Re: $ per word - quantifying TW New Slant

["Dimock, Dick" <red -at- ELSEGUNDOCA -dot- ATTGIS -dot- COM>]

TO TECHWR-L
from  Steve Jong
via Dick Dimock



 ----------
From: Steve_Jong/Lightbridge*LIGHTBRIDGE
To: red
Subject: Re: $ per word - quantifying TW
Date: Thursday, January 18, 1996 3:12PM

Dick,

 [I'm frustrated by my inability to post to TECHWR-L (we changed something
over
the summer and I was cut off mid-sentence).  I still read the digest, and
the
discussion of metrics touches questions near and dear to my heart.  I would
like to offer you some thoughts personally; I ask, please,  that you post
this
message (less its attachment) to the list.]

 I agree that the question of metrics is both elusive and vitally important.

The obvious metrics, like pages per hour or cents per word or readability,
have
obvious flaws, and in some cases are counterintuitive.

 You wrote that you have tried a variety of metrics as a writing manager and
as
an individual contributor, and that none of them was satisfactory.  There is

solution to the problem, but it requires a different conceptual framework to

understand.

 A user information product (a printed document, a help file, a training
aid,
etc.) is a manufactured product.  As such its quality, or its worth, has two

components: its value and utility as seen by the person who uses or
possesses
it, and its value and cost to the agent that produced it.  The two
components
are separable, though not entirely unrelated.  Thus, to take a concrete
example, a document that is timely, accurate, clear, concise, attractive,
friendly, and so forth and so on, can  be said to be of high quality; but
that
same document can be of high quality OR low quality to the agent that
produced
it, depending on a different set of conditions.  Did it satisfy the
customers?
Was it economical to create?  Is it easy to maintain?  Expressed another
way,

given document could fall anywhere in a two-dimensional space along the axes
of
"easy to use" and "easy to make" (assuming for the moment that it's possible
to
measure both these attributes):


Hard to |  X   X
Make |
 |
 |
 |
Easy to | X   X
Make |_________________________________

 Hard to Use  Easy to Use


Viewed as a two-dimensional problem, it becomes easy to see that both
dimensions -- user information as a product, and user information as the end

result of a process -- are important.
Ideally, we would like to produce a document that is both easy to make and
easy
to use; I think in practice one can argue the two factors correlate, such
that
it's hard to make an easy-to-use document -- in short, there are tradeoffs.

It's also apparent that as technical communicators we tend to be interested
in
maximizing the product, while managers naturally tend to be interested in
minimizing the cost.  Should we print this book in full color?  The question
is
not one simply of customer satisfaction; of course they'd like color.  But
neither is it entirely one simply of cost; of course it's more expensive.
Both
elements must be considered to reach a sound decision.

Now, given the two dimensions, it becomes apparent why productivity metrics
such as pages per day (or dollars per word) are attractive, and in fact
*relevant*, to managers.  It also becomes clear why writers don't care for
such
metrics, but are more interested with other factors, such as clarity and use
of
illustration -- and why managers don't care for those.

Which to use?  Obviously, you need a mixture.  You need to measure and track

the costs of producing product, as with any other manufacturing business, to

maximize results and minimize expenses.  At the same time, you need to
measure
those elements of the finished product that result in customer satisfaction
and
efficacy.  (Do they like art?  Give 'em more art!)  So I argue for a suite
of
metrics; none individually amounts to much, but collectively they are
persuasive.

I don't wish to leave you with an empty theoretical lecture, so I will add
that
in my research I have discovered that some metrics are more useful than
others.  For example, an IBM study found that measuring productivity in
terms
of weighted pages per day was meaningful.  A weighted page was a page
(whatever
that is) times the effort to create it, with a page of new documentation
weighted as 1.0, a revision as 0.5, and an update, I think, as 0.25.  (Being

more precise than that is not worth the effort.)  The cost of documentation
is
primarily the cost of labor, so the time to finish is meaningful;
specifically,
time overruns should be tracked and analyzed.  Did the project slip because
of
slow writing?  Or was it slow review?  Or late specs?  3M discovered that
the
three greatest contributors to slippage were no specs, late specs, and
changed
specs.  No surprise there, but having data in your hand is the only way to
effect change.

To create metrics for the finished product, ask your customers what they
like
and don't like, about your products in particular and about all such
products
in general.  Collect statistically significant samples, not just the three
closest customers.  Generate weighted lists of critical success factors
(CSFs)
based on their responses; use them to learn what to measure.  I'm willing to

bet that you'll come up with a very plausible list of important factors.
 The
last time I did it, we ended up with these:

 More examples
 More illustrations
 More headings
 Better index
 Clear
 Concise

Once you collect CSFs, you know what will satisfy your customers; can you
think
up ways to measure your products to see how well they meet these needs?

(Don't take these as your solution.  These were *our* solution, *at that
time.*  You will get your own answer, based on your own customers at the
time
you ask.  But that's good!)

  -- Steve

================================================================
Steven Jong, Documentation Specialist  ("Typo?  What tpyo?")
Lightbridge, Inc, 281 Winter St., Waltham, MA 02154 USA
<jong -at- lightbridge -dot- com>, 617.672.4902 [voice], 617.890.2681 [FAX]
****************************************************************************
***
***********
Great Stuff, Steve!  Yeah!

Foist uv all, TECHWR-L has a new address for subscribed posters:

TECH WR-L -at- LISTSERV -dot- OKSTATE -dot- EDU

The VM stuff dropped out. Eric cut over to that address Saturday.
Try that address. Look for the most recent EJRAY or ERIC post
for subscription directions.

I particularly like your CSFs.  Excellent food for thought!

But I gotta write - deadline tomorrow, vacation after that.
I would like to chat more about this, and I wish I were not
missing the list discussion while I vacate!

Thanks!

Dick Dimock
richard -dot- dimock -at- elsegundoca -dot- attgis -dot- com
red -at- elsegundoca -dot- attgis -dot- com