DMN Adoption – Barriers and How to Break Through

Industry guru Bruce Silver recently posted essays discussing the parallels between the Business Process Model and Notation (BPMN) and the more recent Decision Model and Notation (DMN). Both standards, as he notes, were created with the goal of business user acceptance. In his view BPMN has been a success because it provided a clear, business-friendly way for users to communicate, and (with BPMN 2.0) established a means to tie the modeling and execution language. It was truly “What You Model is What You Execute.”

So where does DMN stand now? BPMN has a decade head-start so it clearly has some catching up to do! Like BPMN, it is a standard that has also established a language business users can appreciate, communicate and understand. More importantly, decisions that are rendered by the execution of rules can be expressed in another business-friendly way – via decision tables. DMN also has Silver’s second success criteria and it’s more important here. While not many BPMN models are executed, most DMN models will be. And the executable component via FEEL and S-FEEL gives us – again using Silver’s wording – What You Model is What You Execute!

Silver stopped here while just briefly touching on the vendor product market for DMN. The marketplace has seen tools come in roughly 3 flavors:

DMN compliant interfaces on some legacy BRMS platforms (e.g., Red Hat Drools and FICO Blaze Advisor)
Tools explicitly meant for decision modeling – some with integration to executable engines (e.g., Decision Management Solutions Decision Modeler)
Platforms with DMN Modeling and Execution (some integrated with BPMN Modeling, e.g., Trisotech, and some even integrated with Salesforce, e.g, Decisions on Demand)

The point here is that the tooling should not and will not be an obstacle. The depth and breadth of the vendor market for DMN is already impressive. Solutions and support are available. But what other hurdles will be faced along with education of the standard and more tooling to support it? While the promise is appealing, the barriers to adoption may seem equally immense.

For the purposes of this missive, let’s consider three different types of potential DMN clients:

Companies that have really done nothing formal with respect to business rules or process and have little to no experience with any platforms.
Companies that have rule platforms in place – perhaps many – but haven’t has as much success with them as they would have liked.
Companies we would consider major players with an extensive application base in place.

The first case should be the easy sell. DMN provides both a rigorous approach / methodology as well as a mechanism to seamlessly integrate with implementation. It’s facilitates structured requirements and engages business. Starting from scratch in this fashion can yield tremendous benefits.

In the second scenario a skeptical consumer must be educated as to why DMN is different. Perhaps previous efforts failed because they weren’t business friendly enough. Perhaps they never found an implementation that provided a smooth transition from requirements to execution. In either case it is easy to show how DMN fills the gap. It can truly be sold as a business transformation and a move from legacy systems.

Before moving to the third case, I want to discuss what I believe is one of the most important aspects of DMN that hasn’t been discussed here or in the Silver essays. A key driver of DMN is to ensure that decision models are interchangeable across organizations. Platform dependence is a thing of the past. And I can tell you from personal experience that demonstrations of this capability (creating a DMN model in one platform, exporting it in a XML format, importing into another platform, and viewing that same model) are eye-opening for most audiences. The potential power in this capability has the mortgage industry standards organization (MISMO) very close to officially naming DMN as their official standard for decision exchange. And remember - that exchange can also occur within organizations. Larger organizations with perhaps multiple legacy rule platforms in place may find DMN the missing link to finally unifying themselves!

This brings us to the third case. Convincing these experienced organizations can happen several ways. As previously noted, multiple platforms in place may be brought together via a central DMN repository. If they have no formal, defined methodology for business analysts, DMN can fill that gap. They can also be educated on the plethora of tasks that go hand in hand with Decision Management (i.e., see James Taylor and his recent discussion on how decision modeling can really bring some clarity to popular data mining techniques). It should also be noted that the majority of these DMN platforms are cloud-based. That alone may pique interest from enterprises anxious to either evolve to this type of infrastructure or use it to re-think all of their processes.

The bottom line: platforms are sufficiently mature to clearly articulate and demonstrate the power DMN brings in modeling, implementation, execution and in exchange