"It is not clear that intelligence has any long-term survival
value" -Stephen Hawking, British theoretical physicist
For our friends dealing in intelligence, this is not good music.
But clichés apart, the consensus is that Business Intelligence or BI as it is
popularly called, is a useful tool; not the Harry Potter-kind magic wand as some
would desire, but still fairly helpful.
The consensus also is that BI is an oft-abused term. Working on
Excel is a kind of BI, because one is making some sense of the data he has. But
for all sensible purposes, we are taking into account and talking of the
essential four-data ware-housing, extract-transform-load (ETL), analytics, and
reporting.
It is here that some CIOs report skeletons in their cupboards.
Take the example of Monsanto, a leading provider of agricultural products and
solutions. It has been using a BI solution since 2000 that helped successfully
plan/manage its forecasts from a financial and sales perspective. But there are
also expectation mismatches, as the company's IT head, Anand Kumar outlines.
I |
Anand Kumar, Two, ensure that organizational change Lastly, keep in mind that reporting is a CN Ram, HDFC |
Conceptual idea 1: Response to user's need for data on a timely
basis.
Reality:Â "While we have been able to develop some good reporting
capabilities, which help in the forecasting and the planning/reporting arenas,
what Gartner calls 'information democracy' has not yet set in. We are still
faced with challenges in terms of data integrity because of the often-used
paradigm 'garbage in, garbage out'. Systemic thinking evades us, but the good
thing about it is that users are becoming more and more aware of the challenges
of working with compromised data." So Anand is now looking at dashboards and
analytics as the way to go forward.
Conceptual idea 2: Delivers Just In Time (JIT) information.
Reality: "The information lifecycle is often prolonged due to
various issues in the supply chain and other customer facing departments, and,
often, infrastructure/training/systems approach are key challenges in this arena
that prevent JIT information to be available. There are also issues surrounding
getting partners/suppliers to participate and collaborate in information
lifecycle management to shorten the time taken for information to travel in the
chain."
Conceptual idea 3: Decreases cost and improves operational
efficiency.
Reality: "We have been able to reduce some cost and improve some
aspects of operational efficiency, but capturing RoI on this has been extremely
challenging since, in a sense, resource management and planning are areas where
we have had the most impact. Operational efficiencies on a large scale can only
be realized once we have all the key stakeholders involved in the BI
process-suppliers, retailers, and our partners as well."
CIO Check |
size=1>MicrosoftÂ
Team
SSA
|
Happy Home: The HDFC Bank, on the other hand, has been
able to overcome such challenges through an enterprise data warehouse solution,
Reveleus, a product of i-Flex. It started building the warehouse about five
years back and has grown in size and complexity now. The bank, says IT head CN
Ram, has seen immediate improvements in profitability through cross selling
opportunities and decreased channel costs through a better understanding of the
customer's channel utilization.
The bank is also using the data to develop targeted marketing
campaigns and to offer new services and products. The data warehouse is now the
primary source of information for its entire retail business. In addition, while
the data volume has doubled over the last year, the resources needed in the MIS
department to maintain the systems have remained the same.
"Using analytics from our warehouse, we have been able to
successfully migrate customers to direct banking channels, bringing down the
share of branch transactions from 60% to less than 35% now. The turn-around for
retail credit risk related MIS as well as information on collections management
has reduced drastically from days earlier to a few minutes now," Ram tells. And
the RoI went much beyond this.
BI @
Oracle
Oracle India recently introduced Oracle
Business Intelligence Suite, which is an open, standards-based software,
designed for all users and business processes throughout the enterprise. Oracle
Business Intelligence Suite unites Oracle's Business Intelligence (BI)
middleware with Siebel Business Analytics. When combined with Oracle E-Business
Suite, Oracle's PeopleSoft Enterprise, Oracle's JD Edwards EnterpriseOne,
Oracle's Siebel pre-built analytic applications, the Oracle Business
Intelligence Suite addresses the complete spectrum of BI and analytical
requirements, including historical, real-time, and forward-looking insight, with
data integration, and management. The suite, along with other software
components of Oracle Applications, Oracle Database, and Oracle Fusion
Middleware, delivers enterprise-wide business intelligence that is
'hot-pluggable' into existing operational
systems.
Customer data integrated in the warehouse enabled HDFC to
understand the customer at an individual level rather than at the portfolio
level. The ACRM tools further helped it to use this data to build profitable
conversions with customers.
Continuing with the lift HDFC's product managers received with
improved credit card activation, higher credit, and debit card spends
significantly improved cross sell ratios; the year 2005-06 witnessing over 650
unique campaigns. More than 60% of these campaigns use a statistical model.
Marketing analytics now focuses on strategic priorities such as proactive
retention and balance build up.
The bank now has over 10 mn customers who avail
of one or more of its products and services. It can now know how many products a
particular customer avails of; how frequently he uses a particular product; and
the correlation between the usages of one product with that of another. Each of
these answers provides it with valuable insights into customer needs, helping
the bank keep them happy.
The analytics supporting credit risk helps the credit manager in
optimizing the risk-reward ratio by monitoring parameters using analytical
domains, reports, key performance indicators, and dashboards in addition to the
following:
Factors that influence the business trends and pattern of various
products
Factors that contribute to the performance of each product
vis-Ã -vis the others
Customer behavior with respect to vintage
Credit risk of lending portfolio
Performance appraisals based on various vintage cuts
Sourcing channel reviews
Right Pickings: If enterprises need the right answers to
difficult questions, the market has to have the right vendors with varied
portfolios. India, when compared to the US or Europe, is still in the nascent
stage as far as BI adoption goes. It's very small, and very new with
Frost&Sullivan pegging the market close to $45-50mn.
So what are the choices for Indian CIOs? The country has pure-play
players such as SAS, Business Objects, Cognos, and big time vendors who bundle
BI solution along with other products-SAP, for instance, bundles BI with
Netweaver and MySAP, and Microsoft bundles BI solutions with SQL server. The
preference for either of these two categories would typically depend on the
needs of the enterprises. "One is functionality. A bundled product that comes in
with SQL server or MySAP would kind of satisfy your desire to do elementary BI
reports. But if you want advanced analytics and reporting methodologies, then
one would start looking at pure play vendors," says Rajat Sharma of
Frost&Sullivan.
Yes Bank, which is planning to implement a BI solution, has
similar aspirations. It is looking for a platform to address its enterprise
requirements of Basel II Reporting, Anti-moneyLaundering, Asset Liability
Management as well as MIS Reporting. The technology framework needs to have a
number of components such as an ETL tool, staging area, data warehouse,
database, cube, and a reporting platform, which support different plug-ins or
adapters to different databases and file formats in order to provide a flexible
way of working. For the bank, it is important for the system to provide a
meta-data layer where definitions about data extraction, translation, and even
the complex analytical queries can be externalized leading to both flexibility
and putting the power in the end-users hands.Â
"A good mathematical analysis and
computation tool should also be provided for doing sophisticated correlation,
regression, and other analysis required for creating campaigns and gaining
business insight. It is very important that different business segments within
the bank can customize the views and reports according to their
requirements. Last but not least, the product should come with some standardized
data models to address Basel II to ensure that the implementation is rapid,"
demands CIO Aditya Menon.
Other enterprises are demanding that their data is represented in
every division or department of their organization in terms of empowering either
divisions or the user to take informed decisions based on a single view of where
the company is. So, if one has divisional tools without a single view, the
decisions made by one department will contradict decisions made by another. So
many companies are actually demanding that there is consolidation of vendors
that they are dealing with.
In a competitive, dynamic world, they cannot shift
the blame between vendors because of the errors that occur often. "Customers are
also demanding that we speak their language; so BI has to be directed towards
their industry. They are also demanding speed to intelligence, time to
intelligence, ability to implement very quickly, deliver short-time RoI. That is
because this investment is not incorporated into their transactional
system-on-time investment needs to derive short-time RoI," says Don
Cooper-Williams, director, Marketing and Alliances, Asia Pacific, SAS.
But are customers really looking for integrated suites? SAP
executives have been saying that BI will cease to be a market in about five
years. Executive Board Member, SAP AG, Shai Agassi told Dataquest sometime ago
that BI is not an isolated thing: "It's like saying that in your business, there
are some people who think and some people who don't. You can't have that. People
who don't, need to think. You need an integrated analytics market, not a
separated disconnected one. That's what we are doing right now.
'Operational efficiencies on a large -Anand Kumar, IT head, | 'The operational systems of Indian - Sudipta K Sen, CEO and MD, SAS  | 'The product should come with some -Aditya Menon, CIO, Yes  | 'Customers are also demanding that we -Don Cooper-Williams, director,  |
Pure play vendors such as SAS counter this by saying that SAP is
making an error in judgment. There is a huge difference in creating data from
transactional applications and driving that data into normal data warehousing
type data warehouses-those data warehouses are not optimized for BI. They are
not optimized to draw data from more than one database. They cannot work across
multiple databases and they do not have the scalability and efficiency that is
needed from a BI point of view.
"They are making these statements because growth
in the ERP market is slow at the moment. And the growth for BI is beginning to
pick up. Secondly, they have not developed the applications for the solutions
that will sit on top of their transactional systems. This requires an immense
amount of R&D. The question I would ask is they are not going to divert
their R&D from transactional area into an area now called BI. They must be
looking for companies for acquisition. However, if you look at the top
acquisition companies in the market place, none of them have an integrated
platform. So, there is a lot of opportunity and growth prospects in the market
place," explains Cooper-Williams.
The third thing SAP and similar vendors claim is that they might
buy the technology they do not have. "But they don't have the culture within
their organization of selling this. So, they would need to drive and build
that," Cooper-Williams reacts. The issue with BI as opposed to typical
transactional-type applications is that there are time issues around them-around
governance and compliance-time schedules about why somebody would want to use
BI.
"By and large, these companies have failed to even give good reporting from
transactional system and enterprises had to use third party reporting
capabilities. You need to build an extreme amount of credibility in the market
place. I believe they will have some traction and the traction will be in the
large accounts where they are dominant. None of them have done anything more
than build CRM from a transactional point of view, build any intelligence, Sibel
included. These organizations are single in terms of databases they work on.
Whereas we, over the years, had a more catholic view of the market place, are
agnostic to databases, are built on world-wide standards, and have huge customer
bases that are recognized," he says.
Fresh tale: CEO and managing director of SAS India
Sudipta K Sen feels that the operational systems of Indian enterprises have
matured and they are increasingly looking to differentiate themselves from
peers. Mature companies are, therefore, looking beyond the classical BI usage of
reporting and are moving into predictive analytics. A good example is an Indian
bank with a network of 560 branches, which implemented SAS enterprise-wide a
year ago.
The bank wanted a reporting platform that is web-based. A lot of
reports went out as Excel. Spreadsheets are typically unprotected and data can
be shared across the organization, even outside the organization. It wanted that
somebody who didn't have the rights to a data should not see it. It should be
protected so that it cannot be shared outside the organization. Third, it wanted
that people get only that amount of data, which was required-since they pull it
over the web, huge reports can waste bandwidth on e-mail.
Earlier, these reports usually were sent during working hours,
which meant that it was a time when the branches were crowded and people were
trying to access the core banking and loan systems. The bandwidth can get
clogged up by e-mail if multiple MBs of files are sent. Whereas, the reading of
the reports typically does not happen during peak hours; it happens after hours.
So having a web-based platform meant that people will only pull reports down
when they needed to read it. The bank managed to shift a lot of bandwidth
consumption from peak to non-peak hours.
The BI solution is yet to be fully rolled out across the
organization. But the bank says the RoI envisaged, has started to roll. The
tangible RoI was bandwidth. The intangibles were increased security, increased
efficiency because only what is relevant is coming-once somebody has to come
down from going through multiple MBs of data to just having the relevant cuts,
it makes a huge difference in efficiency. The bank had a decent enough view of
the business even before the present solution was implemented. What the
enterprise intelligence platform did was to present it much better and target
information to the right people.
The deployment challenges were mainly cultural in nature. People
who have been writing SQL scripts and that kind of queries, churning out magic
out of a database that a layman cannot understand, seem to loose their position
of power. Now, even a business user could tap into the system and pull out
reports. The other thing is insecurity, because there are specific skill sets
that are no longer required.
CIO advice, therefore, is to start managing re-skilling right from
the beginning; be ready for resistance from people whose skills would become
redundant. Two, look at the cost of ownership and not at the cost of
acquisition; ensure that deployment is part of the acquisition plan. Initially,
the bank in question did not put deployment costs as part of the fixed costs. It
had estimated deployment costs, cycles. Cultural change issues come up and delay
things. Then, variable costs are likely to spiral.