CONTENTS Business Rules - Decision Management Solutions · Services replace decision points in...
Transcript of CONTENTS Business Rules - Decision Management Solutions · Services replace decision points in...
1 © 2017 Decision Management Solutions
Maximizing the Value of Business Rules Decision Management streamlines and focuses business rules projects for faster, more effective deployment
Business Rules Management Systems deliver on the promise of costs savings,
agility and happy customers. Yet for many organizations, these efforts
remain point solutions. Decision Management is a proven framework to
drive the widespread adoption of business rules and improved business
performance across the company.
You have realized the benefits of adopting a Business Rules Management System or
BRMS. Now, if you could have these results anywhere else in your business, where
would they make the most difference? Standalone project successes confirm that a
BRMS delivers on its promise of costs savings, agility and happy customers. Yet most
business rules efforts remain point solutions.
A typical first BRMS implementation starts by creating what could be a called a “Big
Bucket O’ Rules.” The team has selected a BRMS as part of a project and they start
capturing their rules. Typically, they interview experts, read policy manuals and
reverse engineer code into rules. Almost always, but especially when they reverse
engineer code, they end up with a lot of somewhat low-level rules. In one, big,
bucket.
This big bucket of rules is not usable. The team will often try to add rules to their
very complex process diagrams—placing each individual rule on the process
diagram. Most of the low-level rules cannot be placed in this way and doing so just
contributes to over-complex business processes. It’s also difficult to manage the
By James Taylor CONTENTS
Decision
Management—
Automating and
Improving Decisions
Decision Discovery
and Modeling
Decision Service
Definition and
Implementation
Decision Monitoring
and Improvement
Decision Modeling
Conclusion
Maximizing the Value of Business Rules
2 © 2017 Decision Management Solutions
rules properly. There’s no organizational structure to the rules so they end up being
grouped by source or by the person doing the rule identification. As policies or
regulations change, or as new business needs are identified, the team often
struggles to update the right business rules.
Typically, the team perseveres and ends up grouping the rules into services that
perform some useful action, generally answering some question or selecting an
action from a list of candidate actions. Groups of rules, rulesets, are deployed as
services or several rulesets are linked together and deployed. These services can be
invoked by existing systems or as part of a new business process.
What the team has done, often without realizing it, is identify the decisions that
their rules support. Not only are these decisions more stable—new decisions within
a business are rare while new rules or changed rules are common—they represent
the connection point that existing systems and new processes need. Connecting
systems and processes to decisions and automating those decisions using business
rules gets them the results they were looking for. They get a positive ROI from
using a BRMS by making these decisions more accurately and more consistently.
They engage business users and empower them to manage the rules. They deliver
agility as it is easier to evolve the rules behind these
decisions as the business changes.
For all the success and enthusiasm coming out of a
BRMS project that takes this “Big Bucket O’ Rules”
approach, companies often stall when trying to expand
adoption of BRMS across the company. Some pigeonhole
the BRMS as a technology only suitable for the initial
project – the BRMS becomes “claims technology” or
“the pricing engine.” New projects in other areas
revert to writing code or using table-driven
parameterization. Some companies try to go
“enterprise-wide” but this is like trying to boil the
ocean as it becomes an effort to find and manage all their business rules.
In both cases the problem is that the decisions being managed using these business
rules are not central to the solution. If the team does not have a way to identify
and classify decisions that are suitable for automation with a BRMS then the BRMS
will be perceived as a solution to the first, specific problem for which it was
purchased. Similarly, without a way to identify high priority decisions and so focus
expansion efforts on the rules needed for them, broad adoption of the BRMS will
often stall.
Decision Management resolves these issues. Business Rules Management Systems
generate an ROI when they improve business decisions – it’s the decisions that
create the value. By focusing you on the decisions that matter to your organization
– the ones that affect your business drivers and measures – Decision Management
simplifies business rules design and implementation while accelerating adoption of
a BRMS where it has the highest impact.
Maximizing the Value of Business Rules
3 © 2017 Decision Management Solutions
Decision Management—Automating and Improving
Decisions
Decision Management focuses on the decisions that create value in your business,
recognizes these decisions as reusable assets, and makes them widely available via
a Service Oriented Architecture (SOA). The three phases (shown in Figure 1) are:
Decision Discovery and Modeling
The first step in adopting Decision Management is Decision Discovery and Modeling.
Decision Discovery finds the decisions that matter to your business and drive
results. Decisions are modeled using the Decision Model Notation (DMN) standard
ensuring consistency and re-use. See Decision Modeling at the end of this white
paper for more information.
Decision Service Definition and Implementation
The second step is the definition and implementation of Decision Services. Decision
Services replace decision points in processes and systems to make those processes
and systems simpler, smarter and more agile. Defined in terms of the decision
model, these Decision Services are built using Business Rules Management Systems
and can be enhanced with the results of data mining and predictive analytics, now
or in the future.
Decision Measurement and Improvement
The final step is to close the loop with Decision Measurement and Improvement.
Ongoing decision measurement and improvement ensures that decision making is
monitored and constantly improved to deliver increasing value over time.
Figure 1.The Three Phases of Decision Management
Source: Real-World Decision Modeling with DMN, Figure 5-3
Maximizing the Value of Business Rules
4 © 2017 Decision Management Solutions
Decision Discovery and Modeling
By understanding which decisions matter to your organization – which ones affect
your business drivers and measures – it becomes straightforward to show a strong
ROI. Decision Discovery is thus the first step in applying Decision Management.
Discover Decisions
Decision Discovery identifies the high value decisions that will provide the biggest
pay-off for your business. Decision Discovery externalizes the operational decisions
in your processes and systems. This allows the decisions to be understood and
owned by the business and allows decision-making to be linked explicitly to
performance measures and KPIs. With this done it will be clear what changes to
decision-making will be required to improve any given measure. And the decisions
that have the greatest impact on the most critical measures will be the place to
start.
Model Decisions
Graphical decision models like that in Figure 2 make it easier to communication and
collaborate on requirements and outcomes. Staff can then focus on value-add
activities that require their expertise, adding further value.
Figure 2. Example Decision Requirements Model
Source: DecisionsFirst Modeler
Maximizing the Value of Business Rules
5 © 2017 Decision Management Solutions
Decision modeling using the industry Decision Model and Notation (DMN) standard
provides a framework that teams across an organization can use. It provides a
common language between business analysts, architects, business owners, IT
professionals and analytic teams. Decisions are more easily tied to performance
measures and to the business goals. This makes it easier to focus teams where they
will have the highest impact and to measure results.
Link to ROI Drivers
Having discovered the decisions that run your business and contribute to your KPIs
the next step is to prioritize them for the application of Decision Management. You
need to know how likely it is to show a strong ROI. Good candidates include:
Manual decisions made by supervisors because automating them will empower
others to make those decisions directly, reducing referrals and increasing first
call resolution.
Manual decisions where more than seven factors must be considered as people
are not good at handling complex trade-offs.
Inconsistent decisions where the website gives one answer while the call center
gives another or where different call center representatives give different
answers are good candidates as the management of those decisions can
eliminate the inconsistency.
Decisions embedded in existing systems can be good candidates too when the
existing system implements regulations, has requirements that change frequently or
has many maintenance projects.
There are several characteristics of a decision that make it likely to show a high ROI
from Decision Management. Any of these characteristics is enough, though decisions
with combinations are more common and will show higher ROI. These
characteristics of high ROI decisions include:
The Decision Model and Notation (DMN) Standard was approved in 2014-2015 by the Object Management
Group. Decision Management Solutions is a submitter of the DMN standard, an open industry standard with
broad vendor and community support.
This continuous manufacturing company automated the process of checking materials masters for its ERP
systems, previously a manual decision, and reduced elapsed time to create a materials master by 95%.
This government agency moved a hard-coded decision into a decision service (calculating the cost of a
license involving some 2,000 rules derived from constantly updated regulations) from its legacy system
and could save 13,000 hours in maintenance work in the first year alone.
Maximizing the Value of Business Rules
6 © 2017 Decision Management Solutions
Rapid or frequent change required
Multi-channel
Highly dependent on outside factors
A need to demonstrate that the decision is compliant
Risk management
Potential for personalization
Reasonably high volume
B2C
We’ll call this Business-to-Consumer company “B2C”. It has an objective to improve
its customer retention this year and has established several KPIs including
“Percentage of customers who renew when their contract expires” and “Save rate –
the percentage of customers who are persuaded to stay when they call in to
cancel.” These are the Retention KPIs.
While the efficiency of B2C’s process for handling customer renewals and the steps
it takes when someone cancels contribute to these KPIs, it is more helpful to
consider the action taken by a call center representative in each case. Specifically,
the retention offer chosen at that moment is critically important.
In other words, the “Retention Offer Decision” is a key operational decision that
impacts the Retention KPI.
Identifying the retention offer decision enables B2C to discuss how they want to
retain customers, how they might determine the value of retaining different kinds
of customers, which departments should have a say in the retention approach and
more. Making the decision explicit gives a focus for these business discussions.
A typical retention scenario involves making retention offers across multiple
channels each with different applications and platforms. These applications are
optimized for supporting a specific channel – a website or a call center for instance.
This means a high risk of inconsistency – the retention decision B2C makes is likely
to be inconsistent across channels. B2C makes thousands of retention decisions
every day and how effective a retention decision will be is highly dependent on our
competitors. Multi-channel, high-volume and dependent on outside forces – clearly
the retention offer decision is a good candidate for high ROI use of business rules.
Once identified, this decision can now be modeled. The various elements that make
up the decision – the sub-decisions – can be identified to give the decision
structure. How the decision and its sub-decisions use data can be clearly specified
and all the various source of decision-making know-how (policies, regulations,
expertise and analytics) can be identified. A complete model of the decision-
making can be defined.
Maximizing the Value of Business Rules
7 © 2017 Decision Management Solutions
Decision Service Definition and Implementation
Decision Services are the implementation of a decision – how systems will find out
what the best or most appropriate decision is for a particular customer or
transaction. A decision service also makes the decision reusable and widely
available.
Decision Services are essentially business services in a Service Oriented Architecture
(SOA) that deliver an answer to a specific question. These services generally do not
update information – they just answer questions such as “how should we handle this
claim?” or “what is the right discount for this order?” Because they don’t make any
permanent changes, they can be used to answer questions whenever they come up
without worrying about potential side effects.
The usual SOA infrastructure is used to access a decision service. The decision
Figure 3. Decision Services
Maximizing the Value of Business Rules
8 © 2017 Decision Management Solutions
service can access enterprise resources and other services (using the same SOA
infrastructure) and can make various business-level interfaces available for other
services to use. Decision Services are built primarily using business rules managed
using a Business Rules Management System or BRMS. While many only need a BRMS,
they also support the integration of predictive analytics now or in the future. Figure
3 shows how these elements come together to provide Decision Services across an
application portfolio.
Business Rules
Having found and modeled high ROI decisions, the next step is to define the
business rules that support them. Business rules represent the expertise, tribal
knowledge, regulations and policies that drive a business. Mining business rules
from software code, reviewing regulations and policy manuals, interviewing experts
and many other techniques can be used to find the rules that support the decisions.
Business rules are the core building block of the decisions in decision services.
Business rules create a language the business and IT can both understand – essential
for effective Decision Management. Business rules are maintained in a repository or
catalog that is updated by both business and technical users.
Decision Models and Business Rules
The decision model is used to scope and bound the decision service by identifying
the decisions that should be included – those inside the automation boundary. The
business rules for each of these decisions are identified and managed in the BRMS.
The decision model acts as a framework for capturing and defining the business
rules, keeping each set of business rules focused on a particular sub-decision. The
knowledge sources identified in the model show where the rules can be found and
the input data show how data is consumed by these business rules in order to make
a decision.
The business rules for a decision are most commonly represented as decision tables
or rule sheets, as shown in Figure 4. These tabular representations allow many rules
to be presented in a compact format while allowing easy verification of the rules
against the decision model and verification of completeness.
Where a tabular layout is inappropriate, business rules can also be controlled by
customizable rule templates that incorporate organization-specific terms to make
them easier for non-technical users to understand and edit. Some BRMS also support
additional business rules representations such as decision trees or decision graphs.
Business rules are also used in workflow, to manage data quality in a database or to control a user
interface. These are effective uses of business rules but are different than the business rules that are
built into a Decision Service. The rules used for a Decision Service are truly about the business and how it
should act. They are independent of a company’s current databases, systems or processes.
Maximizing the Value of Business Rules
9 © 2017 Decision Management Solutions
These representations further ensure that the decisions can be managed by those
who understand the business context of those decisions.
Focusing on the business rules for a specific decision helps avoid a common problem
in business rules implementations. Business rules are everywhere in an
organization. Collecting and managing those rules without an organizing driver can
result in lots of rules being managed for relatively low business impact. Certainly,
everyone understands the rules better and it is easier to find and update them but
the company will not maximize the return from this investment unless business
performance has improved as a result. Focusing on the decisions ensures that this
will be the case. A focus on decisions increases the value of a rule repository and of
an effective system for managing rules.
Business Rules Management Systems
Using a BRMS to encode business rules has several advantages that are critical to
managing decisions, even when the equivalent software code would be simple:
The syntax is clearer to a non-programmer so that decisions built with them can
be managed by non-programmers
The rules can easily be reused across multiple systems that use the decisions
The rules are independent. No sequence is implied making it possible to edit
them, and thus change decisions, to respond to changing business
circumstances, without unintended consequences.
Figure 4. An example Decision Table
Real-World Decision Modeling with DMN, Figure 11-11
Maximizing the Value of Business Rules
10 © 2017 Decision Management Solutions
The basic components of a BRMS are shown in Figure 5. Technical and non-technical
rule management, a rule repository, verification, impact analysis and testing tools.
Decision Services are generated from these other elements. Designing the
repository, giving business users rule management capability, verification, testing
and deployment will all be simplified by focusing on specific decisions.
B2C, Part 2
Once B2C recognizes that a retention decision is required and makes it explicit they
are ready to automate it consistently across their various channels. They develop a
Decision Service that is used by each channel. For example, instead of the website
making its own retention decision, it will use the Retention Decision Service to
determine the appropriate offer for a specific customer when it is needed and
deliver that through the web channel. The call center, outbound marketing and
other channels will also use the Retention Decision Service.
Figure 5. Components of a Business Rules Management System
Source: Decision Management Systems Platform Technology Report
Maximizing the Value of Business Rules
11 © 2017 Decision Management Solutions
Having identified and modeled their retention decision, B2C wants to specify it with
business rules to make it flexible, agile and business-centric. A BRMS enables them
to specify the decision in a way that is easy to update, easy to manage and
accessible to the business users who understand retention. This ensures their
retention offers will reflect their business strategy, best practices and the know-
how of their most successful agents.
For this retention decision B2C gathers rules from the customer service
representatives with the best history of cost-effectively retaining customers, from
the head of customer service, from existing policy manuals and from existing
customer service scripts. Collecting, normalizing and reviewing these rules give
them a first cut rule set for retention decisions.
Predictive Analytics
Although many rules are written based on judgment, tribal knowledge, regulations
and policies, this information can be augmented by using data mining and analytics
to find or refine business rules. Increasingly common, this approach can be an
important complement to business rules for some decisions. Decision Management
is the framework that links business rules to analytical and data mining insights and
uses those insights to improve operational decisions.
Many BRMS support a Decision Tree metaphor and many data mining techniques
create decision trees to divide populations as, for
instance, in customer segmentation. Representing these
models in a BRMS allows them to be seen, managed and
potentially updated by business users. Business rule
management systems can also use decision tables to
implement the models that result from data mining and
analytics techniques as score cards. These represent a
collection of rules that “score” a particular customer or
transaction depending on the value of specific attributes.
In the Figure 6, the scorecard shows that spending two or
more years at your current employer adds 50 to your risk
score. The total score can easily be calculated from the
various elements of the score card. BRMSs execute these
score cards quickly and effectively while the data mining
and analytic tools and techniques find out which
attributes make a difference and what values for those
attributes predict behavior. Implemented as a score card,
the predictive model is easily used in any decision based
on rules.
Figure 6. A Predictive Scorecard Table
Source: Smart (Enough) Systems
Maximizing the Value of Business Rules
12 © 2017 Decision Management Solutions
B2C, Part 3
B2C could do a better job of making retention offers if it knew who was a retention
risk. This would allow them to spend on retention offers for customers who are at
risk not those that will stay anyway. But to be effective they must make retention
offers before a customer decides to leave.
Using data mining techniques B2C uses historical data about which customers left to
estimate the probability of a particular customer failing to renew (their retention
risk). This results in a scorecard that weights and totals these predictors into a risk
score. Implementing this scorecard in the rule technology means the score is used
by rules to target customers who are a risk more precisely. This allows B2C to write
rules that focus their most expensive offers on profitable customers who are at high
risk of leaving.
Decision Monitoring and Improvement
Decision Morning and Improvement involves the application of performance
management techniques and technologies to the monitoring of decisions. With a
Decision Management approach, the business understands how specific decisions
create value. These decisions are linked to the business and individual performance
metrics being tracked. To continuously improve business performance, Decision
Management monitors decision performance, throughput and basic statistics. How
many decisions are made to approve, reject or refer is a measure of decision
effectiveness. Too many referrals will increase the burden on staff doing manual
reviews. Too many rejections, thanks to false positives for instance, will impact
customer service or sales. Similarly, decisions that take too long or that cost too
much (because they use data that must be purchased, for instance) may have a
negative overall impact. Tracking and reporting on this information will help the
business owners understand and thus manage their decisions more effectively.
B2C, Part 4
B2C manages the performance of its retention decision by tracking which customers
are, in fact retained as well as the offer made and its cost. Business users see what
specific offers cost relative to their effectiveness so they can improve the cost-
effectiveness of the decision.
B2C tests out new ideas for renewal rules to see how a particular approach –say a
more expensive and aggressive one – would affect results. If the additional
customers retained would be worth the additional cost they can deploy the new
approach, improving the decision. The monitoring and analysis of the decision,
along with its mapping to key company KPIs, allows them to continuously improve
it.
Maximizing the Value of Business Rules
13 © 2017 Decision Management Solutions
Decision Modeling
Decision modeling is fast becoming a best practice for business rules projects,
replacing traditional rules analysis approaches that were created before modern
BRMS capabilities.
Decision modeling is specified using the Object Management Group’s Decision Model
and Notation (DMN) standard. This industry standard gives users access to a broad
community and a vehicle for sharing expertise more widely. It is also a technique in
the International Institute of Business Analysts (IIBA) Business Analyst Body of
Knowledge or BABOK®, giving business analysts an industry standard approach to
accurately describe decision requirements that are
essential for today’s data-driven projects focused on
improving decision making.
The example model in Figure 7 uses the Decision
Model and Notation (DMN) standard notation. It
shows a marketing offer decision (a rectangle) along
with its sub-decisions (and sub-sub decisions), the
input data entities (ovals) that are required by each
of these decisions and sub-decisions, and the
knowledge sources (document share) that constrain
or guide the decisions.
Decisions First
The rules-first, natural language rulebook
approaches used historically tend to capture many
individual rules in a rush to detail. This results in
large numbers of low-level rules, often with multiple
versions, that quickly become hard to change and maintain.
Often written in natural language and phrased as constraints, these rules must be
converted to the executable rules and rule metaphors that will work in a BRMS. This
is an additional step and creates situations where multiple versions of a rule are
being maintained. This step is no longer needed given the capabilities of a modern
BRMS. The creation of an independent fact model and vocabulary that must be
mapped to the implementation data model can exacerbate this severely,
sometimes resulting in yet more versions of each rule.
The rules-first approaches can seem fine for the initial project, but quickly the
rules can become complex and hard to manage. Teams will often try to add rules to
their business process diagrams, placing each individual rule on the process
diagram. Most of the low-level rules cannot be placed in this way and doing so
contribute to over-complex business processes. There’s also no organizational
structure to the rules so they end up being grouped by source or by the person
Figure 7. Example Decision Model
Source: DecisionsFirst Modeler
Maximizing the Value of Business Rules
14 © 2017 Decision Management Solutions
doing the rule identification. As policies or regulations change, or as new business
needs are identified, the team often struggles to update the right business rules.
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 Decision Model
and Notation (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.
The Role of Decision Modeling
Decision Modeling has a role in all of the Decision Management phases. As shown in
Figure 8, Decision Modeling:
Builds a transparent and unambiguous
definition of the decision at the beginning
of the project, before rules are gathered,
providing a decisions first, top-down view.
Scopes and specifies the interface and
content of Decision Services. The decision
model structures the internals of the
decision service and provides a business-
focused structure for the rules in the
service.
Finally it establishes meaningful and
business-centric measurement criteria and
structures the monitoring and improvement
activities, closing the loop back to the
original decision.
Figure 8. Decision Modeling Supports All Phases
Maximizing the Value of Business Rules
15 © 2017 Decision Management Solutions
Decision Model - BRMS Integration
Integrating a Decision Requirements Model with a BRMS implementation allows
decision requirements models to be linked directly to the business rules that
implement them. Decision Management Solutions’ DecisionsFirst Modeler offers
integration with leading BRMS’s.
The integration automatically generates links between the BRMS and DecisionsFirst
Modeler as a project progresses and are updated automatically. The integration
provides:
Full traceability from business objectives through decision requirements models
to the business rules running in production.
Full access to all the rule editing, validation, testing and deployment
capabilities of the BRMS
Standard security of the BRMS to ensure only authorized users can access the
rules.
DecisionsFirst Modeler does not
capture implementation details
such as business rules or decision
tables in our models as this results
in duplication—there would be one
version in the model and another in
your implementation environment.
Instead it links Decisions to
Implementation Components
representing the business rules,
decision tables or decision trees in
the BRMS. Because the relationship
between Decisions and
Implementation Components is
many:many, reuse of logic between
Decisions is supported.
For example, in Figure 9, the
decision table in IBM Operational
Decision Manager containing the
rules for the Rental Agreement
Surcharge is directly linked to that
decision in the model.
Figure 9. Decision Model in DecisionsFirst Modeler link to BRMSs
A detailed description of how to do decision modeling is described in our free white paper, Decision
Modeling with DMN, available in the white paper section of our website.
Maximizing the Value of Business Rules
Conclusion
Decision Management is the framework you need to ensure the success of your first
business rules project or expand your adoption of business rules. Decision
Management is focused on business drivers and on the decisions that impact them,
so you can apply your BRMS where it will make the most difference for a clear ROI.
Decision Management systematically works through identifying and modeling these
decisions and mapping these decisions to your business objectives. Decision Services
are designed and implemented using your BRMS to ensure accurate decision-making
and business agility. The effectiveness of decisions is constantly monitored so that
the decisions can be continuously improved.
Decision Management delivers clear priorities for applying business rules and a BRMS
for maximum ROI, give each project a clear objective and delivers a robust
monitoring and improvement framework to keep you competitive over time.
Decision Management is a proven framework that ties business rules to business
objectives. A Decision Management approach to business rules begins by looking at
business performance drivers and focusing on the decisions that have the biggest
return. Decision Management streamlines BRMS implementation, ensures business
ownership of the business rules, and delivers agility and continuous improvement.
For More Information
Taylor, James and Purchase, Jan (2016). Real-World Decision Modeling with
DMN. Meghan-Kiffer Press.
Taylor, James (2011). Decision Management Systems – A Practical Guide to
Using Business Rules and Predictive Analytics. IBM Press.
Decision Management Solutions (2013-2015), Decision Management Systems
Platform Technology Report.
Decision Management Solutions (2017), Decision Modeling with DMN.
Taylor, James and Raden, Neil (2007). Smart (Enough) Systems. Prentice Hall.
CONTACT US
Decision Management Solutions specializes in helping organizations build decision-centric, action-
oriented systems and processes using decision management, business rules and advanced analytic
technologies.
www.decisionmanagementsolutions.com Email: [email protected]