Management Perception of System Administration

Fri Feb 22 11:38:53 PST 2002

Interesting...I did not attend last night, but Paul and I have known each
other for, hmm...pushing 10 years.  We have worked together at a couple of
different companies. Sounds like maybe I should have attended.

Anyways, there were a couple of points made below that I thought were worth
addressing.  First, contradiction between operational management and
strategic corporate management is not a function of the dot.com bubble, but
simply the state of being for many companies, and in fact a constant tension
at even the best of companies.  IT managers are contantly asked to provide
the best service for the least money, just as retail companies or
manufacturers are asked to provide the best sales services or highest
quality products for the least amount of money.  In this game, IT managers
are asked to perform continuous risk analysis, trading risk of software or
hardware failure for reducing capital and expenses.  In the early life of
many companies, this really tends towards high risk, low capital decisions.
As the company grows, you would think that the trend would be to shift that
equation towards lowering risk, but in disfunctional management teams, that
is not the decision made, and you end up with all sorts of problems.

Note that disfunctional managment is not a problem that can be easily
quantified and analyzed.  Sometimes the IT management does not do the job of
pushing information up, sometimes strategic managment does not do a good job
of asking for/requiring/listening to the new data.  Many dot.com's could in
many ways just be described as overly aggressive.  In market's like
WebVan's, market share is king, and rapid expansion is one way to get it
fast to keep it from others.  Of course, if you cannot deliver on the
promise, you loose the market share just as fast, if not faster.  I can
easily see how a company bent on rapid expansion would ignore the warnings
of IT and other departments that the infrastructure is not reliable enough,
in favor of spending the requested capital on new marketing campaigns,
regional launches, etc.  I have in fact seen it on multiple occasions in the
past 10 years.

Note that one problem we in IT face is that strategic managment does not
understand ourworld as well as they understand others.  Paul has always
advocated that people moving from our world into the that strategic
management world is one of the best ways to fix that problem, and I have to
say that I agree.  The issue you mention of people sidestepping issues to
avoid blame is a problem as well, but that is a product of bad management,
and only addresses one set of problems that any corporation faces these
days.  Get rid of the managers who foster a culture of blame and punishment,
and replace them with managers who foster a culture of responsibility and
reward, and you can turn that around.

On your last point, your right.  Engineers, systems administrators are all
labor.  The ways management planning is done these days is you figure out
the amount of work, and the amount of resources to the work, and make sure
they match up.  Resources (manpower, equipment, services) cost money, so you
try to minimize them.  Work, in theory, produces profit, so you try to
maximize it. Managers who don't know what they are doing will tend to view
their staff as strictly manhours, which can be effective in some situations,
but generally produces unhappy staff members.   Good ones understand that
their staff is more then just the sum of the hours they worked that week,
and their employees tend to be much happier (and as a result, are better
workers). This too is endemic of all companies, not just ours.

The dot.com's are interesting in that they accelerate everything...fast
growth, fast profits, fast management...and sometimes fast failures.  It is
a rollwer coaster ride.  Find a seat with a seat belt.

bigmac

> -----Original Message-----
> From: owner-baylisa at baylisa.org [mailto:owner-baylisa at baylisa.org]On
> Behalf Of David Dull
> Sent: Friday, February 22, 2002 9:30 AM
> To: baylisa at baylisa.org
> Subject: Management Perception of System Administration
>
>
> Last night Paul Evans spoke at the Bay LISA meeting on lessons
> learned at Webvan during the period of 1999-2000.  Actually, he
> presented a few insights and then opened the floor for public
> discussion.  In his well-thought-out presentation he drew parallels
> between the public political and business climate and that of the
> individual companies participating in the dot-com bubble, and then
> he got down to the specifics of incidents at Webvan itself.  Although
> some of the observations were contradictory, anyone who was reading
> business magazines at the time would remember that the times
> themselves were contradictory.
>
> What came out for me was the disconnect between what Paul called the
> "operational management," or the managers who were responsible for
> the day-to-day business, and the "strategic management," or the
> president and the board of directors.  It appeared that the
> operational management was unable to communicate to the strategic
> management the common business sense that they had acquired through
> their prior experience.  Is this true, or is it simplification?
>
> I tend to think that the missteps along the way were shared by all,
> that the miscommunication occurred on both sides of the arbitrary
> fence.  We heard an example of non-redundant production equipment in
> a mission-critical application failing, without spares on hand.
> Many of the system administrators in the audience saw this as a red
> flag, while Paul seemed to think that this was a necessary evil due
> to the budgetary constraints of the earlier, less extravagant,
> incarnation of Webvan.  Paul also seemed to blame new, more ambitious
> management for ignoring this warning and deciding to roll out its
> services in numerous markets before the concept was proven in the
> first market.  However, a contractor in the audience indicated that
> Paul's team had done exactly the same thing by ignoring the
> implications of an earlier failure on the same mission-critical
> application.
>
> I don't think there's a disconnect between operational management and
> strategic management.  I've seen the scramble for safety and the
> fear of being blamed in the rank-and-file, as well as in the
> boardroom.  I have observed how managers, both of operations and of
> research and development, have side-stepped critical issues because
> they did not want their names associated with an obvious problem.  I
> have seen how corporate politics works at all levels to subvert
> corporate profitability.  I once thought this was a phenomenon
> reserved for old companies, but the Webvan presentation showed me
> that it can happen anywhere, that it is more a phenomenon of society
> than of one company or another.
>
> Paul made a very strong point that the only real way for system
> administrators to apply their personal experience to board-level
> decisions is to climb into the board and participate.  He mentioned
> the suit phenomenon, where employees can be promoted into management
> because of their appearance.  I'd like to keep in mind that I have
> only one chance to make a first impression, and that the suit, or at
> least a business-casual appearance, must be maintained continuously.
>
> I would like to point out, however, that it is often corporate
> expectations that provide the social pressure for technicians and
> engineers to "dress down."  Managers don't want to be confused into
> thinking that their line employees are also managers.  That doesn't
> fit into their mindset.  Yes, in their minds there is a class
> distinction between "management" and "labor," and while engineers,
> programmers, and system administrators are not typically unionized,
> managers tend to think of them as labor.  System administrators will
> not find changing management perception an overnight project.
> Engineers have been trying to change this perception for a hundred
> years, and as far as I can tell they have made no progress at all.
>
> --David R. Dull
>   ddull at ieee.org
>   http://home.netcom.com/~qkstart
>