Mike Gort wrote:
(snip)
>... if the business user works in a very transactional culture, that
>get-it-done-today is likely to collide with the more reasoned, project
>management oriented, culture of a well run IT group. How often have you
>seen development managers deliver a product that is 80% finished, poorly
>tested and poorly documented, because they wanted to delight the customer
>who is pushing hard for a system needed in his or her transactional
>world.
One of HP's greatest strengths is that it is composed of about 100
business units. Each of these business units (sometimes called
divisions or "entities") is headed by a general manager. That GM has
only one job: to make that business unit a success, as defined by a
negotiated agreement with the GM's manager. Likewise, within the
context of HP's value system ("The HP Way"), the GM can do anything
he or she needs to do to achieve that success.
If the business unit does not succeed, it is shut down, and its
products, people, and resources are "recycled" to existing or new
business units. This evolutionary cycle happens often. In some
years, there is 20% turnover in business unit structure.
Innovation comes from these relatively small, focused organizations.
Experiments and risky ventures can be implemented because we know
that we can easily limit the costs of these ventures if they don't
work out.
This structure allows HP to be entrepreneurial and compete with small
startup companies. And business units, when they feel it is in their
best interests, can voluntarily join together with other HP and
non-HP partners to create economies of scale to compete with large
players like IBM, Sun, Compaq, Fujitsu, etc.
In order to be held so completely accountable for their success,
business unit GMs also have essentially complete freedom on what they
spend their money on. With the exception of a few common processes,
the GMs don't have to buy from other HP sources unless they want to.
In particular, business units have (and do exercise) the freedom to
not-buy IT services from HP unless they want to. In IT, the major
exceptions are the communications wide area network and network
security, and even in these, there are some notable independent
actions. To give you an idea of scale, there are about 6000 IT
people in HP. About 4000 of them work on tasks directly controlled
by the business units, while about 2000 of them work on enterprise
wide/infrastructure processes.
Business units usually treat IT investments the same way they treat
investments in other functions such as R&D and marketing. In
deciding on what and where to buy, the criteria usually focus on
"quality attributes" (i.e., functionality, usability, reliability,
performance, and supportability) and then on cost, although sometimes
cost is a primary driver.
This explanation is an extreme simplification, but I hope it gives
you the general idea.
>In IT shops that practice very close developer / client relationships,
>the result can be very expensive. (snip)
At least in our culture, the results are whatever the business units
want the results to be. Most often, they want some sort of enabler
to innovation or time-to-market, even if costs are higher. Costs
become an issue only in repetitive, long-run processes, and more and
more of our business units are building products with very short life
cycles. Cost is, of course, a major issue for enterprise-wide
processes, but there are relatively few of those in HP.
(snip)
>Further, turn-over is likely to increase as developers remain
>responsible for supporting more and more legacy (in some cases, we
>estimated as much as 80% of a developer was actually spent in customer
>support).
It's another subject, but these days, much of our legacy support is
going into Year 2000 efforts.
>My point is that there is another customer, one that is too often ignored
>by everyone in the organization - the Stockholder. The CEO needs to become
>a proxy for the stockholder and force a more rational decision on what
>quality constitutes "good enough".
(snip)
>Building credibility from perceived trust and perceived technical expertise
>is spot on. However, the dynamics of the IT / Business User / Shareholder
>or CEO dynamic is so strong, they may overwhelm the issues of credibility.
>Even after enhancing credibility by following the many sound pointers in
>the jotting and the comments, the external feedback from increasing
>competition for limited resources may set back the credibility of IT once
>again. It is common to hear the business manager complain about how
>expensive IT is, even though their pressure for fast delivery, and the
>impact on costs, is a significant part of the system driving higher costs.
As I indicated above, HP-IT, except for the few enterprise-wide systems,
tries to help the businesses, who, in turn, try to succeed, and, in turn,
their success rewards the stockholder. But, if we just do what the business
managers want us to do, we are adding little unique value, and the business
units should (and often do) get their IT services from non-HP vendors with
cost structures lower than we have inside HP. The goal, therefore, is for
us IT people to use credibility methods to give us the opportunities to give
business unit managers the benefits of our professional expertise, so that
we are not just a "pair of hands" that can and should be outsourced.
Joe
joe_podolsky@hp.com
--JOE_PODOLSKY@HP-PaloAlto-om4.om.hp.com
Learning-org -- An Internet Dialog on Learning Organizations For info: <rkarash@karash.com> -or- <http://world.std.com/~lo/>