The buzz about operational biz intelligence
- By Stephen Swoyer
You can't say Information Builders
CEO Gerry Cohen doesn't speak his
mind. You might be tempted to call him
a straight shooter. For example,
even though IBI is sitting on one
of the niftiest data integration
technologies on the market—its
iWay family of connectivity solutions—Cohen dismisses data integration
as an over-hyped buzz term.
Instead, he's abuzz over what he calls
operational business intelligence—the
use of BI technologies in conjunction
with conventional enterprise applications
and systems. OBI is an untapped commodity,
Cohen says, that can result in substantial
cost savings for adopters. Of
course, competitors such as Actuate,
Business Objects and Cognos might beg
to differ, but Cohen remains undeterred.
If 2003 was the year BI reporting
went big time, 2005 looks to be a year
in which data integration will follow a
You don't think data integration is a little
Well, hype or no, it's been in the news
an awful lot this year. But I'll bite—why
do you think it's over-hyped?
Data integration has been around for
umpteen years. There used to be data dictionary
products. Data integration is largely
a metadata description of a lot of data.
It's getting a lot of play because after IBM
bought Ascential, they put Ascential into
the DB2 group, and then they started to
talk about data integration and started
talking about building a warehouse. Different
people have different understandings
of what data integration means.
If you're MetaMatrix, the software
company, it's describing all of your data.
It's a nice term. It's necessary to have uniform
descriptions of data across different
islands. But what's driving data integration
more than anything else is not the BI
side, but the B2B and what I would call
the application integration [sides].
You've got a heck of a data integration
story in iWay, with
connectivity into an enormous
number of sources, but—from my
perspective, at least—it's a somewhat
muted story. That might be consistent
with your positioning of [iWay] as more
of a B2B or EAI technology, but clearly
it has BI strengths, as well. What's
your strategy with iWay, and do you
plan to kind of amplify what you're able
to do with it in conjunction with your
iWay software is really an EAI company.
Now we call it an SOA middleware
company, because that's what the buzzword
is. In the BI space, yes we do join different
worlds together. We are a very
powerful engine. We have a lot of very nifty
applications that are both iWay
and WebFOCUS. But they're
very specific in some sense.
We have an application called
Syndromic Surveillance: it picks
up all of the hospitals in a radius
that take up emergency-room
people. iWay picks up all of the
various people's information who come to
the emergency room; iWay brings all of
that information, transforms it into something
common and puts it into a database.
WebFOCUS is a reporting engine off that
database. That's the kind of [BI] application
I'm talking about.
There's been some talk about BI
standardization—or, more accurately,
BI tools consolidation—much of it propounded
by the major suite vendors.
The idea, if I can paraphrase them, is
that organizations want fewer discrete
tools and more integration, in large part
to help drive down costs. I wanted to
get your thoughts on a couple of things
in this respect. First, can IBI offer the
breadth of feature and functionality
coverage that a Business Objects, Cognos
or Hyperion can?
I think we're the only ones who can. We
offer what's also known as operational BI,
where to a great extent they offer what I
would call analytical BI. Business Objects
and Cognos, they basically sell systems
like they would in the old client/server
world—small deployments, where you've
got 50 people who need to get information
out of a particular database. In operational
BI, what we're doing is applying BI to operational
systems where you can have
1,000 or 5,000 simultaneous users. That's
a really big system—a total of 2 million
users with as many as 5,000 users online
at the same time.
Isn't standardization somewhat unrealistic
or at least misguided
as a trend, though? That
is, will organizations find
they're going to have to rely
on at least a few best-of-breed
BI tools to get competitive
No company is going to say, "I
want to get more vendors." They're going
to say it's more economical to have fewer
vendors. Companies have been saying that
for years in every category. We're overdoing
this best-of-breed; let's cut back on vendors.
However in the BI space, it's very
likely that large companies will have multiple
vendors. We have been involved in
quite a number of situations where a company
has taken seven down to three, and
that makes sense. And we feel that we're
going to be one of the three, so to speak.
When you think about the requirements
of operational BI, it takes a certain kind
of architecture to do that. We have clustered
machines with failover with automatic
load-balancing. We're real heavy
duty. No one else can support 5,000 concurrent
users like we do.
You mentioned what you call operational
BI. Would you say that's similar
to the operational reporting companies
do every day? If it is, how do you propose
to differentiate IBI's strategy
from those of competitors such as Cognos,
Business Objects or Hyperion,
which have moved more aggressively
into operational reporting and BI?
I think this [operational BI] is quite
important. I also think it's something
that's unique for us. We're applying BI to
operational systems, systems that companies
rely on to run the company. If you
have intelligence in those systems, you
get payback, more return on investment
for those systems. We have a system
where phone companies distribute to the
workmen in the field all of that information
about what they're doing, fixing Mrs.
Jones' phone. So the phone guy can actually
see how many jobs he's completing in
a day, compared to his coworkers.
When you have 5,000 people basically
controlled by a scheduling system that's
the heart of the enterprise, there's a
chance of getting a lot more savings. The
guys in the field discovered an error in the
scheduling algorithm that told them
where to go, so that their service calls
weren't scheduled as efficiently as possible:
that's an enormous savings. The improvement
they've made in that system,
they claim it's $1 million a month. That's
what I mean by operational BI: the unanticipated
but significant savings you get
when you bring intelligence to those [operational]
The big BI suite vendors have
fleshed out their analytic expertise with
operational reporting solutions, however.
And they say they can scale to support
these kinds of requirements. On
the flip side, what kind of analytic
technology can Information Builders
bring to the table?
We have a complete, beautiful set of
tools for analysis. We have an OLAP
methodology that can work on any database.
It's a metadata layer over the database
that makes it look dimensional.
We have end-user tools, and we have a
[technology] that we call structured ad
hoc, where if you want to get any reports
on your own, you bring up a form and just
go click-click-click, and we generate a
form for you.
We can parameteratize and we can get
to anything you ask for on a form; it's just
a matter of designing a form that's appropriate
for you—and that's useful for very
large numbers of users with no training.
The enterprise applications space has
thinned out considerably over the last
30 months. J.D. Edwards is no more.
Nor PeopleSoft or Siebel, for that matter.
For all intents and purposes, there's
just Oracle and SAP, with a host of
smaller players dotting the landscape.
It's probably too early to expect that so
drastic a culling will occur in the BI
space, but there's already a pronounced
trend toward consolidation. I wonder if
you think that history—in the ERP
market and elsewhere—is a reliable indicator
of what's to come?
It probably will come, but it's not as imminent
as it was in the ERP space. Even
in the ERP space—now that PeopleSoft is
gone, J.D. Edwards is gone—we see a lot of
activity in the smaller ERP players.