[Sigia-l] Re: Card sorting - do not like but what else is there

[infoarchitect at ourbrisbane.com]

Quoting Tania <tania_peakusability at optusnet.com.au>:
> From the section of Cooper's book that I have read (..) If you are 
> designing a car I agree but if you are designing a government website, 
> you have to cater for a number of different user groups (..) 

Many of us have faced the business requirement of "design the website for everyone in the 
world". This, of course, is ludicrous - so we identify target markets and design for them.  
Yes, Govt. sites may have a number of disparate target markets, but aren't the user goals 
going to be similar?  Sure, they may want to reach their goals for different reasons, but 
analysis will reveal similar/overlapping goals.

> In a recent project I worked on, we identified 43 different 
> user groups and developed personas and scenarios for all of them! (..) 
> However, I am happy to be further enlightened.

I'm not a staunch advocate of design purely by scenario, but creating useful personas and 
prioritising them can be very helpful in informing design, just as creating scenarios (and 
subsequent tasks) that match the goals of these personas can help inform the IA.  

Of the 43 user groups, did you create goal based personas or just demographic based user 
profiles?  Surely, not all 43 user groups had 43 completely different sets of goals?  Even if 
they are all disparate demographics or user groups, there would have to be some common goals 
and priorities upon which primary, secondary (and, if needs be, even tertiary) personas could 
have been based.  If not, then perhaps the analysis, or even the business goals should have 
been re-visited...  

> Regarding prototypes, in my view they are the "ends" rather than the 
> "means" and reflect the output of the IA processs - unless I missed 
> something.    

Bottom-up card sorting is good for identifying how people group things - this can help inform 
the categorisation of a directory.  Open card sorts are good for identifying how people would 
label groups.  As we have established, neither of these techniques are entirely suitable for 
identifying how people navigate whilst task-centric, as that requires context - yet these 
techniques are all that seems to be used these days... :P  

Of course, different projects require different methods to be investigated/utilised, but as an 
example, here are some suitable techniques that may inform a particular navigation structure:
1. Identify the true users and their goals - Contextual Inquiry (Holzblatt & Jones, 1993);
2. Aggregate common goals and develop primary and subordinate personas - Persona Modelling
(Cooper, 1999);
3. Identify alternative task sets to achieve these goals - Task Modelling (Hofstede & 
Brinkkemper, 1989);
4. Analyse and prioritise the required tasks - Hierarchical Task Analysis (Lindgaard, 1994);
5. Assess the tasks from the persona's paradigms - Cognitive Walkthrough (Rowley & Rhoades, 
1992), (Spencer, 2000), (Wharton et. al., 1994); and
6. Test design assumptions with users - Low Fidelity Wireframes (Rosenfeld & Morville, 1998).

I hope this helps.

Best regards,

Ash Donaldson
User Experience Designer