Tread carefully to find holy grail of clear data for
practical use
Business intelligence is at a crossroads. Commentators are
increasingly asking in what direction the softwareis heading and
whether, amid all the consolidation and competition, its original
purpose is being forgotten.
The basic reason for having business intelligence systems is to
provide a single, consistent and robust view of an organisation.
However, they have changed substantially in the past five
years.
There used to be three basic categories of tool:
- Online analytical processing (Olap) applications were used to
build huge databases called "cubes". These were designed to answer
a wide range of questions.
- Management information systems (MIS) drew on the functionality
of SQL Server to answer particular management questions.
- Executive information systems (EIS) were dashboards that could
pool information from various sources to aid executive
decisions.
But the benefits and limitations of each of these approaches has
led to a merging of the categories.
For example, Olap is powerful but not particularly
user-friendly. If a cube of quarterly sales data is assembled, it
may be out of date by the time it is ready. Should an executive
want to ask a question of it that was not anticipated, this could
cause problems. So Olap suppliers have developed more intuitive
front-ends and geared the tools for the real-time enterprise.
Conversely, MIS and EIS applications have tried to capitalise on
their flexibility by moving in the opposite direction: they have
gained analytical substance by copying Olap characteristics. In the
process, suppliers have been bought and sold, brands have
disappeared and others have gained ground.
Jon Crews, a partner at PA Consulting Management Group, divides
the growing number of products into three types. First is
enterprise-wide business intelligence, from companies such as
Cognos and Business Objects, which are substantial enough to
provide a corporate-wide view, and robust enough to supply data
feeds to the supply chain and customers.
Second, business intelligence is bundled in with other corporate
applications, notably from SAP and Microsoft.
Third, there are niche applications for particular needs. "For
example, Hyperion is often thought of as the business intelligence
tool for accountants, and SAS has found niches in a variety of
areas," said Crews. Other suppliers include Microstrategy, which is
particularly strong in retail, and Applix for finance and
insurance.
Overall, sales of business intelligence tools are growing.
According to the annual Olap Report, the market continues to grow
faster than most other enterprise software sectors, increasing by
16% in 2004.
IDC has predicted the global market for business intelligence
products will grow to £3.1bn by 2009, with Business Objects the
leading supplier. The analyst firm said the company is growing two
and a half times faster than its nearest competitor.
However, it is hard to make comparisons between different
products. Business Objects software may be regarded as the leading
pure-play corporate business intelligence tool, but Microsoft
products are widely found in smaller organisations.
"A continued maturation of this software market is leading to
increasing adoption of business intelligence tools throughout
organisations of varying sizes," said Dan Vesset, research
director, analytics and datawarehousing at IDC.
It is a market replete with choices. There are premium products
that, for the right price, will do everything. There are low-end
products that satisfy the needs of specific business
decision-makers. But analysts are concerned that the basic value of
business intelligence systems is being clouded by market
consolidation and competition.
Frank Buytendijk, research vice-president at Gartner, believes
the central issue is whether business intelligence enables the
chief information officer to present the chief executive with one
version of the truth. He suspects that although business
intelligence has come down from the top of the "hype cycle" (when
supplier claims reach their peak), gaining one view of the
organisation may be as difficult as ever.
The past few years have not been without progress, however.
"What has happened is that corporate infrastructures have become
much more centralised. Companies no longer have multiple business
intelligence tools all over the organisation," said Buytendijk.
"That business intelligence is now infrastructure- heavy and
application-light is the right balance, but the functionality still
needs to be pushed back to the users."
However, other commentators are not so sanguine about the
centralisation of the infrastructure. They point out that if you
put bad data in, you only pull bad information out. "The data layer
is the biggest problem," said Royce Bell, a partner in Accenture's
Technology Practice.
He said much has been spent at the functional level, orienting
business intelligence to business needs - for example, in response
to regulatory demands - and this has led to silos of
information.
"People have put single enterprise-wide databases together in a
few cases, but mostly companies create a warehouse to deal with
financial analysis, compliance needs, and retail requests: multiple
warehouses that then exist in silos," said Bell. "And the amount of
data in the organisation is only going to grow, when you add new
sources of data, such as RFID flows."
This has been referred to as information entropy, the idea that,
left to its own devices, data becomes more and more chaotic. "What
is required is an integration plan at the highest level," said
Bell.
Bell's thoughts are echoed in research from IT services firm
Unisys. According to a survey of 250 European IT directors, 60%
believe their business intelligence investments have yet to
significantly improve customer management; 27% of companies cannot
cross-sell between divisions or products; and 32% of all customer
data still requires manual analysis.
Steve Rawsthorn, Unisys vice-president of sales and marketing,
systems and technology EMEA, said many companies think they have a
business intelligence strategy once they purchase a business
intelligence application.
"CIOs need to ensure that there is an infrastructure in place
that can monitor a mass of data streams in real time, enable huge
databases of information to be shared across departments, and
provide sufficient storage capacity to analyse data quickly and
effectively," he said.
"Inadequate systems specification means that, despite
implementing the most sophisticated software, companies cannot even
mount all their customer data simultaneously for analysis or to
cross-sell."
Although organisations are using business intelligence products
for a range of tasks and are keen to do more, questions remain
about the direction suppliers are taking. Systems may generate
useful corporate knowledge, but to exploit them further companies
might have to get smarter in using them.
Case study: Software powers Powergen campaign
strategy
Energy supplier Powergen is using business analytics from KXEN
to fine tune the way it targets customers and prospects with the
most appropriate energy packages for their needs. Even early on in
its implementation the results are encouraging, according to
Powergen's head of customer relationship management Mark
Perrett.
"The first time we used a KXEN-generated model to support
campaign activity we saw a 20% rise in sales and a direct £150,000
saving in mailing costs. Those figures represent an excellent and
speedy return on a software investment. More importantly, we also
retained 300,000 customer contact opportunities for future
campaigns, having been told by KXEN that a successful sale was
unlikely to result this time."
Only four analysts were needed to do the modelling. Perrett said
he has heard of other organisations where many more analysts were
needed to do similar work. Also, rather than taking several weeks
to build one model, the software took minutes.
KXEN puts its flexibility and speed down to the use of
mathematical rather than traditional statistical techniques when it
comes to building models.
"We did get KXEN initially for a very specific purpose but on
the strength of our early experiences we are going to invest some
time in exploring its other capabilities," said Perrett.
Case study: Peacocks tackles till-based
fraud
Clothing and homeware retailer Peacocks has installed Fraud
Alerter software from Innovetra to help it tackle the problem of
fraud at the point of sale.
Peacocks will use Fraud Alerter to analyse till data from 420
stores. The software will identify suspicious transactions and
highlight excessive use of tills' training mode, which can be used
to disguise fraud.
The increased speed and efficiency with which the firm's loss
prevention team will be able to act based on the data is expected
to cut losses from till-based fraud and provide a strong deterrent
to staff.
"We are confident it will play a significant part in preventing
fraud and stock loss," said Chris Miles, operations director at
Peacocks.
"The data is mined overnight and the loss prevention team will
have indicators of suspicious activity on their desktops first
thing each morning, enabling them to start investigations
immediately. Previously, such information was not available until
two or three weeks after the event."