The distribution chain analogy is extremely helpful
and I think we can all see measures at different
positions in the chain.

The wholesaler will evaluate the cost benefit balance
when considering which fridge trucks to buy because it
will directly impact his ability to deliver to his
customers, many of which will have contractual service
level agreements.

The supermarket will usually have clauses that state
"less than 1% of deliveries to be innaccurate or
incomplete" and will measure that.  Web users measure
sites in the same way - for news sites people judge
based on the timeliness and quality of reports on
issues that matter to them - their implied SLA.  For
product sites the consumer is looking for something
and judges based on their ability to find it.  They
then make a buying decision based on price, brand,
trust etc.

It is the wholesaler's, or business manager's, job
within the wholesale or content-delivery link in the
chain to specify what is needed in order for them to
perform the job as best they can.  These tests will
usually  not revolve around functionality, a fridge
truck doesn't vary that much!  What varies are the
non-functionals or quality metrics, the meantime
between failure, the availability of the systems, the
frequency and scope of loss of data accidents etc.

While I agree that the spoilt lettuces may well be a
human error issue it could also be an issue with the
refigeration unit on the truck.  The balance between
human failure and machine failure will be affected by
the frequency of machine failures.  Hence, buy a
better truck and you should get fewer issues.  But
when I say better you have to take _all_ quality
requirements into account - the MTBF, the ease-of-use
of the unit, the ease-of-maintenance and so on.  If
the system is mis-configured is that human error or an
issue with the quality of the configuration tools and
interface?

What concerns me about those comments that say that
CMS implementations are difficult or impossible to
measure is that business leaders will not continue to
buy that.  When the money was a VC's and times were
flush, maybe, but not now.

Perhaps the lack of intelligent, objective measures is
the reason that _most_ cms implementations are
absurdly expensive and deliver little or no perceived
benefit once implemented?

For those working in an XP or Agile development arena
this is the test.  If you're business user can't
define the test that the system must pass then you are
at liberty to say "it's done" and take the money.  A
trend we seem to see a lot of in the CMS marketplace.

To say that a product has no directly and objectively 
measurable benefits is to say "don't buy this".

cheers

mmmmmRob

__________________________________________________
Do you Yahoo!?
Yahoo! Web Hosting - Let the expert host your site
http://webhosting.yahoo.com
--
http://cms-list.org/
trim your replies for good karma.

Reply via email to