Grinding It Out
How can the #3 apps vendor gain share?
Filling in the Gaps
Several years ago, a well-known financial analyst named Phillips accurately predicted a downturn in the applications market, saying (among other things) that the ERP market was penetrated.
I argued then that the market was not so much penetrated as inoculated. Yes, many large companies had
made some decision and installed an ERP system somewhere. But the ratio of seats bought
to seats that could be sold was still quite small.
History has proved us both right, I think. On the one hand, the large-company customer bases
for most of the application companies were indeed largely established by 1999.
On the other hand, total license revenues keep on going up.
Some of this revenue is from the infill I predicted: people putting SAP or whatever into countries
and divisions and functional areas that weren't hit the first time around. But
more than I expected comes in because the companies have more to sell. They've been busy
improving old products, shifting architectures, and adding lots
of new products.
In the process, the ERP companies have shifted away from being platform companies, which sold a big piece of software that managed some core operation, like financials or HR. Today, they are portfolio companies,
each selling a Chinese menu of products that are built on top of those platforms. At least some of the growth in license revenue comes from the sale of those portfolio products.
Portfolio products have complicated the job of the analyst enormously. Now, if you want
to evaluate the company, you can't just get an hour briefing on the
vision and talk to a few customers. Theoretically, you should be keeping track of
all the products in the portfolio.
It also hasn't made running an applications company any easier. Now,
your highly expensive programmers are no longer working on a single, horizontal app that has an enormous potential market. Most are working
on much more specialized applications, which are intrinsically more difficult
to build, and can only be used by a relatively small segment of the market: an
industry, say, or a functional area that previously hadn't been seen as a
buyer of seats on the original platform.
Even though segments are smaller, going to market
within a segment is more challening. You need expert marketing, or the segment doesn't take you seriously. And somehow or other, you have to train a sales force that can speak to the segment. And how do you do that? Train everybody badly? Or assemble a small group that may have difficulty finding and reaching the opportunities.
It's a hard game to play. But with the core platforms relatively commoditized (at
least from a marketing point of view) and all the application companies
presenting a strong face to the public, it is inside these portfolios that
the battle will be won. Football metaphors occur easily these days to people
from New England, but it is as if all the application companies are football teams,
whose success depends on the footwork of the interior line.
Managing a Portfolio
Are there winners and losers so far? Over the past six years, two companies do stand out as leaders: SAP and Siebel.
SAP was perhaps the first to move to a portfolio structure. It was some 8 years ago
when SAP first set up industry organizations called IBUs or
industry business units. Particularly in industries that SAP dominated, like oil and gas or specialty
chemicals, the IBUs provided a marketing and development apparatus for what I'm calling portfolio products. The specialists who were at the core ran customer advisory boards,
developed industry maps, and told a few overwhelmed programmers what to do.
The IBUs have never been the home run that SAP hoped they would be, and support for many of them has been halfhearted. But over the years, hey have actually been able to develop some quite good products. (See my Short Takes from last year's Sapphire for further discussion.)
Siebel's portolio products were built up rather
more rapidly than SAP's, either because their scope was narrower or because Siebel was better at this than SAP. In my view, the difference lies in the talented pool of product marketing managers whom Tom Siebel seemed occasionally to drive personally.
When these people didn't succeed, this structure made it
easier to pin blame
than it might have been at SAP, and this
fact has been a real boon to many small companies in the Bay Area, whose Siebel
product marketing managers now work for them.
At the same time that Siebel and SAP were making these moves, Oracle also
tried to develop industry verticals.
Some of their talent was very good indeed. George Kadifa, now CEO
of
Corio, and Mike Capellas, now at MCI, were at one time responsible for A&D and Oil & Gas,
respectively, at Oracle.
But for some reason, Oracle did not follow through.
One year, Oracle was going to be the leader in
one of these verticals; the next year, George or Michael are working
somewhere else.
This may have been a very sensible thing for Oracle to do.
Neither Siebel nor SAP has been able to develop portfolio products
with any great reliability. Many, many initiatives at all the companies have been announced, invested in, and forgotten. And as the companies get burnt, they become more reluctant about leading and more willing to leave development until a market is proven. As one developer at one of these companies said, "You just have to get used to being five years behind."
The Oracle Applications Portfolio
Whatever Oracle did in the past, it is clear tht today they are trying
to play the portfolio game.
To a great extent, the presentations at AppsWorld were organized in a way
that made it simple for Oracle to lay out its full portfolio. the first day
was devoted to industry verticals; the next two to functional segments.
Any analyst who could have been in 8 places
at once would have had a good chance to see the entire thing.
Not having that much energy, I focused on three segments: the CPG vertical, compliance,
and sourcing. I know a fair amount about CPG and sourcing, very little
about compliance, so I'll base my conclusions here only on what I know.
With new management only just in place at Oracle, I did not expect to
see many
significant changes in Oracle's approach to its portfolio game, and so it was here.
Power Point slides were excellent, but as solutions, the offerings were not compelling.
There were a few signs of improvement over, say, last year, but if Oracle is
going to play this game at the highest levels, it must do better.
The very rough, subjective measure I use to evaluate a portfolio product is what
I call its "time to value." The time to value is the interval between the time a
product is announced and the time when it has actually returned enough value to
customers that it has justified the investment to all concerned. The time to
value is always long--5 years is an excellent time to value--and very often infinite.
I frankly get excited any time I think
the time to value is finite, that is, that the product might actually
take hold in the marketplace and be used by a significant number of people.
In one segment, CPG, I don't think that Oracle's offerings yet
have a finite time to value. But with some changes, the company could be
close to delivering a set of solutions that matter.
In another segment, Contract Management, I was
less sanguine.
Oracle CPG
Most readers will probably not remember that Oracle CPG was the
name of an extremely ambitious portfolio solution that Oracle announced
with great fanfare in 1997 and lost their shirts on in 1998, 1999, etc.
At that time, Oracle was trying to fight the perception that SAP
"owned" the CPG industry; Oracle CPG seemed to be a
quick way of backing into a competitive position using partners.
We now know that Oracle gave SAP way too
much credit. Yes, companies like Procter & Gamble, Unilever, and Warner-Lambert
had committed to SAP. But even today, penetration is by no means complete
at any of those companies (and at the former Warner-Lamber, it is being de-installed).
It turns out that putting the CPG industry on any platform was a
far more intractable problem than the software companies thought.
So, even today, the industry is a good potential market for an applications
company that is willing to address its problems.
Even today, the competition is beatable.
SAP, for instance, doesn't have the forecasting, warehousing, transportation,
process manufacturing, or CRM tools that American CPG companies need.
It makes sense, therefore, for Oracle to try to make an impact in this industry. And there have been sales, as the badges at the CPG sessions attested.
One reason for the sales, perhaps, is that Oracle has been assembling a team of old CPG hands. Phil Friedman, the major
spokesman during the day, has been inside the doors of many a CPG company, and
Joe Kozak, the most senior executive that I talked to, I had known in a previous
life, when he was heading up the CPG effort for SAP. These are both
people who can open the doors and can talk the talk.
Still, the effort seemed to me to be slipping on the ice. Take
one example, a new effort of Oracle's to develop a promotions
and allowances package. The idea of such a package is to provide
some automation and control for the process whereby CPG companies
pay retailers for prominent product placement, for putting
the product on sale, or for featuring the product in advertising.
The amounts of money paid are staggering, though hard numbers are difficult
to pin down. And the control is very poor.
I've actually discussed what automation might
do in this area with
senior people
at several CPG companies. On the whole, they've been skeptical. A whole lot
of the profit for retailers and manufacturers rides on their decisions about
who gets what promotion money, so you tend to get a lot of people
who are trying to work the system. To get real control, therefore,
you have to make people accountable just when they need flexibility.
You have to get salespeople to report exactly what they are doing; you have to force
yourself not to discount at the end of the quarter, when pressure to
make the numbers is greatest; you have to figure out how to insist that your customer live up
to promises on his end. Not easy.
When the problem is clearly this difficult and this tied up with
the fabric of the organization, I'm not sure that senior
executives believe the software syllogism any more. Software syllogism? It's...
-
This is a huge problem that cost you this much money.
- Our software addresses the problem.
- Buy our software, and you'll
save a lot of money.
The problem with the software syllogism is that software may be necessary in order
to save money, but it is rarely sufficient. By now, in areas which have
been resistant to software solutions for literally decades, the business people
want to understand what else is required for the money to be saved. And when they
themselves have experienced the fact that the relevant processes are very
difficult indeed to control, they are particularly resistant to bland assurances
about software capabilities.
They want the solution, of course, so they're not entirely unwilling to
listen. But they want a lot of substances when you're working through that second step.
Phil knows his business and understands the problem, no question. But
his presentation did not have the substance that would have made
me comfortable about recommending Oracle.
I want more real leadership from a solutions vendor these days than was provided in his message,
a message that was pretty much just the software syllogism
and was pretty much the same as messages that I've heard from every other vendor
over the years.
There was one other problem: this wasn't the right audience. Somebody asked
at one point, how many people in the audience were IT people. The entire audience raised their hands. But solving the promotions and allowances problem will never be an IT-driven project.
Another part of Oracle's CPG offering is new capability in RFID.
Since this is a new problem and a very technical one, I expected Oracle to teach me something, and indeed they did. Oracle's approach is
to start by building RFID communication natively into the application server. This
is an interesting idea.
And when they presented, it was clear that their grasp of the state of the art overall was very good. But we all came away thinking that it is still very early.
The bubble also went a little flat when a person from one of the
Wal-Mart Chosen, a company that is deeply committed to the 11i e-business suite, asked with real interest when Oracle is going to address
the Canadian standards as well as the American ones. And Oracle said, "Well, right
now, we're just dealing with the American."
Oracle CPG is not a disaster area by any means. It's a work in progress. Right now, I'm not seeing that acceleration of time to value that makes me think a product is really exciting. The promotions and allowances capabilites don't seem to have much traction, and the RFID are too early.
Still, it's a break from the past. Those few customers who remember the old CPG will think this is different and will be somewhat more prepared to believe. Still, they're entitled to wonder, "Will Oracle really have the follow-through required to make this work?"
Sourcing and Contract Management
I picked contract management because it's an
area where I've talked to customers and also talked to the
product managers at several competing companies. In this area, unfortunately, I saw little change from what has gone on in the past.
As usual, Oracle's product manager gave good Power
Point. Contract management, he said, is a "really hard problem" that requires a very deep, rich
product with capabilities in many areas. He laid out the areas
clearly. But when it came to the crucial problem--the design
decisions that have to be made when a product has too much to do--it was a flub. He simply promised
everything for everybody.
This
is no more possible than a World Trade Center site that makes
everybody happy. Whatever Oracle actually delivers, it won't be what a lot
of people are expecting. At some point, you need
to be realistic.
This contract management presentation seemed to me a relic of the era
when realism in software sales was considered to be ridiculous.
People
bought based on how much your Power Points promised, not on how much
you actually delivered. To compete, you needed to become the market leader based
on your Power Points, then figure out what the market would put up with
once their money was in your hand.
Maybe I'm naive, but I don't think that this strategy works nearly
as well as it used to. Certainly at Oracle it has been problematic
for
a number of years. There was e-procurement. E-marketplaces.
Online travel. Maintenance. Warehousing. All
had superb Power Point presentations (I learned a lot from them), but
even today, their time to value is indefinite.
There may
be markets where the strategy still makes sense.
But before it is the default, I think you need
a rigorous market analysis that justifies rapid
development of a shell product, a marketing crush, and then rapid
withdrawal if the market doesn't prove out. In contract
management, an area I know reasonably well, it would
astonish me if this were true.
What's wrong with promising everything to everybody? Well,
I think it can extend your time to value out toward infinity.
You end up taking a long time to build a product that's too thin
to provide any individual customer with much value. And when the
customers don't come or don't like it, you give up.
What Does It All Mean?
I said in my last piece that Oracle seemed to be very slowly losing ground to the other major application companies. This piece should be read as a counterpoint to that statement, showing how complicated the problem really is and how difficult it is to make a judgement either way.
The fact is that if Oracle can develop an effective portfolio of products, sell it efficiently into the installed base, and use the successes to sell more suites faster into adjacent markets, it will not decline at all.
And unless we watch these products and gauge their success, we will miss it. We will look at the rough numbers, the sums of all the geographies and industries, and modules, and we will see an upturn, but we won't know why.
In the few areas where I looked, I did not see this happening. In CPG, there was promise, but not a business solution that was compelling. In contracts, there was the plan for a product that seemed too broad and too thin to return value to its customers any time soon.
If Oracle has apparently not yet cracked the code on how to be an effective portfolio company, remember that nobody else has either. Rick Bergquist at PeopleSoft has long argued that it is simply not possible for a broad-based application company to build deep portfolio products. The evidence after Apps World suggests that he might be right.
For other recent Short Takes, see www.b2banalysts.com/new/research/shorttakes/recent.html.
| 
