G00265003

How to Architect the BI and Analytics Platform

Published: 11 August 2014

Analyst(s): Joao Tapadinhas

Many organizations have a business intelligence (BI) and analytics platform

that doesn't deliver according to expectations. Learn how to rearchitect it to
improve the impact on the organization and increase users' satisfaction.

Key Challenges

In many organizations, the BI and analytics platform centrally provisioned by the BI team
is not fulfilling business users' demands, does not clearly correlate with business decisions and
is overly focused on BI's technical aspects. Due to this, business units have started deploying
siloed solutions such as data discovery tools, vertical analytic applications, open-source
analytic workbenches and cloud BI.

BI teams, especially if located in IT, believe that the business analytics platform must be a
tightly integrated solution with as few components as possible preferably from a single
vendor to deliver a "single version of truth."

The capabilities typically offered are limited, including reports and dashboards, OLAP tools, ad
hoc access to data through SQL and eventually a data mining environment. There are gaps in
information exploration and analytics accessible to business users.

BI and analytics data sources are usually stored in a data warehouse and domain-specific data
marts, containing structured data from the organization's business applications and provided to
end users with access restrictions (often through BI reports).

Recommendations

Build a BI and analytics platform composed of three tiers the information portal, analytics
workbench and data science laboratory with varying levels of information trust, analytics
capabilities and information access.

Provision the analytics workbench and the data science laboratory with broader access to more
data sources, according to the needs and skills of the users leveraging each tier.

Define information governance rules and supporting metadata to manage an integrated BI and
analytics portfolio, promote content between tiers and ensure information consistency.

Allow business users to explore the right tier for their needs and skill level, formalizing and
supporting the roles of the information analyst and data scientist as primary users of the
analytics workbench and data science laboratory.

Table of Contents
Introduction............................................................................................................................................ 2
Analysis.................................................................................................................................................. 3
Acknowledge Misperceptions, Evolve Beyond the Monolithic Mindset.............................................. 3
Build Upon the Gartner Business Analytics Foundational Tools.........................................................4
Follow the Business Analytics Framework...................................................................................4
Leverage the Spectrum of Analytics Capabilities......................................................................... 5
Apply a Pace Layer Model Approach to the BI and Analytics Platform........................................ 7
Rearchitect the BI and Analytics Platform......................................................................................... 8
Information Portal..................................................................................................................... 10
Analytics Workbench................................................................................................................ 11
Data Science Laboratory.......................................................................................................... 12
Understand the Characteristics of a Tiered BI and Analytics Platform............................................. 13
Gartner Recommended Reading.......................................................................................................... 15

List of Tables
Table 1. Platform Tiers: Similarities and Differences.............................................................................. 10
Table 2. Characteristics of a Tiered BI and Analytics Platform...............................................................14

List of Figures
Figure 1. Gartner's Business Analytics Framework................................................................................. 5
Figure 2. Analytics Spectrum.................................................................................................................. 6
Figure 3. Pace Layer Model.................................................................................................................... 7
Figure 4. Tiered BI and Analytics Platform...............................................................................................9

Introduction
In many organizations, the BI and analytics platform is not fulfilling business users' demands and is
insufficient to achieve business objectives. The typical platform capabilities offered include reports
and dashboards, online analytical processing (OLAP) tools in some cases, ad hoc access to data
Page 2 of 16

Gartner, Inc. | G00265003

through SQL for a limited number of users, and eventually a data mining environment often
managed as an isolated solution.
An organization's established BI and analytics strategy and deployed tools are seldom enough to
meet users' needs because they are generally focused on systems of record reporting as opposed
to delivering a range of capabilities to a range of users. Enforcing them without additional options
will only worsen the existing problems.
To overcome limitations, business units have started deploying siloed solutions such as data
discovery tools, vertical analytic applications, open-source analytic workbenches, or subscribing to
alternative cloud BI platforms without IT's support or approval.
Due to these constraints, there is often a pressing need to rearchitect the BI and analytics platform
(covered in this document), review the user skills and responsibilities, and reorganize the process of
how content gets created and deployed (to be covered in a forthcoming note).

Analysis
Acknowledge Misperceptions, Evolve Beyond the Monolithic Mindset
There are a number of ingrained flaws in most organizations' BI and analytics platforms, as well as a
misperception of their objectives and how to manage them. BI teams, especially if located in IT,
believe that:

The business analytics platform must be a tightly integrated solution with as few components
as possible preferably from a single vendor to deliver a single version of truth to the
organization.

Information can only be trusted if stored in the corporate data warehouse and delivered to the
information consumer using BI artifacts, such as reports or dashboards.

Information created or manipulated by business users will inevitably produce discrepancies

through different analysis, leading to wrong decisions and generating chaos in the organization
over time.

IT's responsibility for information management stops at the BI semantic layer and IT-driven
content. Business-driven analytic processes are out of scope and not supported by IT.

There are plenty of documented information-driven problems in the BI and analytics world that
compel BI leaders to follow these beliefs and deploy a monolithic, centralized BI environment, which
ends up being enforced on users regardless of its fitness to needs. Incumbent BI vendors, in favor
of their own platforms, will encourage this approach too.
Gartner believes that a successful BI and analytics platform needs to evolve beyond the monolithic
mindset. A transformation must occur to offer different solutions for the disparate needs of users,
with a diverse set of integration levels striking a balance between trust and agility. The purpose is

Gartner, Inc. | G00265003

Page 3 of 16

to help users achieve their business objectives through the use of the proper technology, not to
eradicate the user-driven BI solutions that partially solve their problems today.
The resulting BI and analytics environment will also require changes in analytics and information
governance processes, as well as attribute new responsibilities to different personas in the
organization.

Build Upon the Gartner Business Analytics Foundational Tools

To support the BI and analytics platform transformation, Gartner has three foundational and highly
correlated tools to consider:

Business Analytics Framework

Analytics Spectrum

Pace Layer Model

These tools are to be applied in conjunction, but let's start by analyzing them in isolation.

Follow the Business Analytics Framework

Figure 1 presents Gartner's Business Analytics Framework. See "Gartner's Business Analytics
Framework" for a thorough review of this tool.

Page 4 of 16

Gartner, Inc. | G00265003

Figure 1. Gartner's Business Analytics Framework

Performance
Business Models, Business Strategy, and Enterprise Metrics

Descriptive, Diagnostic,
Predictive, Prescriptive

Social

Documents

Transactions

Information Capabilities

Metadata and Services

Analytic Capabilities

Information
Governance
Processes

Enable

Information

Analytic
Processes

Collaboration, Decision Making, Intelligent

Decision Automation, Applications

Platforms

Processes

People

Program Management

Produce

Decision Capabilities

Decision
Processes

Consume

Describe, Organize, Integrate,

Share, Govern, Implement

IT/OT

Image

Audio

Text

Video

Mobile

Search
Engine

Source: Gartner (August 2014)

The Gartner Business Analytics Framework identifies the people, process and platform components
that support the transformation of information into better performance of the organization. The use
of this tool is done by reading it from the top down, starting with the business outcomes and then
figuring out the supporting analytic compositions and information required to achieve them.
According to users' needs, the platform must be rearchitected with a broader set of technical
capabilities (filling the gaps), new responsibilities and organization. Focusing on tools or vendor
standardization alone is not the answer.
The Framework is also very useful for defining current and future architecture states. The difference
between them is the road map and includes changes in people and processes. The organization
will, most likely, also need to reorganize and retrain the providers and users of BI and analytics. The
business users must gain access to the proper analytic tools according to their goals and skills
and a comprehensive range of data sources with varied data types, suitable granularity and
appropriate access.

Leverage the Spectrum of Analytics Capabilities

Focusing on the platforms pillar of the Gartner Business Analytics Framework, in particular the
analytic capabilities component, we see four analytic styles that are further detailed in Figure 2. See
"Extend Your Portfolio of Analytics Capabilities" for a thorough review of the Analytics Spectrum.

Gartner, Inc. | G00265003

Page 5 of 16

Figure 2. Analytics Spectrum

Analytics

Human Input

Descriptive
What happened?
Diagnostic
Why did it happen?

Data

Predictive
What will happen?

Prescriptive
What should I do?

Decision

Action

Decision Support
Decision Automation

Source: Gartner (August 2014)

The analytic capabilities deployed in organizations are often limited to descriptive analytics, through
basic reports and dashboards. With this capability, the question, "What happened?" can be
answered. After knowing "what," users will most likely also ask, "Why did it happen?" Properly
addressing this requires far more agility and more advanced information exploration capabilities.
Traditional BI deployments tend to have gaps in this area but IT usually overlooks the problematic
impact of this and continues to push the standard vendor and its unfit-for-purpose tools. As a
consequence, users will resort to Excel, ad hoc queries, data extractions and shadow IT teams to
achieve their analysis goals.
BI leaders must extend the BI and analytics platform to diagnostic analytics to complement
descriptive analytics. This is where OLAP and in-memory data models are used to provide easy and
speed-of-thought navigation of data without a predefined query. Leveraging enhancements to the
data access level, we also see the need for improved semantic layers abstracting the complexity of
the underlying physical model. This can make it much easier for self-service discovery without the
accompanying IT bottleneck found in a typical BI team.
Beyond the data layer, we see the introduction of newer data visualization tools, and this is where
the focus of the rapid increase of data discovery tools is concentrated. But traditional tools can also
provide improvements with a greater focus on more comprehensive reporting (such as variance
analysis), integrated planning, dashboards and KPI reporting.
Over time, with an increasingly higher analytics maturity level, the organization should move into
predictive and prescriptive analytics. These require a significant increase in the skill levels of the
business analysts. Predictive models require development and maintenance with complex logic and
business rules. They incorporate sophisticated methods that can also require a deep understanding
of statistical or operational research.

Page 6 of 16

Gartner, Inc. | G00265003

Furthermore, organizations must realize that there is a need to blend all these different techniques
into comprehensive solutions rather than leave them as discrete silos.

Apply a Pace Layer Model Approach to the BI and Analytics Platform

Figure 3 presents the Pace Layer Model. See "Applying Gartner's Pace Layer Model to Business
Analytics" for a thorough review of this tool.
Figure 3. Pace Layer Model
Systems of
Innovation

Analytical
Agility
Dynamic/
Ad Hoc
Systems of
Differentiation

Configurable/
Autonomous

Systems
of Record

Structured/
Repeatable

High

Low

Levels of Integration Between People, Processes and Platforms

Source: Gartner (August 2014)

Going into further detail on how to architect the BI and analytics platform, the Pace Layer Model
differentiates systems according to their integration level and analytical agility. It outlines three
approaches to systems and information:

Systems of record: Usually the tightly integrated, sanctioned legacy application and
infrastructure systems in an organization, these typically support administrative and transaction
processing activities such as finance, HR, asset management or procurement. In BI they largely
correspond to standardized reports and dashboards developed and managed by IT.

Systems of differentiation: As applications that support processes unique to the organization

or its industry, systems of differentiation can be seen as more agile BI platforms. Products in
this category may come from traditional IT-modeled BI architectures with semantic layers, but

Gartner, Inc. | G00265003

Page 7 of 16

with extra capabilities. They are often offered by smaller independent vendors in the form of
data discovery. The capabilities could be described as easier for end users, fostering selfservice adoption, but also as capable of providing the sophisticated analysis for a broader
range of users.

Systems of innovation: These are applications built to support new, innovative business
activities and constructed quickly to enable enterprises to take advantage of these new ideas
and opportunities. Systems of innovation represent the more fragmented and decoupled
solutions more tactical, agile and sometimes forward-looking BI applications from vendors
providing domain-specific analytic applications and capabilities such as predictive modeling.

Most of the existing BI and analytics deployments run by IT are focused on the systems of record
reporting (such as a data warehouse, extraction, transformation and loading [ETL], and BI platform
combination). The systems of differentiation and innovation are usually disconnected and spread
across the organization in line-of-business silos far from IT's control. Furthermore, there is little
ability and no processes to leverage business-user-generated content as the basis of requirements
to enhance the systems of record.

Rearchitect the BI and Analytics Platform

BI leaders should follow the foundational tools described above to successfully rearchitect the BI
and analytics platform. Gartner recommends the setup of a tiered architecture composed of:

Information portal

Analytics workbench

Data science laboratory

This is the representation of the tiered BI and analytics platform (see Figure 4) use it as a generic
guideline that can be tuned according the organization's specific characteristics.

Page 8 of 16

Gartner, Inc. | G00265003

Figure 4. Tiered BI and Analytics Platform

Source: Gartner (August 2014)

To realize the vision of the three tiers (according to the Pace Layer Model) and be able to maximize
their strengths, BI leaders need to deploy new technical capabilities to provide missing analytic
styles, improve the usage of existing tools through a better overall integration, and provide common
metadata and governance.
Processes and people roles and responsibilities, although not detailed in this note, are of utmost
importance for success. They must be addressed in conjunction with the technical platform as
described in the Gartner Business Analytics Framework.
As depicted by the arrows in Figure 4, some of the key processes to address are:

Gartner, Inc. | G00265003

Page 9 of 16

The ability to promote content built by business users in the analytics workbench or data
science laboratory to the information portal

Tight cooperation between the information analysts and data scientists

It must also be noticed that the different analytic styles (from the Analytics Spectrum) tend to align
better with certain tiers, as represented in the top part of Figure 4 ("Typical analytic styles usage").
However, they may be pervasive across the whole platform for example, descriptive dashboards
are more relevant in the information portal but may also be used in the analytics workbench.
The platform tiers must work in conjunction. Table 1 shows the characteristics they share that help
create a coherent global BI and analytics platform and, at the same time, the disparate strengths
they have to support a diverse set of needs and audiences.
Table 1. Platform Tiers: Similarities and Differences
Similarities

Belong to the same overall BI and analytics

landscape and work in conjunction toward
common goals.

Differences

Serve a different set of needs.

Support different user audiences with

different skills, although the ultimate
information consumer may be the same.

Share a base set of information.

Support the execution of integrated analytic

processes.

Add disparate data sources to each tier,

according to user needs.

Share BI and analytics content, metadata

and processes.

Use different technical capabilities and

tools, eventually from different vendors.

Follow a single set of encompassing

information governance rules.

Require different access levels to

information.

Are supported by a common BI and analytics

team and follow a single, integrated strategy.

Produce distinct outputs and impact the

business differently.

Source: Gartner (August 2014)

We'll expand on each tier to understand how to integrate and leverage them in conjunction.

Information Portal
Closely follows the characteristics of systems of record from the Pace Layer Model.
The information portal is the workspace where business users can quickly and easily find the key
trusted metrics with which the organization measures its performance. It is usually made of
reporting and dashboard capabilities that provide content to information consumers.
Its outputs are the result of a formal development process that includes a business user
establishing requirements and a technical specialist (typically from IT but increasingly from the

Page 10 of 16

Gartner, Inc. | G00265003

business) implementing them. This can take days or months depending on complexity and
workload. The information can be trusted and is used across the organization, but has low flexibility
and reduced associated interactivity capabilities.
Typical platform capabilities:

Reporting

Dashboards

Microsoft Office integration

Mobile BI

Embeddable analytics

Sample tools and vendors:

SAP BusinessObjects

IBM Cognos

Oracle BI

Microsoft Reporting Services

MicroStrategy

Information Builders WebFocus

Analytics Workbench
Closely follows the characteristics of systems of differentiation from the Pace Layer Model.
The analytics workbench is the workspace used to investigate trends in trusted metrics or to detect
patterns in other datasets from multiple sources that may turn into opportunities or risks. It is
an agile tier to explore information and has access to a broad range of data sources, with limited to
no support from technical experts. Toolsets should include a data discovery tool and a number of
other capabilities to help business users extract value from information autonomously.
In the Analytics Spectrum, the workbench is able to provide descriptive analytics but will usually
focus on diagnostic analytics. In some cases namely through the use of more analytics-focused
data discovery tools it can extend to a basic level of predictive analytics and will gain data
modeling and more advanced analytic capabilities going forward.
Typical platform capabilities:

Data discovery

Ad hoc reporting/querying

Gartner, Inc. | G00265003

Page 11 of 16

Geospatial and location intelligence

Embedded advanced analytics

OLAP

Business user data mashup and modeling

Collaboration

Data filtering and manipulation

Sample tools and vendors:

Tableau Software

Qlik

Tibco Spotfire

SAS Visual Analytics

SAP Lumira

Oracle Endeca Information Discovery

MicroStrategy Visual Insight

Alteryx

Microsoft SQL Server Analysis Services and Power BI

Data Science Laboratory

Closely follows the characteristics of systems of innovation from the Pace Layer Model.
The data science laboratory is the workspace where advanced analytics takes place and is the ideal
incubator for big data initiatives. It is a flexible environment where experimentation with trial and
error is actually encouraged to generate impactful insights for the organization.
A broad set of technical capabilities is expected and often provided by specialized tools with
minimal IT integration, meant to deliver agility and the ability to answer unforeseen questions. This is
why IT tends to overlook this area in favor of investing in the information portal.
Users are skilled and experienced, often more than the technical experts in IT. Their toolsets include
data mining capabilities, forecasting and other complex statistical and analysis tools.
Typical platform capabilities:

Advanced data access

Support for big data sources

Page 12 of 16

Gartner, Inc. | G00265003

Advanced descriptive analytics

Predictive analytics

Forecasting

Optimization

Simulation

Further advanced analytics

Although not BI capabilities, Hadoop and other NoSQL databases must also be referenced here

Sample tools and vendors:

SAS Enterprise Miner

IBM SPSS

SAP InfiniteInsight

Revolution Analytics and R

RapidMiner

Knime

Alteryx

FICO

Dell StatSoft

Cloudera

Hortonworks

MapR and other Hadoop distributions

Understand the Characteristics of a Tiered BI and Analytics Platform

The following table summarizes the characteristics of the three tiers. BI leaders should try to
understand the gaps in their current BI and analytics deployment and change their strategy
accordingly.

Gartner, Inc. | G00265003

Page 13 of 16

Table 2. Characteristics of a Tiered BI and Analytics Platform

Characteristic

Information Portal

Analytics Workbench

Data Science Laboratory

Key Objective

Deliver standardized and

trustable information to the
organization.

Provide information exploration

capabilities to a broad range of
business users.

Allow the production of

advanced analytic
processes.

Audience

Technical specialists
produce the BI content;
business users (decision
makers) consume it.

Business information analysts

produce content for decision
makers, as in the information
portal.

Data scientists produce the

content for consumption by
business operations (such as
call center staff) through
embedded analytics in
business applications. May
be consumed directly by
customers (e.g., via
websites).

Data Sources

Structured information from

the enterprise data
warehouse (EDW) or
domain-specific data mart.

Mostly structured information

from the EDW, domain-specific
data marts, user-generated
information (often in Excel),
business applications, and
external and open data. Starting
to use unstructured data
sources. Inputs from the
information portal.

Structured and unstructured

information from all available
internal and external data
sources. Inputs from the
analytics workbench.

Trust vs. Agility

Reliable and structured

information, but static and
inflexible.

Interactive, agile and

customizable according to user
needs, but may show
discrepancies in user-generated
information.

Reliable, in-depth and factdriven. Customizable to

answer or solve a single
issue and inflexible
thereafter.

Time to Deliver
Content

Can take days to months

for development, but can
be consumed in seconds.

Minutes to hours.

Days to months.

Level of Skills
Required

Intermediate to advanced
technical skills for content
development. No particular
BI skills for information
consumption.

Basic to advanced data

manipulation capabilities at a
business-user level. No major
technical or statistical skills
required.

Advanced technical and

mathematical skills.

Information
Access Required

Narrow access to
information for
consumption, according to
user role and profile.

Broad access to multiple data

sources, according to areas of
responsibility.

Very broad access to

information, according to
areas of responsibility.

IT Support
Required

High content
development.

Medium data source

availability and content
validation.

Low data source

availability.

Page 14 of 16

Gartner, Inc. | G00265003

Characteristic

Information Portal

Analytics Workbench

Data Science Laboratory

Typical Analytic
Capabilities
Produced

Descriptive analytics and

output of predictive and
prescriptive analytics.

Descriptive, diagnostic and

basic components of predictive
analytics (such as forecasting).

Advanced descriptive,
diagnostic, predictive and
prescriptive analytics.

Source: Gartner (August 2014)

Gartner Recommended Reading

Some documents may not be available as part of your current Gartner subscription.
"Gartner Business Analytics Framework"
"Extend Your Portfolio of Analytic Capabilities"
"Applying Gartner's Pace Layer Model to Business Analytics"

Gartner, Inc. | G00265003

Page 15 of 16

GARTNER HEADQUARTERS
Corporate Headquarters
56 Top Gallant Road
Stamford, CT 06902-7700
USA
+1 203 964 0096
Regional Headquarters
AUSTRALIA
BRAZIL
JAPAN
UNITED KINGDOM

For a complete list of worldwide locations,

visit http://www.gartner.com/technology/about.jsp

2014 Gartner, Inc. and/or its affiliates. All rights reserved. Gartner is a registered trademark of Gartner, Inc. or its affiliates. This
publication may not be reproduced or distributed in any form without Gartners prior written permission. If you are authorized to access
this publication, your use of it is subject to the Usage Guidelines for Gartner Services posted on gartner.com. The information contained
in this publication has been obtained from sources believed to be reliable. Gartner disclaims all warranties as to the accuracy,
completeness or adequacy of such information and shall have no liability for errors, omissions or inadequacies in such information. This
publication consists of the opinions of Gartners research organization and should not be construed as statements of fact. The opinions
expressed herein are subject to change without notice. Although Gartner research may include a discussion of related legal issues,
Gartner does not provide legal advice or services and its research should not be construed or used as such. Gartner is a public company,
and its shareholders may include firms and funds that have financial interests in entities covered in Gartner research. Gartners Board of
Directors may include senior managers of these firms or funds. Gartner research is produced independently by its research organization
without input or influence from these firms, funds or their managers. For further information on the independence and integrity of Gartner
research, see Guiding Principles on Independence and Objectivity.

Page 16 of 16

Gartner, Inc. | G00265003