|
30 Minute Interview
The Power to Predict
Ram Menon, Executive Vice President, Worldwide Marketing,
TIBCO Software Inc talks to Prashant L Rao about his company’s predictive
technology that’s based on both static and dynamic data where you try to
anticipate what is happening by mapping dynamic data onto historical information
to detect patterns as they emerge.
Ram Menon |
There’s a lot of buzz around service oriented architecture
(SOA). Is SOA all there is to enterprise IT architecture today or is there something
beyond it?
|
Predictive business is about gaining
insight into the future. Users are trying
to decompose their black box applications into services
|
In the past, people have approached the whole idea of an application from the
standpoint that it’s hard-coded, you get a black box and the application
vendor tells you how to use it. This black box sits atop a database and the
enabling architecture was a database-oriented architecture.
The traditional killer application in an enterprise has been ERP while the
analytical tool has been what we call Business Intelligence today. That has
been the stack at many companies for the last 20 years.
However, data and information are not static any more. People want to know what
happened a few minutes ago and then compare it with events that took place last
month or last year. Data can be static or dynamic; it’s in motion.
The enabling architecture of the future is SOA. It confers the ability to use
Web Services to create services that are spread across the enterprise and can
be combined like Lego blocks at will.
The killer application of the future will come from a customer
and not from a software company. The business process where a customer takes
an order and converts it into cash can be automated using BPM. The killer application
of the future will be enterprise IT encapsulated into a BPM tool.
Predictive business is all about gaining insight into the future. This is the
new stack. Users are trying to decompose their black box applications into services.
The heterogeneous nature of an enterprise architecture requires a company to
create Web services on all these platforms.
Is there a limit on what a company can hope to accomplish
with predictive technology?
You are only limited by your imagination, creativity and technology smarts.
We have folks who are trying to predict the speed of trains. In the railway
system, if a train goes faster by even a mile per hour, the amount of diesel
that is consumed goes up considerably.
Folks are trying to predict breakdowns in the telecommunications system. Financial
services companies are attempting to manage customer loyalty in credit cards.
You have BPM, business processes and predictive technology.
The use to which this technology can be put to is limited solely by the business
that you are in and what it is that you want to anticipate. This is not stargazing
or astrology. It is about looking for patterns in static and dynamic data and
trying to anticipate threats and opportunities.
There’s an emerging area of computer science called
complex event processing. Tibco has been working for many years with Professor
David Luckham at Stanford University who is known as the father of complex event
processing. What we do is basic complex pattern recognition.
| Harrah’s Casino is the largest casino chain
in the US. Customer loyalty is important to its business.
When a customer enters the casino, swipes his card
and plays the slot machine, has lunch at the buffet and attends shows,
all these events are connected by the software system.
With a high probability of losing money, it is
difficult to get a customer to keep coming back. The information about
how much a gambler lost is put together in real time and the software
tells the casino that once this particular gambler loses x$ he stops gambling.
The software might tell you that the customer likes magic
shows. A lady turns up as a good luck ambassador and gives the gambler
two tickets to the magic show. Maybe then he’ll come back to gamble.
|
| Becoming a predictive business demands more than
speedy technology. A predictive business is agile. It must proactively and
rapidly address the signals that it receives about opportunities and threats.
This requires a new style of leadership, a renewed and closer focus on customers
and close collaboration between business and IT managers. This transformation
is not an event; it is a process that takes time. Introducing predictive
techniques is best done in stages, addressing one or two business processes
and then rolling it out throughout a company’s operations.
— Excerpt from The Power
to Predict by Vivek Ranadivé
|
|
