If it makes sense to try to bring your BPM and enterprise architecture (EA) efforts into alignment, you'll want to consider several key issues.
COMMUNICATION COMES FIRST
Fundamentally, in integrating BPM and enterprise architecture, the main thing you need to consider is how the two will communicate, notes Rodrigo Sousa Coutinho, senior product marketing manager at OutSystems, an agile software-development company.
For example, he says, if you've got an application that manages expenses, and you also have a well-defined process for expense approval, the BPM engine must be updated when a manager approves an expense.
In a perfect scenario, both the application and the process are tightly integrated, not only at run time, but also during design and development, he explains. “This is the only way you can be certain that changes on the application won’t break the process, or that the process doesn’t assume a flow that’s not supported by the application,” he says.
“BPM fits in EAs at the business layer, something that’s present in most EA definition frameworks, so there’s a good fit there,” says Coutinho.
Returning to his original point, Coutinho says the main issue in terms of integrating BPM in an EA is the communication between the layers--for instance, between the business and data layers, or between the business and application layers. “It really doesn’t matter if you designed a very optimized process for your organization if the data and applications aren’t able to support that process. That’s why it is important to make sure there’s a tight integration between BPM, the data, and the applications,” he says.
TRUE MERGER REQUIRED
In most cases, though, continuous business improvement cannot happen without effectively merging the holistic planning aspects of EA with the process improvement focus of BPM, notes Sean Narayanan, chief delivery officer for the iGate Patni consulting and outsourcing company. "We need to work smarter throughout the enterprise, transforming organizations so people can make more informed decisions, build deeper relationships, and work with more agile and efficient business processes,” he says.
However, working smarter requires more than simply aligning efforts, he warns. It requires a deep understanding of the business processes of the enterprise and the ability to execute change on these processes by collaboration between business and IT. “This melding of planning and delivery is exactly where BPM and EA are strongest when positioned together,” he says.
The effort can pay off because BPM can bring a great deal of insight to EA, even at the design stages. Coutinho cites information architecture as one area that can benefit significantlyf from systematic analysis of an organization’s processes.
METRICS, MONITORING AND MODELING
But one of the biggest values of BPM, when properly connected to the applications that support it, is that it allows you to measure the different steps of a process quite accurately. “This gives you an insight to your process strengths and inefficiencies, and uncovers things that you can adapt or improve, that before were blind spots,” he says.
That means that after a process is put in place, the work isn’t finished. You need to monitor that process and look for opportunities to improve, both at the BPM level and at the EA level.
Coutinho says models and modeling can be helpful to all layers of EA. At the business layer, it's much easier to communicate with your organization about how the processes work. In other words, diagrams are more effective than a simple textual representation. Even when you’re thinking of things that are usually owned just by IT, such as the data and applications, using models can provide much greater efficiency, he says.
“The big win with models is that they are so much easier to read and understand than a textual language," Coutinho says. "And this knowledge transfer is a huge deal, because the cost and time span of maintenance is much bigger than the cost and time span of creating an application." Even when the people involved change over time, the model will persist--and a model provides an important edge when it comes to learning what a system does.
BEGIN WITH BABY STEPS
Coutinho says the best way to begin is to aim big, but start small. “Be agile. Pick a problematic process in your organization, or a process you believe can be highly optimized, and start there,” he says.
Of course, you need the right tools to support such an approach. For that reason, he recommends making sure that you pick a tool that is strong in delivering through an iterative approach, in terms of process, application and data design, and that tightly integrates the process with the application that employees are using.
You also need to check the analytic capabilities of the tool you plan to use. “One of the great things about BPM is the ability to optimize the process as you see it running in the real world,” he says. You need to be able to quickly understand what tasks are blocking the process, if it’s an internal or external problem, so that you can optimize your processes and EA accordingly.
Drew Rockwell, CEO of the MDS Group, which focuses on enterprise process analytics, echoes that point. In Rockwell's view, any effort to integrate BPM and EA probably should include a role for analytics, too.
“In the system design phase...you need to analyze your existing process because you are expressing how things are done today in your BPM system,” he says. Then, further along, Rockwell says you need to make sure that everything works as expected, and that you can answer this question: “Once that architecture is in place, how do you ensure that it is operating accurately in an environment with continuous change?”