sommai - Fotolia
If digital transformation is the process of optimizing business operations through effective use of IT applications,...
it's hard to see how it can be done without enterprise architecture. However, three of every four businesses don't have formal EA, and almost half of those still don't use it effectively. Making the most of EA in your own transformation and app modernization planning starts by reviewing the three critical areas where EA models can help, taking a systematic path to addressing these areas and integrating EA not only into software architecture but into application lifecycle management (ALM).
Transformation and modernization are evolutionary terms, and it's easy to forget that they not only have to start with the present state of IT in a business but move forward in harmony with business goals and technology tools. Too much focus on the tools-and-how process risks losing sight of the goals-and-where-to, and only EA practices can restore the balance. Given that most businesses don't have formal EA modeling or enterprise architects, it's smart to look to the principles of EA, exploiting what you have in formal modeling but not depending on it.
Most IT professionals will admit to a kind of tunnel vision with respect to software projects. These projects normally target a narrow business function, and often, they're driven primarily by technology change. That creates three specific problems with transformation and modernization projects, any of which are enough to derail the project completely if not addressed.
EA project concerns
Problem number one is the failure to recognize common requirements across multiple operating departments. This happens because transformation and modernization projects take aim at specific changes, usually generated by a single organization at a time. Without a broader vision, you can end up doing the same thing several times and inconsistently. Problem two is selecting a path for change without considering the future. Business trends drive IT requirements trends, and if you don't have a source of the former, you'll probably take too short-term a view for the latter. Problem three is failure to address the complexity explosion that usually accompanies transformation and modernization. Many surveys have shown that transformation and modernization of apps will increase the interdependence of line departments, and the extent to which information has to be integrated follows that same path.
All of these problems can be addressed by taking a business-goals-and-benefits perspective on projects, but that requires a systematic view of business goals and how they translate into operating practices. EA models and the people who maintain them aim to link the structure and operation of a company (usually on a per-organization basis) with business goals and objectives. In most cases, this is done by framing business processes as the specific operations practices that achieve these goals. These processes are then the source of requirements for IT organizations, which provide the computing and networking tools.
Since it is the common business processes that drive converging strategies for transformation and modernization, organizations without formal EA models (or even EA personnel) can achieve some of the benefits of EA by reviewing the way that workers are currently empowered, seeking out common business processes that create opportunities for a shared technology strategy. Common processes should be supported, as much as possible, with common architecture and tools.
Trends in EA models
Evolutionary trends are harder to uncover by looking at current empowerment strategies. In EA, evolutionary trends are usually reflected in the changes made to the EA model over time and in response to business changes. This may not be captured fully, even where EA modeling is done, unless records of past model states or major drivers of change are maintained. If you have an EA team and model, make sure this information is captured and maintained.
If you don't have formal EA or have failed to capture the critical trending data, look at the record of application changes across the business, and track the impetus for these changes back to the business requirements. That should provide you with a trend line that you can then project forward. When in doubt, assume that future applications will have to present more information to workers at a given point and that the information will have to be formatted flexibly to support per-worker customization.
Just getting an EA-driven transformation and modernization plan into place won't insure it will work. Enterprises have found that, even when EA input is available to address those three critical areas covered above, it's easy to lose your way during development and easier still to lose it as applications are changed over time. You'll want to ensure that the documentation you prepare on how business processes are driving overall IT evolution is maintained and fed back into the ALM cycle, as well as into new projects.
One good way to do this is to have EA cooperation in preparing the testing procedures and test data needed for ALM. Testing should push the boundaries of business practice evolution, particularly in terms of transaction volume and mix. Workers tend to gravitate to resources that are helpful and avoid those that are not, and that can invalidate historical data on what specific information they access and how often. EA input offers a clue to how worker behavior will change with transformation and modernization tools available, and testing can validate application performance under the new conditions.
There is little sense talking about transforming or modernizing applications if you can't ensure that the applications you develop will meet your company's business goals throughout their expected useful life. Linking goals to business processes and business processes to application requirements is the only sure way to accomplish that, and it's a path that EA is especially designed to follow. In fact, it's unlikely that you can succeed without it.
Top five DevOps steps for EA
Quiz yourself on enterprise architecture models
Connecting EA with DevOps
Discover more about what digital transformation tools and strategies actually mean