Dancing with Bears
On Wednesday, May 8, i2 will hold e-Day 2001, where it will show
off its new order management software to financial analysts and will
announce new strategic relationships with Procter & Gamble and
VF. Experienced software people are likely to discount the order management
piece because i2 has little experience in that area, but there I think
they underestimate i2. Financial people are likely to take heart from
the existence of big customers, and there I think they overestimate
i2. i2's recent experience with Kmart suggests that there are perils
in these relationships which i2 has yet to learn how to deal with.
In August and October 2000, at the last two big i2 events, i2 announced
a new strategy: to form partnerships with large companies whose aim
was to transform those companies' infrastructure. Among the companies
was Kmart, whose CEO, Chuck Conaway, announced at i2 Planet (the user
conference) that the partnership would enable Kmart to "leapfrog
It wouldn't be too bad for i2 and its consulting partner, Deloitte
& Touche, either. Conaway announced that $50 million would be
put into the project.
In the last couple of months, Kmart changed direction.
Conaway originally announced that i2 and Kmart developers would work
together onsite at Kmart on 12 (later 24) different supply chain products,
the first of which would be released in the first quarter of 2001.
Now, the i2 (and Deloitte) team devoted to those projects has gone
home. The Kmart initiative, Play to Win, has been refocused on operational
improvements and pragmatic process changes that have nothing to do
with the joint i2/Kmart projects that Conaway described.
Both i2 and Kmart continue to have a relationship, because Kmart is
a heavy user of Inforem, an IBM product that i2 bought several years
ago, and both i2 and Kmart have indicated that they are happy with
that relationship. Deloitte still has some people at Kmart, including
Randy Allen, the Kmart CIO, but the total has been decimated.
What It Means
We should expect these things. Software is just a tool, and every
software company has sold the tool to a company that is unable to
use the tool effectively. When the company is big and expects to have
things its own way and when it is having lots of other operational
issues, the chances of success go down even more.
Brian Sommer, one of the gurus of the industry, puts it more picturesquely.
Alliances of large companies, he says, have a limited life. "When
a large company forms an alliance with many smaller companies, it
works pretty well. The large company plays the music, and the smaller
companies all dance to it. But when two large companies, each with
their own agendas and their own will, try to form an alliance, it's
like two bears dancing. They make a take a few steps together, but
sooner or later they step on each other's toes, and it falls apart."
That alliances end in divorce may be a fact of life; even so, i2 has
been put in a difficult position. Wall Street is not so attuned to
the nuances of this industry that it can be relied upon to judge fairly.
As the Wall Street Journal article about Kmart and i2 said, "Any
setback would be a black eye for both companies." Bear in mind
that SAP has yet to live down the Hershey's debacle, even though Hershey's
is now able to forecast and deliver product quite effectively. i2
has already had its Nike; it doesn't need a Kmart.
That's why i2 executives have been so eager to downplay the Wall Street
Journal article and to be dismissive of later questions about the
Kmart relationship. Rather than laying blame on one party or the other,
and rather than punishing i2 harshly for something that is to be expected,
we should remember the following truths:
- Some strategic partnerships will fail. Whatever i2's
capabilities, some organizations are ill-equipped to take advantage
- Scaling up is a non-linear process. Software that can
handle 20,000 SKUs very effectively may have to be rewritten before
it can handle 150 million SKUs. Processes that are supported well
in some industries (e.g., replenishment planning in manufacturing)
may have to be rethought when you try to support them in more
complex industries (like retail).
- New technology projects take longer than you think. A
marriage of software expertise and industry expertise does not
necessarily produce a serviceable product on a predictable schedule.
i2 has long used industry partnerships as a way of developing
software. But the software produced by these marriages often has
limitations that must be worked out. It is often the case that
pilots produce questions or issues that were not thought of originally.
And it is often true that the software is cast in the mold of
the founding company and must be revised when it moves out to
- Part of software installation is change management, and change
management for unproven software is much the hardest kind.
Given these fairly obvious truths, it is hardly appropriate to
call a change of commitment by one of the parties, "a black
A longer version of this report is available to our clients
For other Short Takes, see our archive.