1. Business Analytics
IBM Software IBM Analytical Decision Management
Making your enterprise
ready for optimized
decision making
Better outcomes in real time, every time
Introduction
Workers within your organization make thousands—maybe even
millions—of decisions daily. However, to meet today’s performance
goals—increases in revenues and profits, cost reduction, risk
management and other improved outcomes—organizations know
that they cannot continue to operate the way they have in the past.
In most organizations, transactional decisions need to be made in an
instant. Because there is no time to process and analyze the staggering
volumes of ever-changing and increasingly varied data, these decisions
are often based on a mix of knowledge of corporate policy, experience
and gut instinct.
Individually, each of these decisions has a relatively small value to the
organization. But, taken together, they can add up to have a significant
impact on the successful execution of your strategy and bottom-line.
Contents:
1 Introduction
2 The IBM Analytical Decision
Management architecture
5 IBM Analytical Decision Management
design environment
8 IBM Analytical Decision Management
execution environment
10 Making decision management part of
the enterprise
15 Conclusion
16 About IBM Business Analytics
2. IBM Analytical Decision Management
IBM Analytical Decision Management, a platform that
combines predictive analytics, local rules, scoring,
optimization and data management automates the high-
volume, high-value decisions that organizations make every
day. This can vastly increase the benefits that flow from
good decisions and minimize the risks associated with bad
decisions. The platform helps organizations in a range of
industries connect strategy to execution, especially those
with large volumes of transactional decisions—such as in
retail, banking and financial services, insurance and
manufacturing as well as government agencies and
academic organizations.
Organizations that use IBM Analytical Decision Management
to automate and optimize decisions have a significant
advantage over competitors. With the globalization and
accessibility of resources today, businesses that ensure smart,
accurate decisions have a greater chance of prevailing on
the competitive playing field.
This paper focuses on the technical aspects of an end-to-end
IBM Analytical Decision Management solution. The first
section describes the architecture of the solution; the second
and third sections explain the design and execution
environments required. A final section provides guidance
on managing the implementation of this solution.
The IBM Analytical Decision
Management architecture
IBM Analytical Decision Management creates an analytical
foundation for an organization’s entire decision-making
ecosystem. Because it leverages existing technology, those
who implement IBM Analytical Decision Management can
generate results quickly and achieve a high rate of return on
their technology investment.
The IBM Analytical Decision Management
framework
IBM Analytical Decision Management is a decision process
framework that leverages key enterprise technologies to
optimize outcomes within point-of-interaction systems.
Business people define decision patterns—leveraging
predictive models and local rules—which lead to optimal
outcomes at the point of impact. People simulate what will
happen before it happens and then, when they are satisfied
with the results of the simulation, move configurations
through robust rollout procedures.
This ensures proper governance of changes to mission-critical
systems and processes. The basic architecture includes a JEE
application server framework (the core of the IBM Analytical
Decision Management product) which runs in IBM SPSS
Collaboration and Deployment Services (the IBM SPSS
platform). The framework orchestrates processes within
the IBM SPSS Modeler analytics engine by “consuming”
enterprise data to generate predictive models which are
then deployed into an organization’s systems. The underlying
platform of the solution includes an SOA-based integration
layer and real-time scoring system for connecting to point-of-
interaction systems.
Business Analytics
IBM Software
2
3. three critical components drive IBM Analytical Decision
Management’s ability to optimize and automate decisions and
recommendations at the point of impact for better outcomes
in real-time.
Predictive models in IBM Analytical Decision Management
are developed using a variety of sophisticated analytical
techniques. Existing models can be leveraged inside IBM
Analytical Decision Management; the solution leverages SPSS
Modeler as its core analytic infrastructure. Predictive Model
Markup Language (PMML) models may also be imported
into the solution and facilities exist to enable the business to
create new models from scratch.
Built on predictive analytics
Predictive analytics—products such as IBM SPSS Modeler
and IBM SPSS Statistics—help people anticipate what is
likely to happen next. IBM Analytical Decision Management
leverages the power of predictive models built with these
technologies. A complete IBM Analytical Decision
Management solution must include at least three critical
components: SPSS Modeler 15 (which serves as the predictive
engine), SPSS Collaboration and Deployment Services 5
(which enables scoring) and SPSS Decision Management 7
(which combines predictive analytics and rules, drives the
what-if simulations, applies optimization and more). These
IBM Analytical Decision Management
Business Analytics
IBM Software
3
Figure 1 : The architecture of an IBM Analytical Decision Management solution includes a JEE application server framework running in IBM SPSS
Collaboration and Deployment Services. It includes an SOA-based integration layer and a real-time scoring system.
JEE Architecture
4. IBM Analytical Decision Management
Business Analytics
IBM Software
4
The solution also incorporates the ability to continuously
and automatically refresh these models, to guarantee that
the best-performing model is in production. IBM Analytical
Decision Management leverages and supports the use of
business intelligence for managing ongoing solution
performance. Built-in reporting provides visibility into the
underlying efficacy of deployed IBM Analytical Decision
Management solutions. Configurations of ongoing
performance can also be made accessible to enterprise
business intelligence technologies (via ODBC/JDBC) to
provide richer dashboards or scorecards.
Leveraging critical enterprise data
IBM Analytical Decision Management leverages critical
enterprise data—both structured and unstructured. Typically,
data stored in transactional data stores and data warehouses
is combined with other valuable data which lives in
spreadsheets, call center notes, social media and surveys.
It also supports a lightweight virtual data layer called
Predictive Enterprise View (PEV). This virtual layer can
bring together multiple data sources in one virtual layer
to streamline data access for the business and can be
integrated with existing data management technologies of
the enterprise—including enterprise data warehouses,
entity management and master data management.
Because of IBM Analytical Decision Management’s open
architecture, there are no technical limitations on the number
and type of data sources that can be leveraged for making
recommendations at the point of interaction. The platform
can leverage data characteristics or metadata on all types of
data to support the decision process; data attributes associated
with the defined data sources can be leveraged within rule
creation, predictive models or prioritization/arbitration
equations. Practically speaking, most organizations
consolidate data in an analytical data warehouse and present
it through the solution via configurations in the virtual
data layer.
To make recommendations to website visitors, the user of IBM
Analytical Decision Management might want to begin by analyzing
web clickstreams. However, most business people don’t have the
know-how to manually add these complex data sources. A data
source combining historical clickstream data, previous purchase
history and key customer demographics ready for analysis.
Adding a new data source—such as a customer’s purchase
history—involves mapping the new source through the
platform’s administrative interface by defining a Data Provider
Definition (DPD). Once that has been defined, the new
attributes can be leveraged within the IBM Analytical
Decision Management interface for rules, arbitration inputs
or model inputs. For sources that do not support JDBC/
ODBC, the platform supports additional mechanisms such
as Web Services, Remote Method Invocation and Java
Messaging Services.
Incorporating local rules
Local rules translate the organization’s experience into
practical operational policy for use in the operating
environment to drive specific decisions. IBM SPSS Decision
Management supports the creation of these business-centric
rules directly (with a built-in simple rules engine). Or IBM
Analytical Decision Management can be architected to
include the capabilities of an enterprise rules engine (like
IBM WebSphere®
Business Rule Management System and
IBM Operational Decision Management).
Rules systems, like predictive analytics, can be used independently to
optimize business processes. However, solutions which combine the
uplift value of predictive models as well as the business context of
rules generate the highest and fastest ROI.
5. IBM Analytical Decision Management
Business Analytics
IBM Software
5
Integration with enterprise rules systems is typically via
Web Services or via multi-stage data level integration.
Most customers find the existing rules capabilities of the
solution appropriate for getting started with their first
implementations, and explore larger scale rules infrastructures
after solutions begin delivering return.
Connecting to point-of-interaction systems
Customer-facing applications and processes optimized
with IBM Analytical Decision Management provide
up-sell, cross-sell, risk assessment and other business
recommendations that steer action and drive desired
outcomes. IBM Analytical Decision Management supports
multiple ways for actions to be customized, integrated and
deployed to business users—without requiring extensive
staff training or invasive IT work. Inter-operability with
business process management and complex event processing
technologies enables optimized decisions to be leveraged
in long-running or fast throughput situations.
IBM Analytical Decision Management
design environment
Focused on outcomes
IBM Analytical Decision Management enables business
users to configure and manage all of the different decision
outcomes which are relevant to the decisioning challenge they
are facing. At a typical company, for example, each product
category or campaign might have its own associated offers,
qualification criteria, allocation definition and prioritization
formula.
The IBM Analytical Decision Management framework walks
the business user through a structured decision-making
process which helps them define how best to arrive at the
appropriate outcome.
Structured decision-making in seven steps
IBM Analytical Decision Management provides a console
from which business users can administer and maintain
business scenarios, including a number of common use cases.
This console maintains access control to all the business
scenarios and can be integrated with a directory server such
as LDAP. Upon login, the business scenarios (applications)
will be visible to the users based upon their access rights.
Building an IBM Analytical Decision Management application
typically involves seven steps:
1. Connect to data.
2. Define decision scope via Global Selections.
3. Define desired outcomes.
4. Define operational decisions with rules and predictive
models.
5. Test through champion and challenger models.
6. Deploy.
7. Report.
The deployment of decision management solutions usually
requires integration with many different enterprise
technologies: from enterprise resource planning systems
and customer relationship management to business process
management and enterprise service buses. The underlying
platform exposes standards-based deployment options for
all of these technologies.
In website optimization, deployed solutions can be integrated with
content management systems to return digital assets, such as
collateral content, based on sales recommendations or other offers.
This content is rendered in context of existing web processes and
infrastructure. Integration in this case would be via Web Services or
standard JDBC/ODBC bridges.
6. Business Analytics
IBM Software
6
IBM Analytical Decision Management
Step 1: Connect to data. The business user starts by specifying the data
sources from the global repository or from their own data repository. These
intuitive user interface with tabs helps guide the user to the next step.
Step 2
sets the guidelines for the type of decision that will be rendered. At this point,
created to include high risk customers and this rule is based in part on a
decisions to desired outcomes.
Step 1
Step 2
Step 3
this example, attributes such as days aging and outstanding balance dictate
different in other regions.
Step 3
Step 4
to balance objectives, determine the maximum number of offers
and optimize outcomes. Rules, based on historical experience, might,
for example, give certain customer segments priority contact. Predictive
models, which are good at picking up subtle patterns in the data, can
complement rules to predict more accurately whether sub-groups may
be more or less likely to respond to the contact.
Step 4
These steps are configurable to support different business
scenarios such as customer analytics and operational costs.
The example shown on these pages demonstrates the
application of IBM Analytical Decision Management to
solve a business problem related to Accounts Receivables.
7. Business Analytics
IBM Software
7
IBM Analytical Decision Management
Step 5: Test through champion and challenger models. Before deploying
decisions operationally, it is important to be able to simulate or forecast
what the likely impact will be. “What if” simulations help the business test
analysis is complete, business users will be able to see the estimated
results of the offer. They can test different scenarios and optimize the
accuracy of decisions to deliver the optimal outcome. In this example, the
contact decision has been optimized with constraints to include a the
probability of a response and the related recognized revenue minus the
cost to make the contact and offer.
Step 5
Step 6
Step 6: Deploy. With a single click, a business user alerts IT to move the
solution from testing into production. The business has indicated its
approval, now IT can perform the relevant activities needed to migrate a
solution into production. The Deploy button marks the project stored in the
repository as ready. IT is in control of the actual deployment and integration
of the solution into enterprise systems. Any project marked for production
can be registered with the real-time scoring service of IBM SPSS Collabo-
ration and Deployment Services and called via Web Services.
Step 7: Report on outcomes. From the IBM Analytical Decision Management
solution’s Report tab, users can monitor the status of deployed applications.
IBM SPSS Decision Management provides the ability to output and report
statistical and graphical results of campaigns and their supporting predictive
models. This is essential for business users so they have insight into the
business intelligence technologies, as well, and provides a feedback mecha-
nism to start the process all over again.
Step 7
8. Business Analytics
IBM Software
8
IBM Analytical Decision Management
IBM Analytical Decision Management
execution environment
The IBM Analytical Decision Management solution supports
a closed-loop approach to information management, using
the holistic analytical data view—incorporating descriptive,
behavioral, transactional and attitudinal data—to provide
inputs for decision optimized predictive models. Advanced
model governance capabilities such as champion/challenger
and threshold alerts ensure that IBM SPSS solutions support
informed decisions by business users and automated decisions
at the point of interaction.
Closed-loop decision-making process
The figure below shows a closed-loop process. This process
supports decision-making across departments, applications
and channels. The IBM Analytical Decision Management
analytical engine ensures that customers are receiving the
optimal recommendations even in a high-volume
environment.
While the management and delivery of the recommendations
is certainly an important component, the rest of the
components, including the IBM SPSS Collaboration and
Deployment Services and the IBM SPSS Modeler
environments, are required for this closed-loop/real-time
support.
Figure 2: The closed-loop process supported by IBM Analytical Decision Management enables information from current interactions to be used to refresh
predictive models, which, in turn, guide future offers and recommendations.
9. Business Analytics
IBM Software
9
IBM Analytical Decision Management
These are the typical steps in this process:
1. A customer makes contact with your company—through
a website, in a call center or in person.
2. The point-of-interaction system (e.g. website) invokes the
real-time scoring service passing runtime parameters inside
of a Simple Object Access Protocol (SOAP) call.
3. The IBM Analytical Decision Management solution
(running inside of the scoring service) processes rules,
models, and prioritization routines and recommends a
specific action.
4. The point-of-interaction system renders the recommended
offer, potentially pulling assets from a content management
system.
5. This process repeats until a customer reacts as anticipated
or abandons the interaction.
6. The solution captures the outcome of the interaction in an
Offer History database.
7. The model that generated the initial recommendation is
refreshed, on a set schedule, with the latest Offer History
information and other relevant data about all customer
interactions.
8. The performance of this model is compared with that of
other models, with the best-performing model stored in
the solution’s repository and flagged for reuse.
9. With the next customer interaction, this model will be used
to generate recommended actions.
The business user, meanwhile, accesses the performance
results of the solution in the IBM Analytical Decision
Management portal.
Real-time scoring
IBM Analytical Decision Management employs the scoring
service of IBM SPSS Collaboration and Deployment Services.
Scores are generated in real time using predictive models
developed in IBM SPSS Modeler and enhanced with IBM
SPSS Decision Management. Figure 3 shows the scoring
service layer running in IBM SPSS Collaboration and
Deployment Services to support mission-critical applications
that have real-time, high throughput requirements.
Figure 3 : The scoring service in IBM SPSS Collaboration and Deployment Services supports mission-critical applications with real-time, high-throughput
requirements.
Scoring Service
10. Business Analytics
IBM Software
10
IBM Analytical Decision Management
The scoring service framework is designed so that the
consuming applications do not need to worry about the
underlying implementation techniques such as model pre-
cache, data pre-cache, inline-transforms and scoring itself.
An important aspect of the scoring service interface is that it
provides these operations independently of any particular
underlying model implementation. Models are invoked by
reference, and the underlying framework handles all of the
mechanics needed to fetch a specified model, load it, invoke
the correct scoring implementation and return the result to
the consuming applications.
Models are invoked by reference, and the underlying
framework handles all of the mechanics needed to fetch
a specified model, load it, invoke the correct scoring
implementation and return the result to the consuming
applications.
Architectural drivers of performance
The performance of IBM Analytical Decision Management
solutions deployed in real time is directly dependent on
the scoring service layer of IBM SPSS Collaboration and
Deployment Services. There are two distinct metrics used
for monitoring performance.
Scores per second: the number of scoring requests
successfully executed and returned to the requesting client
by the scoring service
Response time: the time elapsed from the client initiating a
score request until the client receives a response from the
scoring service
Effect of solution con guration
In addition to the impact of different application server
architectures, IBM Analytical Decision Management
implementations also can vary in performance based upon
model and solution complexity. In some scenarios, real-time
data retrieval is needed to fully execute a model and return a
score—using what is known as a “real-time data provider
definition” within Predictive Enterprise View. In other
configurations, all data is gathered prior to invoking a score.
Both approaches have technical validity depending upon
the business scenario.
Different IBM Analytical Decision Management applications
can be configured to support multiple outcomes. Each
outcome may be driven by different rules and different
models.
Making decision management part of the
enterprise
A number of key issues need to be addressed to ensure that
your implementation of predictive analytics and decision
management meets the needs of your organization’s IT and
business objectives. Whether you are concerned with optimal
model lift, are optimizing systems in the Web 2.0 world or
are trying to build a successful project team, decision
management must be treated in the context of any other
mission critical IT system.
11. Business Analytics
IBM Software
11
IBM Analytical Decision Management
Support for high availability
In many of the enterprises that might select IBM Analytical
Decision Management, the number of decisions or
recommendations to be made daily might number in the
thousands, the tens of thousands, into the millions.
The related scoring, offer delivery, outcome recording and
model refreshment will require significant computing
resources. IBM Analytical Decision Management solutions
are often architected for high availability—including analytic,
test/acceptance and production environments, as illustrated
in Figure 4.
The analytic environment is designed for creating modeling
assets that may or may not be destined for deployment.
This is where the “heavy lifting” occurs—tasks such as data
integration and preparation, model creation and validation
of the best-performing algorithms. The test/acceptance
environment is the pre-deployment/QA environment. It is a
mirror configuration of the production environment (without
server capacity) and is used for testing and QA purposes.
The production environment is the deployment environment,
in which real-time customer interactions occur with pre-
calculated propensity scores or a “live” predictive model
integrated with other operational systems that support
closed loop recommendations. IBM Analytical Decision
Management is designed to scale horizontally, meaning that
components can be deployed onto additional server resources
as needed in order to accommodate increased demand.
Figure 4: IBM Analytical Decision Management solutions typically include analytic, test/acceptance and production environments.
12. Business Analytics
IBM Software
12
IBM Analytical Decision Management
Ensuring model performance in decision
management solutions
Predictive model governance is a key component of IBM
Analytical Decision Management. The underlying IBM SPSS
Collaboration and Deployment Services platform supports
model governance in the following ways:
Model performance is automatically monitored and can be
configured to provide an alert when a model’s accuracy
degrades to a pre-determined level. Such a level may be
based on business rules and/or on target variables.
While an instance of a model is deployed into production,
the model itself can be configured to automatically challenge
the production model against other algorithms to ensure that
the best performing, most accurate model is in production
based on new data. If a more effective model is found, an
alert can be sent to notify users, or it can be automatically
deployed into production to replace the existing instance.
Models and the related performance metrics are version-
controlled within the decision management platform.
This gives users insight into model characteristics and
performance.
Considerations for optimizing web interactions
IBM Analytical Decision Management can be configured to
capture web click-streams in real time. This is critical for
optimizing recommendations to customers contacting your
company through an organization’s website. Several different
approaches can be provided, depending upon a customer’s
particular environment.
The mechanism works by using network taps to monitor
full duplex (two-way) traffic. It monitors details of the web
pages viewed by “listening” to both incoming and outgoing
traffic at the network level. This does not require any
changes to a customer’s website pages and does not impact
the end-user experience.
Since it can see the entire detail of the page being returned it
can capture information not available to web server log files. It
gives a total control of the capture and aggregation of
information—only relevant data is collected. The following
are some examples of click-stream data that can be captured:
Standard web log statistics—pages viewed, dwell time on
each page, referential information, cookie items etc.
Session, page and hit (i.e., file) level summarization of
web browser information (by visitor if known)
SSL traffic can be de-coded and captured
The click stream data is stored in a relational database that
will be accessed as needed by IBM SPSS Collaboration and
Deployment Services via SOAP to provide real-time
recommendations.
13. Business Analytics
IBM Software
13
IBM Analytical Decision Management
Considerations for optimizing marketing processes
When optimizing marketing decision processes, two primary
technology touchpoints exist. First, most organizations have
separate marketing automation systems which manage
ongoing marketing campaigns. IBM Analytical Decision
Management is called to provide targeted offers specific to
the individual customer. Solutions often retrieve campaign
and offer history as part of rules processing to ensure the
proper cadence is followed.
Event-based email can be managed within IBM Analytical
Decision Management in so far as the events that should
trigger an email can be monitored and a time scheduled for
the release of the message. However, the solution does not
provide a mechanism for executing mass email distribution.
For example, a list of email recipients might be generated
and saved to the software’s Campaign History database.
A scheduled process within the IBM SPSS Collaboration
and Deployment Services Automation Service can then
retrieve this list, determine the best customized offer for
each recipient, and write the offer details back into the
Campaign History database prior to email execution.
Typically, users of IBM Analytical Decision Management
define target outcomes which blend the application of target
systems as appropriate. This enables operational email,
marketing email and other touchpoints to be leveraged
appropriately given the uniqueness and objectives of each
channel.
Best practices for on-premises deployment
Depending upon the organization’s resources and level of
familiarity with predictive modeling, IBM Analytical Decision
Management can be implemented either with direction
and training from IBM staff or by engaging IBM consultants.
A typical implementation might span twelve to sixteen weeks,
depending upon the scope of the business issues, the amount
and complexity of data sources, and other factors. IBM
consultants and training staff are ready to transfer their
knowledge and expertise to your organization—and are
experienced at collaborating with client organizations to
develop a suitable project roadmap.
interaction history to a supporting database that captures what offers
have been extended and accepted. This database can also include
any supporting customer/visitor information that would be required to
-
ness reports can be exposed within the IBM Analytical Decision
Management interface on the Reporting tab or scheduled for wide-
spread distribution through automation facilities provided through
the platform.
The second point of interaction is with email systems. Email
integration usually takes place at the data layer. There are
essentially two types of systems: operational email systems
and campaign-based systems.
Operational email integration might require interfacing
with an organization’s website. This is because operational
email is generated in a dynamic, near-real-time fashion.
For example, when a customer completes an event that
triggers an operational email (registering on a company’s
website, for example), an “event” in a decision management
solution would be initiated—typically through Web
Services—resulting in a particular offer being made to
the new registrant.
14. Business Analytics
IBM Software
14
IBM Analytical Decision Management
After the initial implementation, IBM Analytical Decision
Management requires little IT support unless there is a need
to add an additional data source, or add a new website page
that will deliver recommendations. Post-production support
typically involves staffing roles defined as business/marketing
users and analytic users. The business users are the primary
users of the application and have ultimate responsibility
for the maintenance of the configuration to support an
organization’s goals and initiatives. The number of users
will vary depending upon the scale of deployment and the
number of departments involved.
IBM Analytical Decision Management is designed to
empower the business to participate in much of the project
implementation. From an analytics standpoint, it provides
professional modeling services to organizations with no
experience with predictive analytics. Most organizations
have at minimum one staff member working with the IBM
SPSS analytics team on the analytics portion of the solution
to ensure that maximum experience is obtained from a
predictive analytics perspective. While models can be created
by those with little to no analytic experience using IBM
Analytical Decision Management (via IBM SPSS Modeler),
there is significant value in having dedicated resource(s)
focus on the analytic process and results.
In order to be successful, it is recommended that the primary
business users have extensive product usage understanding.
In addition, the IT organization should have a solid analytical
backbone in place to prevent project failures.
When starting the project, it is also recommended that there
is a documented analytical process flow, which will ensure
that the initial project, as well as subsequent projects will
be a success.
The recommended training for existing staff should be on-site,
during a series of hands-on sessions for each business analyst
user. While training is under way, an IBM consultant should
“shadow” the professional analysts/DBAs, ensuring the future
success of the transition of data and predictive models between
them and the business analyst user. Finally, the IT staff should
undergo system administration training at your IBM facility,
which will prepare them to maintain a properly running
infrastructure.
Key Services
The following services are typically leveraged when an
organization purchases IBM Analytical Decision Management.
The projects vary in scope and not all organizations will need
all services.
Installation and con guration of staged deployment
infrastructure
IBM consultants install IBM software and related technologies,
ensuring interoperability with enterprise systems, including
LDAP, MDM layers and other data
access and integration frameworks.
Roadmap creation
IBM consultants work with your business analyst team to
indentify key operations decisions which can be optimized
by IBM Analytical Decision Management. Prioritization of
opportunities lays the groundwork for ongoing analytics
adoption.
Architecture planning
Working with IT, IBM consultants recommend hardware
and software layouts appropriate for the scale of your
implementation.
15. Business Analytics
IBM Software
15
IBM Analytical Decision Management
Data preparation
Common data services include ETL, statistical data
manipulation, and data infrastructure planning.
Professional model building, optimization and model
drift analysis
IBM specializes in applying advanced analytical techniques
which incorporate robust mathematics and CRISP-DM
methodology to ensure maximum predictive power of
deployed models.
Analytic Solution Template development and deployment
IBM consultants can craft solutions specific to your industry
and business problem using a vernacular familiar to the
business constituency.
Point-of-Interaction systems integration
IBM consultants work with your IT organization to embed
analytics via SOA or other best practices integration
techniques to ensure maximum availability and performance.
Software as a Service (SaaS) deployment
If your organization is looking for an alternative to
purchasing, hosting and administering IBM Analytical
Decision Management, then you are a candidate for the
Software as a Service (SaaS) deployment. With this option,
IBM SPSS Online Services acts as the ‘back office,’ providing
the infrastructure, the security and the software. All you
need is Internet access.
This offering works well for organizations that
Lack the resources to focus on IT implementation
Haven’t yet developed solid skills in database and model
building
Prefer no upfront software purchase and would rather pay
as you go
Organizations that select a SaaS deployment pay a monthly
subscription fee which includes server software, installation,
infrastructure hosting and support and no limit on the number
of users. Pricing is tiered, based only on storage requirements.
The SaaS Jump Start consulting option
If you’d like assistance getting started with the SaaS
deployment for IBM Analytical Decision Management,
your organization also has the option of adding IBM
Jump Start consulting through IBM Business Analytics
Lab Services.
Jump Start consulting for the SaaS deployment provides
additional expertise and proven practices to help your
organization get up and running quickly—and make smarter
decisions sooner. Typically, organizations can expect to
be production-ready in approximately 45 days—rather
than months.
Conclusion
This paper has described IBM Analytical Decision
Management and how its open, service-oriented architecture
helps organizations optimize outcomes at the point of impact.
It includes a solid business-driven experience based on best
practices for analytical decision making. It allows IT to
integrate the use of real-time analytics with existing business
processes and systems.
Implementing IBM Analytical Decision Management will
help your organization improve insight into customers,
processes and business patterns to drive better real-time
decisions and actions in every corner of the organization.
This is made possible by establishing well-constructed
processes and empowering individuals throughout the
organization with predictive analytics. The result: rapid,
informed and confident decisions and actions across the
organization, based on consistent, trusted information.
