[Sigia-l] Making assumptions

[Jonathan Baker-Bates]

I've been thinking about research and design recently, and have had an
idea I'd like to pursue. But I thought I'd ask for a sanity check on it
here first.

The practical effect of research is to give me design assumptions with
which to work. But no matter how much research, persona building and
general selfless listening I do prior to any design activity, I find it
impossible to exclude personal (sometimes I like to call it
"professional") opinion about how something should be designed. This may
or may not have anything much to do with what I've observed or read
about in the research, but are generally things I would like to do
(something about trying to innovate, perhaps).

So, what would happen if I embraced my assumptions for what they are:
ideas which may or may not be correct, but that need to be validated.
So, in creating, say, a web site, I would be able to list all the
underlying assumptions that directed the design of that site. For
example, "People won't scroll below the fold on the home page, but will
do so on the result page", "Price is the only determinant of
conversion", "People don't mind paying credit card fees", "Slider
controls are confusing" etc. 

I can then verify those assumptions in whatever way I can, as and when I
get the opportunity later. Every time I find something that either
agrees with or contradicts an assumption, I note it. Some assumptions
will be strengthened, others weakened or disproved entirely. Still
others might need to be split into separate assumptions, and so on. All
the while, the assumptions list gives me ideas for design directions,
ammunition to support my arguments, and a general discussion guide for
any research I do. Some of the assumptions might be ones I initially
disagree with (the ones those marketing droids believe in, for example).
But I'd include them too.

As an aside, this would also bring a good deal of continuity across the
research programme overall. All too often, different types of research
reveal different things, or point to contradictory conclusions. Just
because I see most people in one user test ignore the fold, it doesn't
mean I can conclude anything much from that until I see the same thing
happen in, say, a remote test, and again in some clickstream data, or
another test, etc. It's not exactly "three strikes and you're out" but
it's getting that way.

So - is anyone else catalogue design assumptions about the use of their
systems, and using that to structure research over time? Does it sound
like a decent approach?

Jonathan