[Sigia-l] From Research to IA

[Marc Rettig]

Hello Dan,
You have hit on a crucial bit of work that is getting too little attention
in the literature. I've been giving this a lot of attention in projects over
the past many years, but have been too busy to write it all down!

I recently gave a talk at the Institute for Design's "About, With and For"
conference on the topic of translating research data into design. Talks were
limited to 20 minutes, so I had to hit the highlights. Still, you might find
some helpful clues. The slides mention and show examples of cluster
analysis, insight-implication-solution chains, and something we call an
"alignment wall." 

Here is a pdf of the slides:
http://www.marcrettig.com/writings/rettig.atomsBetterThanBits.pdf

The alignment wall is more thoroughly documented in this case study from the
DUX conference:
http://www.marcrettig.com/writings/DUX_Herzfeldt_Rettig.pdf

Someone on the list mentioned persona. While these are certainly useful, I
have found them to be weak candidates if you only have time to do one thing
during data analysis. They are good for communicating insights, but leave
the team with yet another translation job: "here are the persona, great; how
should that make a difference in the design?" At their worst, persona are
works of pure speculation. At their best they do a good job of presenting a
synthesis of insights about people. In my experience they are terrible
vehicles for *generating* that synthesis. And, in my experience, they are
poor vehicles for describing insights about patterns of behavior,
relationships, and context. Maybe they're better at Cooper. I know many
people have good success with them, which is great. But when I've seen them
in use elsewhere I've found them quite shallow. Like anything (including the
alternatives I mention below), it's not the tool, it's the people *using*
the tool. 

Typically I try to facilitate teams to produce most or all of the following
artifacts during data analysis and synthesis:
o  an annotated task or activity model (by "annotated," I mean the sorts of
comments mentioned in the description of the alignment wall, perhaps also
supplemented with technology and business implications, by way of catalyzing
collaboration across areas of responsibility)

o  a cluster analysis of the data, usually with three levels: data,
clusters, neighborhoods of clusters. These are usually the first step in
analysis. They may not be documented, they may not be presented to the
client. But the work of creating such an analysis is great for immersing
everyone in the data, getting your head around it all. The clusters inform
the synthesis work that follows. 

o  a set of visual conceptual models, which vary almost every time to suit
the needs of the project. These represent the forces, constraints, people,
objects, systems,... the shape of the design space, the context of use. I
find this work to be the hardest to teach and the hardest to facilitate. But
it's gold when it's right. 

o  something like persona, which characterize the variations in people and
behavior in ways that are useful to the rest of the project. Rather than
rely on narrative alone, I often work to identify "dimensions of significant
variation," and show them as settings on a set of bars or continua. 

o  a chain of arguments or inferences:
   -  insights from research
   -  for each insight, one or more implications for the project or design
      (capture these *without* saying how you'll address the implication)
   -  for each implication, one or more possible solutions or ways to
      address the implication in the design. 

o  a series of workshops, tours of the data, role plays, posters, highlights
videos, co-design sessions, critiques... various immersive experiences for
the team and stakeholders. Whatever it takes to give everyone a sense of the
people, activities, and world for which they are designing and developing. 

Most of these artifacts build up in parallel, and are constructed by many
different collaborating members of the team. 

For each project, it's important to shape the results of analysis and
synthesis to suit the needs of the rest of the project. Look forward across
the project plan, and ask: "For each of these activities, what will we need
at hand to make sure our work is well grounded in the research?"

It's a little frustrating to try to capture a large area of practice in one
little email, and I imagine it might be frustrating for you to imagine what
I mean by some of these brief paragraphs. But I wanted to suggest a few
alternatives to persona, and thank you for raising this key challenge in the
work we all discuss on this forum. 

One last comment. You mention that you have data from several sources, in
several different forms. Different kinds of data benefit from different
kinds of analysis. Often my project plans show several different analysis
activities for data from different sources, all feeding into a period of
synthesis in which we try to pull it all into a coherent story / tool set.  

I'll stop there. Much success to you and your team. 

Marc Rettig
Fit Associates

marc at fitassociates.com
or,
mrettig at well.com



-----Original Message-----
From: sigia-l-bounces at asis.org [mailto:sigia-l-bounces at asis.org] On Behalf
Of Dan Linsky
Sent: Monday, January 03, 2005 2:44 PM
To: sigia-l at asis.org
Subject: [Sigia-l] From Research to IA

I am fortunate enough to be working on a project that has the
appropriate time and budget for necessary research before building a
large-scale website.  The research plan includes online surveys, mail
surveys, phone interviews, contextual inquiry, competitive analysis,
internal proprietary data and usability testing.

Now that I have all of this raw research data, I am having trouble
translating research findings into functional requirements.  Can anyone
suggest resources, charts/graphs, documents or other means of outlining
this to the company and client?

Any help is greatly appreciated -
Dan Linsky

------------
When replying, please *trim your post* as much as possible.
*Plain text, please; NO Attachments

Searchable list archive:   http://www.info-arch.org/lists/sigia-l/
________________________________________
Sigia-l mailing list -- post to: Sigia-l at asis.org
Changes to subscription: http://mail.asis.org/mailman/listinfo/sigia-l