[Sigia-l] MVT: Another Kind of Testing

[Mitchell Gass]

An interesting article on Forbes.com about multivariable testing 
(registration required):

   http://www.forbes.com/forbes/2003/0512/132.html

Here's an excerpt:

----------
Circuit City randomly selected 16 stores for the experiment. The changes 
were either/or, and not all ideas were implemented in every store. 
Commissions were either flat or varied, based on product; employees wore 
either uniforms or street clothes; 0% financing lasted for 12 months or 24 
months.

Each of the 16 stores tried a different combination of the 15 ideas. In 
this way a lot of interactions can be tested at once (maybe the commissions 
only motivate uniformed sales clerks?). Data on sales and net margins at 
each store were tallied weekly. After the best-performing recipe was 
identified, Circuit City and QualPro fine-tuned it by eliminating those 
changes that had no impact when compared with the status quo.
----------

Mitchell Gass
uLab | PDA: Learning from Users | Designing with Users
Berkeley, CA 94707 USA
+1 510 525-6864 voice
+1 510 525-4246 fax
http://www.participatorydesign.com/