[Sigia-l] Classification is an essential skill%26In-Reply-To=%26lt;000901c2caf0$d86a4480$50aa5bd1 at satellite>

[Adam Greenfield]

> If you want to appeal to people and facilitate their processes you will
> architect towards 'who' they are, which is a 'category of person' full
> of the difference which defines them as unique individuals. Going about
> things the opposite way is 'contrary' and asks humans to behave as
> machines so 'the system' can continue to work. In short, your
> categorical paradise is de-humanizing and if you read my recent post on
> 'the decline of information' you might recognize this downward trend.

Very often I'll have someone on the client end of things push back a little
on the comparatively few personae I use to rough out a site structure,
generally three to five.

"How can you be sure these are the right personae?" they'll ask. "Our
marketing numbers show that the prime audience for this brand is 40-55, and
your persona here, Mrs. Yamaguchi, is 42. We're a little bit worried that
she's at the young end of that range, which will skew the numbers."

I always answer these challenges by explaining that my primary personae are
designed to capture, broadly, the range of roles in which users interact
with the site. They are not intended to be 100% accurate depictions of
actual users, nor will they speak to every nuanced response of every
demographic using the site.

Of course, if there is a meaningful difference in on-site behavior between a
42-year-old and a 50-year-old, then maybe this warrants an additional
persona. But the intent is never, precisely, to design one structure to
account for all the infinitely nuanced needs of each member of a mass
audience, which would be neither economic nor practically achievable.

Instead, the humanistically-inclined IA - which, Derek, I would think is
most of us - needs to design a structure to minimize confusion and assist
self-directed information seeking. That's "minimize" and "assist," not
"eliminate" and "guarantee." We can only do so much.

I think most of us are acutely aware of the frustration and even anguish bad
IA practices cause. I think most of us do everything we can to avoid
inflicting bad experiences on users. I think most of us know how sadly often
users blame themselves for their "inability" to use a poorly-designed site
"correctly."

While I'm glad you've decided to join the discussion, I still have to point
out that what you ask is not achievable, even in principle. We can never
architect sites to meet the individual, momentary needs of each member of an
arbitrarily large audience with any fidelity.

What we *can* do is (a) admit this and (b) work towards the development of
adaptive interfaces that allow users to configure their experiences in ways
that make sense to them. The aim, as ever, would be to enhance the user's
agency (and sense of agency) in navigating whatever body of information
confronts them.

I think this is a splendid goal, one worth working toward. I can't see how
it's dehumanizing.

A.
Tokyo