[Sigia-l] Real World UI Design Failure
Jonathan Baker-Bates
jonathan at bakerbates.com
Thu Jan 13 17:54:24 EST 2011
On 13 January 2011 00:27, Alexander Johannesen <
alexander.johannesen at gmail.com> wrote:
> What if you gave the pinned area a
> slight zap (you know that "go stronger color, then .5 second fade back
> to normal), or changed the color permanently when it goes into pinned
> mode?
>
>
We do have what we call a "throb" style for other elements, so I could
consider that on this pinned header bar. Thanks for the idea.
> > So we put them to the right as you see them today. So far the test
> results
> > haven't reached statistical significance so I can't say if that has
> restored
> > their usage yet.
>
> I'd jump in here and note that IMHO the loyalty program logo and text
> is *too* much like an ad, and your drop-downs (which has *no* relation
> to the ad itself) sit right in that area, which may be another reason
> it gets ignored more.
The pinned header was what we offered to allow the visual design to be less
ad-like. What you see there is the result of what the business regarded as a
big compromise - if they'd had their way it would have been much *more* like
an ad :-) But I take your point: this may be just a simple case of banner
blindness.
> Hmm, we the obvious thing here is that UX isn't as integrated into the
> development cycle as it possible should be, with the added problem of
> the time it takes to design and test any substantial UI. The only
> thing here is to get everybody on board with all things going through
> the same channels in the development cycle, even tiny maintainence
> change needs a paranoid UX guy before it gets unleashed.
>
>
Not sure what you mean. The paranoid UX guys spend substantial time on each
UI variant that gets built and tested. Nothing that faces the customer is
designed by anyone other than somebody in the UX team. Not sure if I gave
the wrong impression by mistake there.
I've developed a couple
> of sites that don't use A/B testing as such, but have the underlying
> concept built into the site structure, and made contextual based on
> different parameters like country, user preferences, market testing,
> etc.)
>
OK, but I'm not talking about the *implementation* of the tests (which is
perfectly simple from my point of view), I'm talking about interpreting
their *results*, which is usually far from clear because it can get
statistically very complicated extremely rapidly. And that's part of the
reason why we have 15 analysts!.
>
> > But on the other hand, the occasional universally clear result of a test
> can
> > sure settle any arguments :-) Would that were a more frequent event is
> all I
> > can say!
>
> Shorter cycles, more testing, long-term cyclic planning of larger
> changes. Yeah, when done right and when people get the concepts of
> test-driven development (and not the geeky version that only deals
> with code) in a more agile environment, life can be pretty good.
>
>
Again I'm not entirely sure what you mean. Could you expand on "test-driven
development" here? Customers do not produce a stack trace in my experience
:-)
Jonathan
More information about the Sigia-l
mailing list