[IxDA Discuss] Rationale for *not* using UCD

[Jim Drew]

On Feb 1, 2007, at 2:21 PM, Jared M. Spool wrote:

>
> On Feb 1, 2007, at 10:13 AM, Phillip Hunter wrote:
>
>> Is it not a little more accurate to say that quantifying the
>> correlation is
>> difficult?  Certainly there is correlation.
>
> Quantifying the correlation is easy. You measure the amount of
> resources consumed by UCD. You measure the relative usability,
> consumer satisfaction, or customer engagement (your pick -- doesn't
> matter for the measures) of the resulting products. You plot the  
> chart.
>
> What you get is a scattergram with no clear correlative line. (The
> correlative line would be points clustering on the line itself,
> showing the more spent on UCD, the more the output measures increase.)

The more spent on UCD with respect to what, though?  Absolute  
dollars?  Dollar per schedule day?  Dollars per man hour over the  
entire team?  Weighted based on where it occurs in the cycle, or the  
development type (extreme, agile, waterfall)?  Weighted based on the  
skill level of the designers, or that of the programmers and testers?

There is probably little or no observable correlation because there  
are too many variables amongst the myriad projects you could get data  
from.  That doesn't mean there is no correlation (nor that there is  
one), only that you haven't seen one.

I would say that it's probably true that at some point, there's a  
decrease in the additional benefit from adding more UCD into the  
mix.  As a corollary to that, there may also be a point before which  
there's isn't much added benefit, either: you have to throw enough to  
money/time/people to make an observable effect beyond that which the  
rest of the team will supply just from their own best practices.

-- Jim