have a much greater failure rate than small projects and that most of the reasons for the failures can be addressed by an effective EA capability. In today’s post I will explain the conclusion that the business value of EA is $72.7M for an organization that has an annual capital budget of $100M.
One of my favorite books at the moment is How To Measure Anything by Douglas Hubbard. I’ve long been a fan of evidence-based continuous improvements and as a result have always been looking for, and frequently found, ways to measure things that others thought was impossible. More recently, in my role as Chief Architect for Informatica’s Business Transformation Services team I’ve been asked multiple times by clients to help them articulate the business value of their architecture program. We certainly know how to measure the value of a specific business initiative for which we architect a solution such as the value of a customer master data management solution that helps cross-sell products, reduces call center handling time and saves mailing costs by eliminating duplicate accounts for the same customer. But that is measuring the value of the MDM solution and not the value of architecture.
So here’s how I went about to measure the value of EA. I started with the premise that if we had perfect information and did perfect planning, each project would cost exactly what we projected, it would finish on time, and it would deliver the expected business results. If we take a simple scenario of a $100M annual budget for capital projects and assume that on average the projects take 12 months, have a 12 month payback (business benefits in the first 12 months of operation equal the initial project cost), and have a maintenance cost of 20% of the original project cost, and the cost of capital or the internal rate of return is 10%, then the net present value (NPV) of the $100M investment is $143M which represents a 43% ROI over a 5 year period.
I call this the Perfect Project Portfolio. Note that any of the assumptions in the model can have a significant impact on the results. For example, a discount rate of 6% rather that 10% increases the ROI to 71%. The point is that if things go as we expect, the business benefits are quite significant.
So let’s apply the same model to the CHAOS survey data. Remember, the Standish Group surveyed thousands of projects around the world – so these reflect the “normal” or typical results. If we plug in the data that (on average) reflects the reality that only 39% of projects are successful, that 18% fail and 43% are challenged, the numbers are very different. And to be as accurate as possible, the model I created assumes that 1/3rd of challenged projects run over budget by 189%, 1/3rd are late by 222% and 1/3rd are missing 39% of the features. The 1/3rd division is my assumption – but the other percentages all come from the CHAOS data. The reality is that some projects that are not failures actually experience two or three of these challenges so the reality may actually be worse.
When we crunch the numbers the results are mind boggling. If your organization is typical, then the net present value of $100K is only $23K which is a negative ROI! In other words, the normal kind of chaotic planning is actually destroying, not creating, business value. The net annual difference between typical/chaotic planning compared to “perfect” planning is $120K. Finally, if we look at reasons for poor results in the Chaos servey, the top 5 reasons are directly impacted by proper EA planning and represent 60% of the impact. So when we do the math, the value of perfect planning on an investment portfolio of $100K creates additional business value of $72.2K.
Please remember this analysis is conservative from several perspective. For example, we are assuming that only the top 5 root cause reasons for project failures can be impacted by an effective EA program while in reality EA has an even broader impact.
Also, please keep in mind that this does not reflect the value of any or every EA program – only an effective one. A good architecture will ensure that the defined solution is aligned with corporate product/channel/brand strategies, improves existing business operations to create a seamless customer experience, consolidates or integrates with other IT systems and leverages or re-uses existing technologies and design patterns. In other words a good architecture takes a holistic enterprise perspective and considers all possible impacts and at the same time creates a solution blueprint that is delivered quickly, on-time and on-budget.
For a good architecture practice, check out the BOST™ Toolkit. I first learned the approach 15 years ago when I was at Best Buy and have been a fan ever since. BOST is now available from Informatica so I am bit biased – but not because I work for the company, because I’ve experienced BOST and I know it works!
 Douglas W. Hubbard, How to Measure Anything: Finding the Value of Intangibles in Business, 2014, John Wiley & Sons