[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Heresy? (was Re: User demands)
While everyone thus far has touted the benefits of a ticket system, let me cast
a lone dissenting voice with our *specific* situation. While ticket systems done
properly are IMHO the way to go, practical considerations have forced me to
advise my user community to come to me (and my fellow sysadmins) directly. The
user community is extremely pleased with our work (e.g., our boss came for
feedback and we got spontaneous applause). Won't work for everyone, but here's
the background.
Here is where outsourcing rears its ugly head.
In 1994, the company formerly known as Hughes Aircraft outsourced a major chunk
of its IT infrastructure to Computer Sciences in an 8 year sole source pact.
Part of this deal involved setting up a trouble ticket call center
infrastructure. The other major part you need to know is that the Network
infrastructure and computer hardware maintenance was handed off, while the
SysAdmin labor was mostly held back (i.e., not outsourced).
Fast Forward to 1999. I'm one of the sysadmins. A very current discussion in our
group is whether we wish to continue using the Help Desk or establish an
alternative of our own.
Problems with our Help Desk:
1. By it's very nature, there is an overhead with monitoring and updating
trouble tickets. In this case, we're dealing with a mainframe based system with
arcane codes, with little collected data and trends value.
2. To get hardware fixed, the ticket needs to be routed to the outsourcer's
agent du jour. We need to be very careful when we open the ticket and say for
example: "Printer X is broken. Please refer this ticket to Josie who will then
call Grumman to have the ticket fixed". Otherwise, it winds up back in our
Engineering queue, i.e., referred back to me to fix it when I'm the guy
initiating the call.
3. Help desk personnel turnover is sky high. The outsourcer basically hires the
cheapest people it can find to meet the main job requirement, i.e., answer a
phone, transcribe the input, and figure out what group should handle it based on
keywords in the conversation.
4. It's very PC centric. Say Unix and they might go into a fog.
5. There is inevitable data loss when an engineer calls to describe the problem.
E.g.:
Engineer: I can't seem to print correctly on the m230 plotter from Cadence.
HelpDesk: What's the bar code on the plotter?
Engineer: Bar Code? I don't know. What concerns me is that the layers on my
design are coming out in different colors when I know they're all the same.
Furthermore, I need to be able to see all the layers in my drawing.
HelpDesk: Umm.. OK, let me try and write that down.
Resulting ticket description: Engineer wants drawing to have one color.
Real Fix: HPGL2 Merge needs to be on so that topmost layer in a schematic does
not opaque blind the overlapping metal layers and all layers are visible.
In the case of problem #5, the Help Desk is negative value added.
Our solution ("come directly to me") probably doesn't scale beyond the 80+
people in the engineering workgroup, but the customer satisfaction (and
productivity) is very high. My productivity is enhanced too since cutting out
the middleman and getting the problem description directly let's me deal with
the issue in 1/4th the time. And I don't have to waste my time updating the
calls.
What about the downside?
Yes, we have one very demanding guy who thinks we are at his beck and call. He's
toned down since we raised the red flag with his boss - and his winning
personality is well known. Yes, we have days where the flood occurs and we have
to tell people that their request needs to wait. That's independent of the
method used to resolve the flood. If anythingg, our method get's us out of the
flood faster. Overall, the user satisfaction is quite high as they mostly see
their requests for help resolved in minutes.
As I said, we're evealuating what to do with the existing help center, and yes,
wreq is a candidate replacement. When evaluating in an Excel spreadsheet, it's
hard to translate the value of me having the same interest (i.e., same paycheck
source) as the engineer, versus a Help Desk person who vaguely cares about SLAs
that are easily met with surveys and statistical tricks.
Mario Obejas
mario@fea.net