An enterprise collects enormous amount of data from various verticals like
sales, marketing, inventory, service, etc. The question is, why do you want to
take and secure this historical data? Obvious reason being that historical data
helps you make better decisions in the current scenario. The historical data
could be in the form of charts, spreadsheets, graphs-to sum it up a sort of
decision history which tells about the kind of decisions taken in the past. Each
and every decision taken by the management at any point of time is clearly
marked on the data in terms of negative or positive outcome.
|
For example, whenever there is any change in marketing strategy of any
company, the sales are affected. Either they rise or they fall, stating market
reaction for changed decision. Every organization takes decisions based on their
best logical approach by using every possible resource like market survey, past
project reports, and historical data. Despite using the best possible resources,
some of these decisions turn out to be a failure. The reason being the
non-utilization of future analysis. That is where Business Intelligence (BI)
comes to play.
BI is one technology which is worth implementing for each enterprise. You
surely want your decisions to be more accurate and precise. Right. We'll tell
you exactly how to go about it. Earlier BI has been looked upon as an enterprise
solution that can manipulate historical data into a more presentable form,
merely to take decisions based on past results. But that's not what BI is all
about.
It arms you with the power to visualize future growth and help you take
better decisions based on logical and analytical reasons, calculated and made
ready for you by the BI suite. All enterprises perform according to their
management capabilities. Similarly, each BI suite provides capabilities
according to its defined structure.
Now, the question arises, as to whether an organization really needs to have
a BI solution? The answer is yes. Not only to help in present decisions but also
in your future and long term decisions.
What makes it 'happening'
The BI market is on a growth curve as far as India is concerned. From large
enterprises with existing ERP, SCM, and similar implementations to SMBs that are
ready to improve their IT infrastructure, the scope for BI implementations lies
wide open. The growing potential for BI can be attributed to the reason that
many organizations that already have systems in place to collect data and gather
information, many a times find themselves in a situation where they have no
tools or roadmaps to put the vast data and information into use for strategic
decision making.
For example, an organization might generate reports out of the existing CRM
or ERP solution. But there would be a lot of man hours required if it wanted to
analyze both these reports in perspective. It's places like this where BI
based solutions come into play. Databases at best concentrate on collecting
data; and storing and optimizing it to get information. But with increasing need
for organizations to adapt quickly to changes and the voluminous information
they collect, the information collector needs more than just properly defined
and laid out data.
For a quick analysis of such complexity one has to go for BI based solutions.
The potential customers for vendor products include organizations across various
verticals. Most of the pharmaceutical houses have already gone for BI products
and are harnessing the positives of BI solutions already for OLAP, ETL, or even
information management. But there is still a lot of scope in other segments such
as production houses, retailing, and even in the public sector as systems such
as warehouses are yet to be harnessed by many companies in this part of the
world.
Benefits
The very first thing that must be made clear is that BI will not happen on its
own. It will perform as good as it has been configured. True benefits will be
obtained only when a skilled and analytically sound employee has structured the
overall process according to the management needs. Similarly you need to have
proper constructing team that will understand and change the implemented
structure and structural process according to changed market propositions and
companies' current policies and strategies.
BI will give you edge to make better decisions more efficiently. There are
different scenarios where BI can make a potential mark, as in case of production
you can have all the relevant data that will be able to show you the correct
production of each and every inventory item not only in terms of time but also
about each production belt and batch.
If you talk of sales then you will be able to see through the facts of
figures the current market shifts and along with that you can easily visualize
the future prospect of any business and process according to not only its
present rating but also accounting market trend. You would be able to forecast
the demand or stocks that will arise in any period of time depending upon the
season, buyer and customer behavior.
The advantages of using a BI solution are: easy data representation; complete
start to end logistic data information, much better than that from reporting
obtained by knowledge management; and management information systems. You will
be able to seamlessly integrate various dimensions such as customer, product,
employee, etc. to discover information that was not possible earlier. It lets
you have access to database analysis instantly, to view which product, market,
customers and employees are showing growth prospects, so that you can focus your
energy on them, and which ones have been slowing down, so that you can address
the situation before it gets out of hand.
Implementing BI increases collaborative work among the various departments
for better work efficiency and productivity as all databases are linked, hence
each minor or major change to a database reflects across every department and
the company's entire structure.
This in turn leads to a transparent work flow and smoother processes through
increased collaboration. BI can also cause reduction in time taken by any
enquiry or report due to various access and check points in the organization
providing improved business planning and decision making for each short, middle
and long term scenarios. After better analysis and understanding of the market,
organizations will see increase in sales and profits. It will be easier for an
organization to actually reduce overall costs and expenditure. It also enhances
your customer support and services, which in turn provides better sales and
customer support. Organizations will also be able to predict the nature and
behavior of the market according to the present market strategies and decisions.
Sectors such as Pharmaceutical, IT/Telecom, etc have starting using BI
solutions. In addition to utilizing an ERP or a CRM solution, BI further
enhances your information reservoir for better productivity.
Preview
In this series we will take you through the most prevalent BI implementations in
each of the following parts. We look into implementations of Data warehouse,
OLAP (On Line Analytical Processing), ETL (extract, transform and load), Data
Mining, dashboards, reporting and web based analytics. We take a sneak peek into
all of these before we get into thick of things in the succeeding parts. A Data
warehouse is a repository of data structured in a way so as to provide
comprehensive details of all the aspects important to the organization's
information user. These aspects can be customers, vendors, business processes,
or even transactions. The data is provided with a time dimension in view, and
also with some volatility.
Data warehouses are quite often constructed by combining Data Marts in a
particular architecture. These data marts that form the units of a warehouse,
contain a subset of the data of the warehouse and each of them serves as the
repository for a particular set of users, or community, or even a process within
an organization. A Data Warehouse forms the first part of most of the BI
solutions that are implemented, and organization now even looks at the warehouse
in conjunction with existing data collection sources such as ERP. The process
defined for building a warehouse requires a thorough analysis of the
organization and its processes as each of the marts being implemented in the
warehouse target a specific subset of data as aforesaid.
ETL refers to the methods involved in accessing and manipulating source data
and loading it into target database. An ETL process takes care of the fact that
the data reaches the warehouse in a state that has been defined for the
warehouse. An ETL process starts with mapping the data between the source system
(that can be an application, a file, database, or any data collection source for
which the warehouse is conceived), and the target warehouse or mart. The
mappings we referred to over here are definitions of the relationships and the
data flow between source and target systems and objects.
What follows next is 'Cleansing' the data in a staging area, ie an area
where data is processed before it enters the warehouse. Cleansing of data refers
to removing anomalies in the data with respect to a warehouse. Data mining forms
one of the most critical processes that run on a warehouse. It is defined as the
extraction of information that can be trivial, or even previously unknown. The
process is also popularly known as 'Knowledge Discovery'. Mining involves
usage of, statistical and visualization techniques for discovery of information
and presenting knowledge in a form which is easily comprehensible to humans.
One of the most common implementations of a BI solution is an OLAP (On Line
Analytical Processing). It is one term that lacks a standard one size fits all
procedure of implementation. In fact its definition is also something that
varies depending on the vendor's providing tools for it.
In a nutshell, OLAP can be defined as a process that provides analysis of
structured and multi-dimensional data in quick time. If there is one word that
characterizes OALP precisely it is 'Multi Dimensional'. Querying on tables
usually involves working with a two dimensional flat structure which databases
call 'table.' OLAP on the other hand is far more complicated than that.
Typically in OLAP we optimize the tables into one of the schemas that we'll
define in detail in later parts, and then analyze the structure to define
multiple 'Dimensions' over which the tables will be queried. These
dimensions are nothing but those variables, in whose relation, we need to find
out information. For example for an OLAP implementation that uses tables from
sales department for an organization, the dimensions defined could be
Salesperson, Product, Sales Region, Distributors, or even a particular city and
most importantly a time unit.
Dashboards have caught up as one of the important sources for decision
analysis and they completely depend on accurate visual modeling. They help an
organization to go beyond usage of spreadsheets in the sense that they get an
interface that is a single point of entry from which all relevant information
can be viewed, interpreted and acted upon. Though spreadsheets are very easy to
deploy and provide a dashboard type of functionality, they mostly generate
static reports.
This is where dashboards score over spreadsheets as they provide you a visual
model which is highly interactive, and can provide relevant information as per
the specific needs of a particular information user. Dashboards typically track
performance or decisions or similar actions that are the KPIs (Key Performance
Indicators) for an organization. These KPIs are similar to the variables we
talked about while previewing OLAP. In fact the Dimensions used over there are
almost always the KPIs.
Web-based analytics is a fairly new concept, not that it has not culminated
into implementations, but it's still pretty early days for this concept. As
the name suggests it delivers standard analytics of data to the user over a Web
browser using Web technologies. Web-based analytics, unlike other technologies,
we described earlier can offer both the internal and external users the ability
to analyze data specific to their particular needs.
The analyzed data is then used to create graphics to display the results as
needed from associates to senior executives to sales and marketing guys. For
example, sales representatives can analyze customer purchases generated through
normal sales channels. This information can be analyzed, graphically displayed
and sent on to product managers, manufacturing personnel and senior executives
who can then make real time decisions as needed.
What's available?
There is a wide range of products available for BI implementations which
includes some of the big names in the game. To start with we have SAS, which
have their own BI platform, and also Enterprise Intelligence platform which
includes Integration, analytics and storage tools. SAP also provides BI
solutions based on AG platform and Netweaver. IBM also has a comprehensive range
of products and tools for BI. The DB2 database has a Data Warehouse Edition that
also allows you to define warehousing processes.
IBM also provides Cube Views, their tool for building cubes for OLAP and also
a tool for aiding Web based analytics, called Web Services for Cube Views.
Microsoft has integrated BI capabilities in MS SQL Server itself. The MS SQL
Analytics Services, and the Visual Studio Team Foundation System also includes
features for BI solutions development. Among these is Pentaho, an Open Source
initiative that provides a detailed set of tools and a BI suite as well. Their
tools include a complete BI suite, a Design Studio, and a Reporting Wizard.
Cognos offers business application tools available under Cognos8 BI. It has
architecture for reporting, analysis, scoreboard, dashboards, business event
management and Data Integration. Oracle has a BI suite to cater to corporate
performance management system to assist you in making analytic and tactical
decisions. Suite includes performance management, interactive dashboards and of
course the analytics. It supports ad hoc query, analysis, alerts, reporting and
predictive analysis. A medium order BI solution is MicroStrategy8. Business
Objects has a couple of applications for businesses. These are catagorized under
a BI platform that has business object enterprise, integration kit to integrate
with PeopleSoft, portal integration kits, SAP, Seibel, etc, and analytic
engines.