Re: Are User Surveys the best approach for defining your audienc

[John McDermott <jjm -at- JKINTL -dot- COM>]

Doug, great answer to Kelly who wrote:

>**********************
>     My department has decided to "better define our end-user".  We are
>     considering taking the survey approach. Can anyone make suggestions as
>     to some good questions to ask on a user survey? And how can we get
>     them to send it back? We've got to start somewhere, but is a survey
>     the best approach?
>**********************

>My off-the-cuff response to that last question is "no."  IMHO, the best
>approach to "knowing thy user" is to walk a mile in his or her shoes (or
>boots, as the case may be).  I do not mean "observe," I do not mean "hang
>around with," I do not mean "conduct a personal interview."  I mean get out
>there and BE A USER.  Bend tin, enter data, tune control systems, whatever
>it is that your user does, you should do to.  There is no substitute for
>this experience.

[[lots of good suggestions deleted]]

>If you can't get out among your users, at least spend some time working on
>the support line.  Find out what frustrates people and what you have to do
>to talk around it.  It's a substitute for a site visit the way TVP is a
>substitute for beef, but it's better than sitting back and making a guess,
>or relying on second-hand reports.  Remember, a bad day doing is better
>than an good day theorizing.

>Anyway, once you and some other people in your department have this
>experience, you will have some context to ask questions about the things
>you didn't observe, or to discover how widespread some of the behavior you
>observed is.  Until you have this context, your survey is pretty much a
>blind man in a dark room searching for a black cat that isn't there.

This is a fantastic suggestion.  One major computing facility I know used to
(maybe still does) make its senior personnel man the help desk phones for a
few hours a month just to keep up with the "common folk".  The benefits were
terrific:  the top guys got to know what the *real* problems were and what
frustrated the users.

The cost of sitting in at the trouble desk or with an engineer should be
MUCH less than designing a survey and the results should be much, much more
concrete.  Survey results tend to fit into the "lies, damn lies and
statistics" category...

Good luck and let us know what happens.
--john
>Skoal,

>Doug                            "Did you really think that you could
>ENGSTROMDD -at- phibred -dot- com           conjure up the Devil, and then expect
>                                 him to behave?"
>                                                --Fox Mulder

>***********************************************************************
>The preceding opinions and positions are mine alone, and are only
>coincidentally related to those of Pioneer Hi-Bred International, Inc.
>***********************************************************************

-------------------------------------
Name: John McDermott
E-mail: jjm -at- jkintl -dot- com (John McDermott)
Voice: +1 505/377-6293  FAX: +1 505/377-6313