Case Study: No Changes on Friday
One of Tom’s customers did a lot of SA tasks, and Tom was affected by his customer’s
changes and schedule, rather than the other way around. When his customer made
a mistake, Tom was blamed because some of the group’s systems depended on the
customer’s systems. The customer and Tom had the following debate.
Within his own area, Tom instituted the rule that no changes should be made on
Friday, because if mistakes were discovered over the weekend, response times would
be slower, increasing the risk of adversely affecting his customer’s work. Tom also
didn’t want his weekend ruined, and he thought that the customer too should abide
by this guideline.
The customer believed that no matter what day Tom’s group made changes, the
SAs should check their work more thoroughly and then wouldn’t have to be concerned
about making changes the day before a weekend or vacation. Even though
the customer’s own changes often caused problems, he refused to acknowledge that
even when changes are carefully checked before they are made, the unexpected may
happen or an error may occur. He also refused to acknowledge the increased risk to
other customers and that lowering that risk was in everyone’s best interest.
The customer also felt that the SAs should make their changes during the day,
because there’s no better test of a change than having live users try to work on the
system. He felt that if changes were made during off-hours, the SAs wouldn’t find
the problem until people who used the system returned in the morning or after the
weekend. He preferred to make changes during lunch, so that there was a half day of
testing before he went home. There was no lunch hour at that company. The cafeteria
was open from 11:15 to 1:30, and at least a third of their customers were active on
the network at any time.
Both Tom and this customer had valid points, and no one rule will do for all
situations. However, ignoring the risks to other customers is not appropriate, and
saying that more careful checking is sufficient is not a valid point. More careful checking
is always good, but it is not a sufficient reason to ignore a higher risk to the.