What is the Best Practice for Usability Testing a Registration Process.

13 May 2009 - 1:35pm
5 years ago
3 replies
1723 reads
pcofrancesco
2009

Does any have any suggestions for how to best design a test for a
registration process.

We have 3 different scenarios involving a step process. If we take a
group of people and perform a usability test, should we put them
through all three scenarios? We're thinking this might skew results
as they'll become familiar with the scenario (because each scenario
has relatively minor nuances for differences). Also, they could
become fatigued.

Any suggestions how to produce the best results and most accurate
findings?

Thanks, Paul

Comments

14 May 2009 - 5:26am
dirtandrust
2008

I'll just quote Bill Buxton here from "Sketching User Experiences":

"Those who saw multiple designs gave us more critical comments. A
few even rejected one of the designs (something extremely rare in
studies that only evaluate one design, and therefore very
significant). But constructive suggestions? There was no difference
between users who saw one design and those who saw three."

Further, Bill makes the point that testing one design tends to shut
down criticism; the users tested don't want to appear to be too
negative.

Hopefully you can see that you'll get neither "Best Results" or
"Accurate Findings". What you will get are non-designer's opinions
about the interactions you've designed. It's up to you to take that
data and fashion an interface that best connects the user to your
registration process through to completion.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Posted from the new ixda.org
http://www.ixda.org/discuss?post=41992

14 May 2009 - 7:14am
James Page
2008

Just randomise the order of each scenarios.
It also can be quite interesting to ask people there expectation of how
hard they think the task will be before each task, and then ask them how
hard it was. See: the book Measuring the user experience.. Also see
http://measuringux.com/Tips&Tricks-Boston-UPA-2008.pdf
<http://measuringux.com/Tips&Tricks-Boston-UPA-2008.pdf> page
26.

All the best

James
http://blog.feralabs.com

2009/5/14 Nathaniel Flick <natobasso at gmail.com>

> I'll just quote Bill Buxton here from "Sketching User Experiences":
>
> "Those who saw multiple designs gave us more critical comments. A
> few even rejected one of the designs (something extremely rare in
> studies that only evaluate one design, and therefore very
> significant). But constructive suggestions? There was no difference
> between users who saw one design and those who saw three."
>
> Further, Bill makes the point that testing one design tends to shut
> down criticism; the users tested don't want to appear to be too
> negative.
>
> Hopefully you can see that you'll get neither "Best Results" or
> "Accurate Findings". What you will get are non-designer's opinions
> about the interactions you've designed. It's up to you to take that
> data and fashion an interface that best connects the user to your
> registration process through to completion.
>
>
>
>
> . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
> Posted from the new ixda.org
> http://www.ixda.org/discuss?post=41992
>
>
> ________________________________________________________________
> Welcome to the Interaction Design Association (IxDA)!
> To post to this list ....... discuss at ixda.org
> Unsubscribe ................ http://www.ixda.org/unsubscribe
> List Guidelines ............ http://www.ixda.org/guidelines
> List Help .................. http://www.ixda.org/help
>

14 May 2009 - 9:02am
gMulder
2009

What are you looking for - a qualitative or a quantitative result?

With sign ups etc. I was usually looking for quantitative results -
because I wanted to get the largest possible number of users through
it after launch. So in that case I would suggest showing them one
only and slightly increasing the "n" to get enough data.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Posted from the new ixda.org
http://www.ixda.org/discuss?post=41992

Syndicate content Get the feed