An Introduction to Decision

Modeling with DMN

Building a Decision Requirements Model using the new
Decision Model and Notation (DMN) makes for better
business analysis and improved system requirements.

The Object Management Group has recently adopted the Decision

Model and Notation standard. Here’s why you should use it to
define your decision requirements.

Why Model Decisions

Experience shows that there are three main reasons for defining decision
requirements as part of an overall requirements process:
1. Current requirements approaches don’t tackle the decision-making that is
increasingly important in information systems.
2. While important for all software development projects, decision requirements
are especially important for projects adopting business rules and advanced
analytic technologies.
3. Decisions are a common language across business, IT and analytic organizations
improving collaboration, increasing reuse, and easing implementation.

Gaps in Current Requirements Approaches

Today organizations use a variety of techniques to accurately describe the
requirements for an information system.
Most systems involve some workflow and this is increasingly described by business
analysts in terms of business process models. Experience shows that when process
modeling techniques are applied to describe decision-making, the resulting process
models are over complex. Decision-making modeled as business process is messy
and hard to maintain. In addition, local exceptions and other decision-making details
CONTENTS can quickly overwhelm process models.
By identifying and modeling decisions separately from the process, these decision-
Why Model Decisions
making details no longer clutter up the process. This makes business processes
An Overview of Decision
Modeling simpler and makes it easier to make changes.
Suitable Decisions for Describing decision-making as a task in a process (or as a step in a use case or as a
Decision Modeling requirement) can result either in long, detailed descriptions that are confusing and
Getting Started with
Decision Modeling
References

© 2016 Decision Management Solutions 1

 An Introduction to Decision Modeling with DMN

contradictory or short descriptions that lack the necessary detail. All this has to be
sorted out during development, creating delays and additional costs.
By modeling the decisions identified, a clear and concise definition of decision-
making requirements can be developed. A separate yet linked model allows for
clarity in context.

Decision modeling is a new technique in the International Institute of Business

Analysts (IIBA) Business Analyst Body of Knowledge (BABOK® 3.0).

Business Rules & Advanced Analytics Projects

Successful business rules and analytic projects begin by focusing on the decision-
making involved. For business rules projects, clarity about decision requirements
scopes and directs business rules analysis. Today, business rules analysis can seem
never-ending, with teams trying to capture all the rules in a business area. The result
is often a “big bucket o’rules” that are poorly coordinated and hard to manage.
Instead, by understanding which decisions will be made, when and to what purpose,
it is now easy to tell when business rules analysis is complete.

“Decision modeling enables us to model our business by dividing it into concrete parts that are
understandable to business people without being too detailed. It also helps us not to lose sight of
the overall picture of the process while delving deep into the details of business rules.”
Timo Laukkanen, Process Director, Finnish Tax Administration

The decisions-first decision management approach begins with the higher level
concept of the decision. This provides:
 A higher, overarching principle to impose some business oriented structure on
this complexity.
 The separation of a declarative definition of these rules from the sequence-
oriented business process – improving both.
 A structure that can be expanded in a series of iterations, allowing progress to
be made in a more agile and less waterfall approach.
Specifying a graphical decision requirements model based on the DMN standard
provides a repeatable, scalable approach to scoping and managing decision-making
requirements, making it easier to:
 Draw the automation boundaries.
 Re-use, evolve, and manage rules beyond the first business rules project.
 Consolidate business rules across multiple implementations and platforms.
 Assign ownership, governance and sources appropriately.

© 2016 Decision Management Solutions 2

 An Introduction to Decision Modeling with DMN

Decision Management streamlines and focuses business rules projects for faster,
more effective deployment. Learn more in our white paper, Maximizing the Value of
Business Rules.

For advanced analytic projects a clear business objective is critical to success.

Evidence is growing that specifying this objective in terms of the decision-making to
be improved by the analytic is one of the most effective ways to do this. Established
analytic approaches such as CRISP-DM stress the importance of understanding the
project objectives and requirements from a business perspective, but to date there
are no formal approaches to capturing this understanding in a repeatable,
understandable format. Now business analysts have the tools and techniques of
decision requirements modeling to identify and describe the decisions for which
analytics will be required. How the data requirements support these decisions, and
where these decisions fit, is clarified and the use of analytics focused more precisely.

The value proposition of analytics is almost always to improve decision-making.

Read our white paper, Framing Analytic Requirements, for more information.

In both cases, then, it is essential to first define the decision-making required and
only then focus on details like the specific business rules or predictive analytic
models involved. Specifying a decision model provides a repeatable, scalable
approach to scoping and managing decision-making requirements for both business
rules and analytic efforts.

“This is the critical path to monetizing advanced models.”

Head of Analytics, North American Insurance Company

A Shared Framework and Implementation Mechanism

Decision modeling provides a framework that teams across an organization can use.
It works for business analysts, business professionals, IT professionals and analytic
teams. Decisions are also easily tied to performance measures and the business
goals of a project. This makes it easier to focus project teams where they will have
the highest impact and to measure results.
Many business analysts have known all along that decisions, and decision-making,
should be a “first class” part of the requirements for a system. Systems that assume
the user will do all the decision-making fail to deliver real-time responses (because
humans struggle to respond in real-time), fail to deliver self-service or support
automated channels (because there is no human available in those scenarios) and fail
front-line staff because instead of empowering them with suitable actions to take it

© 2016 Decision Management Solutions 3

 An Introduction to Decision Modeling with DMN

will require them to escalate to supervisors. What business analysts have lacked
until now is a standard, established way to define these requirements.
Enterprise Architects meanwhile are chartered with fitting business rules and
analytic technologies like data mining and predictive analytics into their enterprise
architecture. A service oriented platform and architecture, supported by integration
and data management technology does not have obvious “holes” for these
technologies. Decisions are both the shared framework and the technical
mechanism to easily implement these technologies.
Decision modeling is a powerful technique for business analysis and for enterprise
architecture. Using the standard DMN notation to specify Decision Requirements
Diagrams and so specify a Decision Requirements Model allows the accurate
specification of decision requirements. Decision Modeling is also central to Decision
Management, a proven framework that ties business rules and analytics to business
objectives.

Most process models today are developed using the Business Process Model and
Notation (BPMN) standard published by the Object Management Group. The
DMN standard has been designed to work alongside BPMN, providing a
mechanism for modeling the decision-making represented in a Task within a
process model. DMN need not be used with BPMN but it is highly compatible.

By focusing on the decisions that matter to an organization Decision Management

accelerates adoption of a Business Rules Management System or Predictive Analytic
Workbench, focuses you where you will have the highest impact, ensures business
ownership of the business rules, and delivers agility and continuous improvement of
the decisions across the company.
Decision modeling is an iterative approach to defining the requirements for decision-
making in your systems.

“What used to be one week of requirements work was done in a few hours with decision
management.”
Lead Business Analyst, North American Insurance Company

© 2016 Decision Management Solutions 4

 An Introduction to Decision Modeling with DMN

An Overview of Decision Modeling

Decision modeling has four steps that are performed iteratively:
1. Identify Decisions.
Identify the decisions that are the focus of the project.
2. Describe Decisions.
Describe the decisions and document how improving these decisions will impact
the business objectives and metrics of the business.
3. Specify Decision Requirements.
Move beyond simple descriptions of decisions to specify detailed decision
requirements. Specify the information and knowledge required to make the
decisions and combine into a Decision Requirements Diagram.
4. Decompose and Refine.
Refine the requirements for these decisions using the precise yet easy to
understand graphical notation of Decision Requirements Diagrams. Identify
additional decisions that need to be described and specified.

This process repeats until the Figure 1: Iterative Decision Modeling Cycle
decisions are completely
specified and everyone has a
clear sense of how the decisions
will be made.
At this point you can generate a
requirements document,
packaging up the decision-
making requirements you have
identified. This can act as the
specification for business rules
implementation work or for the
development of predictive
analytics.
Alternatively you can extend the
model with decision logic, such as decision tables, to create an executable
specification of your decision-making.

For a detailed discussion of decision modeling with DMN, download our white
paper Decision Modeling with DMN.

© 2016 Decision Management Solutions 5

 An Introduction to Decision Modeling with DMN

Identify Decisions
Decisions are identified as part of the overall requirements gathering process.
Decision-making appears in business processes, in use cases or as a specific
requirement. Not every decision is equally appropriate for decision modeling as we
will see in the Suitable Decisions for Decision Modeling section and as you iterate you
will also identify additional decisions.

Describe Decisions
Decisions are the core of the Decision Model and Notation
standard. Decisions are shown using a simple rectangle that
contains the name of the decision.
A Decision Requirements Diagram contains one or more Decisions and it is
common to create one such diagram for each area of focus on the project. Critical
information is captured about each Decision:
 Question and Allowed Answers.
 Business context.
 Organizational context.
 Application context.

A decision model also puts decisions in context by linking them to the KPIs and
objectives they impact, the organizations involved in the decision-making as well as
the processes and systems within which the decisions are made.

Specify Decision Requirements

A decision model describes the requirements for a decision in terms of
 Input Data requirements.
 Knowledge about the Decision.
 Related Decisions.
All of these are displayed on a Decision Requirements Diagram to define the model.
Each of these elements can be developed sequentially but you are more likely to
iterate through the various elements.
Decisions often require input data, information about
customers, parts, orders etc. Input data is generally
described at the level of business concepts or business
objects—things the business cares about. Input Data is
shown in DMN using an oval containing a name. The Input Data required for a
Decision is linked to the decision with an Information Requirement shown using a
solid line with an arrow head pointing from the Input Data to the Decision.

© 2016 Decision Management Solutions 6

 An Introduction to Decision Modeling with DMN

Decisions are driven by policies, regulations, best

practices, human expertise and analytic insight. DMN
calls these Knowledge Sources and they describe how
the decision is made and how it uses the Input Data
identified. Knowledge Sources are shown in DMN using
a document shape containing a name. The knowledge sources that impact a Decision
are linked to it using an Authority Requirement– a dashed line with a circle
All of these elements are be combined into a Decision Requirements Diagram. Such
a diagram contains Decisions, Input Data and Knowledge Sources along with
requirements links between these objects. A diagram can be for a single top-level
decision or a couple of closely related decisions. In the example below a high level
diagram shows a single Decision, Select Marketing Offer, along with two Input Data
and three Knowledge Sources.
A High Level Decision Requirements Diagram

Source: DecisionsFirst Modeler

Decompose and Refine the Model

The final stage in Decision Requirements modeling is to iteratively refine the model.
Decisions often require information that comes from making other decisions. You
can think of these decisions as sub-decisions or pre-cursor decisions. These
decisions must be made before the top-level decision(s) because the results of these
decisions must be available before the top level decision can be made. Each of these
decision is linked to the top level decision as Information Requirements because the
decision requires the information produced by making these additional decisions.
This analysis is repeated for each decision in the diagram, including any new ones
identified. Each new Decision identified may require Input Data (new or existing),
new Knowledge Sources and further additional Decisions. As the model is
developed it becomes more complete, such as the model shown below.

© 2016 Decision Management Solutions 7

 An Introduction to Decision Modeling with DMN

A More Complete Decision Requirements Diagram

Source: DecisionsFirst Modeler

This iterative refinement can be repeated as often as necessary to flesh out a

complete, coherent and useful Decision Requirements Model. If the diagram
becomes cluttered, then multiple diagrams can be developed showing overlapping
views on the same overall model. It is common for some parts of the model to have
more layers of decisions than others because the decision-making is more complex
or because there is more value in understanding it in detail. Modeling is complete
when the model is fit for purpose so different projects will require different levels of
detail and all are useful decision models.

Business Knowledge Models (BKMs)

There are two key knowledge objects in DMN—Business Knowledge Models
(BKMs) and Knowledge Sources. They have quite different purposes despite the
similar names.
 Knowledge Sources are for documentation.
 BKMs are about implementation.
Because both objects involve “knowledge”, there is some confusion about how to
document source documents. These are Knowledge Sources not BKMs
Some DMN examples show a BKM for every Decision, giving the impression that
decision logic can only be specified in a BKM. The DMN specification makes it clear
this is not the case.

© 2016 Decision Management Solutions 8

 An Introduction to Decision Modeling with DMN

The value of BKMs in allowing reuse is real when one is trying to store the
implementation definition outside the BRMS and then generate code for it. If you are
linking directly to a BRMS for implementation, then BKMs are unnecessary.

For more information, download our Best Practices Brief, Knowledge Objects in
DMN.

© 2016 Decision Management Solutions 9

 An Introduction to Decision Modeling with DMN

Suitable Decisions for Decision Modeling

Organizations make decisions of various types as shown in Figure 2.

Figure 2: Different Types of Decisions

Strategic Decisions
Organizations make infrequent
but large impact strategic
decisions. Much analysis is done
before the decision is made and
the implications for a business
can be dramatic.

Tactical Decisions
Regular tactical decisions
involving management and control are also made. There is generally still time and
energy for significant analysis but there is time pressure too, a need for consistency
and the opportunity to learn what works.

Operational Decision
Finally, every organization makes large numbers of operational decisions about
individual transactions or customers. Time pressure is often extreme and these
decisions must generally be embedded into operational systems and processes.

Characteristics of Suitable Decisions

Decision Requirements Modeling can be used for any decision. Because Decision
Requirements Models take some time and energy to build most organizations will
only do so when a number of things are true of the decision:
 There is clear value to defining them because they will be made multiple times
(often many, many times) and have enough complexity to warrant modeling.
Some combination of being complexity and repetition is what generally makes it
worth building a Decision Requirements Model.
 They are non-trivial because many policies or regulations apply, there is a wide
range of options to select from or lots of data to consider for instance. If the
way a decision is made must change often, if it is very dynamic, or if there is a
mix of drivers combined with a more modest pace of change then that will also
make the decision non-trivial. Almost any decision that involves an assessment of
fraud, risk, customer opportunity or similar through the analysis of historical data
is non-trivial.

© 2016 Decision Management Solutions 10

 An Introduction to Decision Modeling with DMN

 The value of the decision must be measurable and should be definable in

advance. We should be able to identify the KPIs and metrics that will be
improved by a better decision or weakened by a poor one.

Decision Categories
We can categorize operational decisions into various types, though some decisions
include characteristics of several types. For instance:
 Eligibility or Approval—Is this customer/prospect/citizen eligible for this
product/service?
 Validation—Is this claim on invoice valid for processing?
 Calculation—What is the correct price/rate for this product/service?
 Risk—How risky is this supplier’s promised delivery date and what discount
should we insist on?
 Fraud—How likely is this claim to be fraudulent and how should we process it?
 Opportunity—What represents the best opportunity to maximize revenue?
 Maximizing—How can these resources be used for maximum impact?
 Assignment—Who should see this transaction next?
 Targeting—What exactly should we say to this person?

Decision Services
Having identified high- Figure 3: Decision Services At The Center
ROI operational
decisions and modeled
them so you know what
to do to improve them,
your next step is to
design and build an
automated decision
making solution. The aim
is to create independent
Decision Services to
replace the decision
points currently
embedded in business
processes and
operational systems.

For a detailed discussion of the decision management architecture, download our

Decision Management Systems Platform Technologies Report.

© 2016 Decision Management Solutions 11

 An Introduction to Decision Modeling with DMN

Getting Started with Decision Modeling

Decision modeling is easy to adopt whether your project is kicking off or underway.
Now is the time to optimize and streamline your business, building in agility and
gaining operational efficiency.

Learn More
We offer a comprehensive set of white papers, webinars and other education
materials See the Reference section for suggested further reading.

Hands-On
Get started quickly with decision modeling. Go to www.decisionsfirst.com for
information about our decision modeling software DecisionsFirst Modeler. Our
online DecisionsFirst Support site includes free tutorials and training videos.

Training
Register for our IIBA Endorsed live online training courses. Email us at
[email protected] for information about IIBA member
discounts. See our website for session dates.

Free Consultation
We offer strategic guidance on when, where, and how to adopt Decision
Management and Decision Modeling best practices and technologies, coupled with
concrete tools for becoming analytically-driven.
We are one of the most experienced decision management and decision modeling
consultants and trainers in the world. Many of our clients use our decision modeling
software DecisionsFirst Modeler and many do not – our approach and training are
vendor neutral. We are also one of the original submitters of the Decision Modeling
and Notation (DMN) standard. Contact us for a free consultation today.

© 2016 Decision Management Solutions 12

 An Introduction to Decision Modeling with DMN

References

Additional White Papers and Reports

 Decision Management Solutions (2013-2015), Decision Management Systems
Platform Technology Report.
 Decision Management Solutions (2015), Decision Modeling with DMN.
 Decision Management Solutions (2015), Framing Analytic Requirements.
 Decision Management Solutions (2015), Maximizing the Value of Business Rules

Other Publications
 Taylor, James (2011). Decision Management Systems – A Practical Guide to Using
Business Rules and Predictive Analytics. IBM Press.
 DeBevoise, Tom and Taylor, James (2014). The MicroGuide to Process and Decision
Modeling with BPMN/DMN.

DMN Standard
 Object Management Group. Decision Model and Notation (DMN) Specification 1.0
Current version at http://www.omg.org/spec/DMN/Current
1.1 is approved but not yet publicly available

Contact Us

Email : [email protected]
Phone : +1 650 400-3029
Web : www.decisionmanagementsolutions.com

© 2016 Decision Management Solutions 13