One of the issues that should be considered in discount usability
engineering is that these discount methods still need some rigor.
Take the discount method of "think-aloud usability testing" (TAUT),
one of the prominent discount methods.  Doing a good TAUT requires
careful attention to the choice of tasks that you use (or if you let
the user define the task, a careful interview to understand what
he/she would like to do).  Choosing tasks and deciding on the order of
tasks is a single line in a checklist, but it requires some thought
about biases of stakeholders, the appropriate granularity of the task,
the degree to which the task reflects the real world in terms of it
complexity, data inputs, etc., whether one tasks will affect the next
task, etc. What are the criteria for choosing tasks?  You can choose
tasks by frequency of use, criticality (rare, but devastating), new
features that marketing will advertise about heavily, and many other
criteria.  The choice of tasks is even an ethical issue in that you
could choose tasks that make the team look good (I've seen this when I
was a lab director and rented out lab out to colleagues who wanted to
please their managers or vice presidents).

Oh, and each task should be tied to both a business goal and user
experience goal which is often not done by those with limited
experience.

Thinking-aloud (which can be used for a variety of methods) in a
usability test can be affected greatly by the verbal and non-verbal
prompting and interview style of the facilitator.  Facilitating a
think-aloud usability test is often portrayed as a simple discount
method, but the reality is that it is a complex interaction that
requires some rigor and constant self-reflection.  Studies at the
University of Washington have shown that the variability in how
facilitators conduct sessions is extreme.

Analyzing the results of think-aloud usability testing seems easy,
just list the problems that you hear and see, but extracting problems
is difficult.  If I show a 10-minute tape of a typical session of
something that is a bit complex the results will vary considerably.
There are few places where people are trained in the art and science
of taking good notes (or in the case of automated logging - marking
the places where you then have to take good notes).  Note-taking is
one of the dirty secrets of our field.  People often see the same
event in different ways and inter-rater reliability is often low.  One
of the issues here is that it isn't always clear when a problem
starts.  Theoretically, a person might do something that does not
prompt a "problem alert", but later something happens because of this
unnoticed thing and a visible problem occurs.  Another subtle issue
that I've seen is that some people take notes on what they hear (the
thinking aloud) while others note what they see on the screen.
Sometimes a subtle hesitation or oscillation of the mouse between a
few links or menu items might be a problem, but in my experience, few
people will write that down.  Since the 1980s, when nearly everyone
watched hours of videotape, there has been a general trashing of
videotape viewing because it takes too long.  Well, yes, verbatim
videotaping is probably not cost effective, a quick scan of tapes,
especially with complex products will reveal things that people
missed, sometimes serious issues that are not obvious when a person is
focused intently - sometimes the issues might show up across task, but
there is often not a good way to connect that quickly in the
notetaking.  A colleaue and I did a little experiment where he took
notes during a session that last an hour on a complex network
performance product.  I watched a tape of the session and listed the
problems that I saw and the difference between our two lists was
significant.  I saw about 50% more problems - including some serious
ones.

The issue of what constitutes a "usability problem" is hard because
things that might be a problem for one user group might be a positive
thing for another group so even defining a problem requires some
reflection and consideration of the participants.

Oh, and then there is the severity rating issue - ratings of severity
by different people of the same thing are often quite unreliable. One
reason for the unreliability is that there is often limited training
on rating severity with examples and discussions of context. Much of
the research on this topic shoves a 1-4 or 1-5 or 1-10 scale at
participants with no real training or background and those who use
scales in the real-world fail to work on a convergence of
understanding in some semi-structured way.

So, even a "discount" or "guerilla" requries some rigor and training
and reflection (like having usability teams review a tape together to
help each other understand why they listed something as a problem or
not), and practice (there are very few HCI or Usability programs where
people get extensive training on the basic types of interviews -
structured, semi-structured, and unstructured interviews). Moreover, a
good HCI/UUX/IxDA plan should consider a mix of discount and
full-featured methods (for example, conducting field interviews for
the next version while conducting simplified thinking-aloud and user
interface inspections on the current product).

Discount methods still need rigor, practive, review, and reflection.

Chauncey



On 10/31/07, Peter Boersma <[EMAIL PROTECTED]> wrote:
> Alan asked:
> > [..] However after talking to them for a while it becomes clear that
> > they really don't know what Usability is [..]
> > So I ask you: Is this the price we pay for growing popularity of
> > Usability and User-Centered Design?
>
> Yes, and we should fight it with (drumroll...) discount usability methods.
> Jakob Nielsen wrote an article on Guerrilla HCI in 1994:
>
> Guerrilla HCI: Using Discount Usability Engineering to Penetrate the 
> Intimidation Barrier
> "I used the term "guerrilla HCI" in the title [..] because I believe that 
> simplified usability methods can be a way for a company to gradually build up 
> its reliance on systematic usability methods, starting with the bare minimum 
> and gradually progressing to a more refined lifecycle approach."
> (http://www.useit.com/papers/guerrilla_hci.html)
>
> If these simplified methods come to the people with less knowledge in a 
> well-documented and illustrated way, there is hope that they won't mess it up 
> and bring our practice down again.
>
> And, since this is IxDA and not CHI-Web, I believe each field under the UX 
> umbrella should have its own set of simplified methods, as I explained for 
> IAs on my blog in "A piece of IA pie: little, micro, lite or guerrilla?" 
> (http://www.peterboersma.com/blog/2005/02/piece-of-ia-pie-little-micro-lite-or.html
>  or http://tinyurl.com/c7tw4).
> IxDA.org's Resource Library (http://resources.ixda.org/) and the IA 
> Institute's Learning IA section (http://www.iainstitute.org/en/learn/) should 
> be examples of collections of these well-documented, simplified methods.
>
> Peter
> --
> Peter Boersma | Senior Interaction Designer | Info.nl
> http://www.peterboersma.com/blog | http://www.info.nl
> ________________________________________________________________
> *Come to IxDA Interaction08 | Savannah*
> February 8-10, 2008 in Savannah, GA, USA
> Register today: http://interaction08.ixda.org/
>
> ________________________________________________________________
> Welcome to the Interaction Design Association (IxDA)!
> To post to this list ....... [EMAIL PROTECTED]
> Unsubscribe ................ http://gamma.ixda.org/unsubscribe
> List Guidelines ............ http://gamma.ixda.org/guidelines
> List Help .................. http://gamma.ixda.org/help
>
________________________________________________________________
*Come to IxDA Interaction08 | Savannah*
February 8-10, 2008 in Savannah, GA, USA
Register today: http://interaction08.ixda.org/

________________________________________________________________
Welcome to the Interaction Design Association (IxDA)!
To post to this list ....... [EMAIL PROTECTED]
Unsubscribe ................ http://gamma.ixda.org/unsubscribe
List Guidelines ............ http://gamma.ixda.org/guidelines
List Help .................. http://gamma.ixda.org/help

Reply via email to