In 1996, the College of Mount St. Joseph began migrating its network
from Novell NetWare to Windows NT Server. To prepare my network
administrator for the task, I sent him to take the initial sequence of NT
classes, at a cost of around $5,000. Three weeks after completing the
third course, he tendered his resignation. With a sigh, I sent his
replacement through the same training sequence. He lasted a bit over a
year, then also resigned. A third individual went through the training,
quitting after six months.
The reasons for leaving were varied. Number one left for a large salary
increase. Number two left due to the pressure of the position’s
responsibilities. Number three went to a Microsoft Solution Provider for
"the professional challenge." Whatever the reason, after hiring
and training three network administrators only to have them leave before
the training investment paid off, I decided we needed to find a better
way. Stopping the training was out of the question; as one of our
colleagues has said, "The only thing worse than training them and
having them leave is not training them and having them stay."
"Let’s just outsource the whole network," I told my staff.
It was a naïve response, born of frustration and desperation, because it
didn’t consider all the ramifications of such an action. Yet there was
the germ of a good idea there.
Reasons
There are persuasive reasons for outsourcing network support. Everyone
has a network; everyone needs network support. As a result, network
professionals are in short supply, and competition for them is fierce. The
bidding war that results from this competitive atmosphere may lead to high
turnover, as we raid each other’s technical crews. For example, right
now, individuals in my college’s technical support average slightly
under 20 months of service in their current positions.
Experienced network professionals are highly paid, more highly paid
than a college or university may be able to afford. An alternative is to
take a good technician and train him or her for the position. This can
also be expensive. Besides the class costs, newly trained administrators
are only minimally competent. It takes months for them to effectively
begin applying their training. This not-insignificant ramp-up time coupled
with the high turnover further reduces the productive months an
organization can expect from these high-in-demand professionals.
Ironically, the training you pay for makes the technician more
marketable. It’s a vicious circle. You can try to hold onto them by
making them agree to stay a set amount of time after training—say two
years —or to repay a prorated amount based upon the number of months of
their obligation that was not met. I’ve tried this too, but these
agreements are difficult to enforce. Besides, even if you are successful
with this, the last thing you want is a disgruntled employee, forced to
remain at the institution, who has power over your entire network.
There is another incentive for outsourcing that should not be
overlooked. A certain amount of angst is associated with the support of
the network, pressure caused by balancing the demands of day-to-day
activities, the emergency service calls, and the special projects. When
you outsource, you get rid of all of this. Many problems—recruiting
talent, training, supervising, and retaining them—become those of the
outside firm. The stress of network support and completing special
projects goes away as well. You are the client, not the supervisor, of the
employee.
Sounds wonderful, doesn’t it? But does it work? Well, yes and no.
First, let’s consider the options.
Options
You can outsource day-to-day service. This would include creating and
removing accounts; changing network privileges; making backups; restoring
lost data; installing and updating software; etc.
You can outsource emergency service. Networks have to run 24/7. Banker’s
hours won’t suffice when servers crash, wires get severed, and hubs die.
You can outsource special projects. Networks have an organic quality.
They grow, they evolve, they develop new capabilities. This evolution
results from the addition of new hardware or software or the
reconfiguration of the network architecture. The special work is more
expensive on an hourly basis than paying your own staff. Outsourcers will
tell you it really isn’t; they cover the costs of benefits, recruiting
and training staff, and supervising them. However, even accounting for
those elements, you pay a premium for contract labor.
With an outside firm, it’s more difficult to predict and contain
costs. If you pay consultants on a time-and-materials basis, they work
until the job is done. If it takes longer than estimated, too bad, and a
$5,000 project can easily become a $10,000 headache or a $15,000 disaster.
Using outsourced staff to handle special network projects can remove
stress from your own technical crew, but it can cause other organization
problems. Your own employees want to grow professionally, but if all the
challenging work is handled by the outsourcer, they may not get the
opportunity. College technicians may resent the "good projects"
going to consultants.
As with most consultants, outsourcers have little institutional
loyalty. They aren’t really your employees; they work for somebody else.
In addition, while they may know the business world, they usually don’t
understand higher education, and this lack of understanding can cause
problems. For example, the outsourcer may recommend a proxy server to
provide firewall protection for the network from potential Internet
hackers. All well and good, but as everyone knows, security is typically
at odds with functionality. In order to make the network
"safer," the consultant may block certain types of traffic, such
as streaming audio or ftp. That may be fine in a corporate setting, but in
an academic environment, where exploratory inquiry is the name of the
game, this lost functionality can adversely affect instruction and
scholarship.
There’s one other problem dealing with outsourcers, something I
describe as the "hot dog" syndrome. Many technical staff get so
focused on the technical challenge of a job that they sometimes lose sight
of the organizational effect their tinkerings can have. Because
outsourcers do not know the institution as well as a college staff member,
they might be less sensitive to the ramifications of their actions. In
addition, they are not your employees, so you have less control over them.
Some advice
Unless you have a whole lot more money available for consultants than
we do, you will probably not be able to afford to outsource all network
services and support. What we have finally settled on is using outsourcers
for special projects.
This works wonderfully well. The stress of trying to do a special
project on top of day-to-day work is effectively exorcised from the
organization. Sending your own staff through a steep learning curve is no
longer necessary, because even a mid-sized computer consulting firm has a
host of technical people with different specialties. Because of this
breadth and depth in knowledge and skills, they can implement special
projects much more quickly, and then train your staff to maintain the new
system.
Some advice to help ensure a successful outsourcing experience:
Carefully define your project. It is important to identify all the
deliverables you expect. What is your minimum acceptable product? What is
in the scope? What is out of scope? What is your deadline?
Try to cap the price of a project to control costs. There are two
ways a consultant might bid a project. One is to offer a set price; the
other is to work on a time-and-materials basis. Because of the uncertain
nature of technology projects and their tendency to expand beyond the
initial scope of the project definition, it may be difficult to get a
consultant to offer a set price. If it is necessary to contract a job on a
time-and-materials basis, try to negotiate that it will not exceed a
certain price.
Make certain the consultant performs a knowledge transfer after the job
is done. This should include training your staff and documenting the
work.
Get a guarantee. These technologies are inherently unstable, so
after the implementation occurs and the company says its job is done, have
some agreed-upon "watch" period, during which if it crashes,
they will come back and fix it for free.
Consider developing an ongoing relationship. Switching consultants
on a regular basis may be a way to keep costs down on a project-by-project
basis, but it may be penny-wise and pound-foolish. A sustained partnership
with one or two vendors builds a sense of mutual commitment that can pay
dividends over the years. But ...
... beware of locking yourself in. You need to check out the
competition periodically. Your long-term consulting partner may begin to
take you for granted, the price may go up and the service may go down. It
pays to occasionally put projects out to bid just to make certain you are
getting your money’s worth.