Computers — A Tragedy of Errors: Part Three
| Published date | 01 July 1985 |
| Date | 01 July 1985 |
| DOI | https://doi.org/10.1108/eb057405 |
| Pages | 3-5 |
| Author | William Jacot |
| Subject Matter | Economics,Information & knowledge management,Management science & operations |
Computers
A Tragedy
of Errors:
Part Three
by William Jacot
Managing Director
DBA Computer Consultants
Half of all computer systems must be assessed as failures.
Failures in that they fail to achieve the business objectives
set for them in the time scale planned. Many of these failures
are caused by lack of a disciplined approach to the selec-
tion of hardware and software by the user, often compound-
ed by a failure of the suppliers to accept and demonstrate
their legal responsibilities. However, many carefully chosen
systems bought from responsible suppliers still fail or, to a
lesser extent, have serious problems during the implemen-
tation phase.
In the first two articles of this
series[1,
2], we have examin-
ed the prime causes of failure resulting from hardware and
software problems, assessed the reasons for failure and
sug-
gested how such failures might be avoided — how best to
avoid the pitfalls so unhappily encountered by so many first-
time computer users. In this third article, we look at the pro-
blems that arise between signing the contract and "going
live"
on a fully working and tested system. This is often
call-
ed the implementation phase. The definition of requirements
phase and the supplier selection phase were integral to our
first two articles. While this article is intended to be sup-
plementary to the first two, it closes by summarising the
main overall recommendations.
"We were hoist with our own petard," sighed the manag-
ing director of a small manufacturing company in the
Midlands. "We have had the system for 18 months and we
still haven't gone live on the first jobs yet. With the benefit
of hindsight we should never have begun until we were ab-
solutely clear what we wanted to do.
"Some of my colleagues said that they understood, and did,
and some I'm afraid were so embarrassed by their ignorance
of computers as to nod their assent and hope for the best.
Consequently, it was only when the suppliers sent their
detailed specification of what they thought we wanted to
do that we began to see how different it was from our real
objectives."
Though rolling stones may ignore the moss, successful
computer systems depend on a firm base and any attempt
to make major changes to basic specifications will certain-
ly be disastrous to anticipated start dates if not lethal to the
total project. Modifications will at the least be expensive.
"Yes,
we have discussed our problems with the software
house,"
explained the somewhat frustrated solicitor in
charge of computer development within a major South
Lon-
don practice. Its new computer system designed for client
accounts, time recording and general word processing was
already over a year late with little to show for its investment
of £25,000. "The trouble is every time we ask them to come
and explain why things are hot going as expected they want
to charge us another £500 for changing the software quite
regardless as to whether such changes are required in order
to mend the bugs caused by their programming mistakes
or whether we have asked for changes. If we complain about
the response times — now sometimes over 20 seconds —
they offer to supply us with their latest machine — for an
extra £10,000 of course." Suppliers often encourage the ac-
ceptance of their bigger and better machine with apparently
generous trade-in offers for the previous model.
Where there has been a genuine major change of require-
ment which in consequence requires more powerful
hard-
ware,
or software with enhanced facilities, of course such
overtures deserve serious consideration. However, where
such approaches merely reflect a change in sales tactics
designed more to achieve equipment orders than help the
existing system, one cannot but suspect the worst.
The contract is the key. As discussed in our previous
arti-
cle,
a balanced, fair and reasonable contract is vital to the
success of all computer projects. The linking of acceptance
tests payment terms and the original specification provides
the foundation for allocating responsibilities, tasks and time
scales which would have alleviated the lawyer's problems.
"We did everything right as far as the contract and then
forgot all about ourselves," — a West Country distributor
of machine tools had bought a computer system to help
with its stock control procedures, order processing and
book-keeping. As the financial director
said,
they had made
a sensible approach and taken the right decisions and had
achieved a reasonable and fair contract for a system which
stood every chance of meeting the firm's requirements. So
far so good. "We knew some would be for it and some
against it, but we hadn't really appreciated the effect of the
system on our own staff," he explained as we were about
IMDS • JULY/AUGUST • 1985 3
To continue reading
Request your trial