Inkwell: Authors and Artists

    
Inkwell welcomes Clay Spinuzzi, a professor of rhetoric and writing at
the University of Texas at Austin. He studies how people organize,
communicate, collaborate, and innovate at work. Spinuzzi has conducted
multiple workplace studies, resulting in several articles and three
books: Tracing Genres through Organizations (MIT Press, 2003); Network
(Cambridge University Press, 2008); and Topsight (Amazon CreateSpace,
2013). He blogs at spinuzzi.blogspot.com.

We'll be talking about his latest book, _Topsight: A Guide to
Studying, Diagnosing, and Fixing Information Flow in Organizations_
(http://clayspinuzzi.com/book/topsight/).  From the website for the
book: "Topsight—the overall understanding of the big picture—is hard to
achieve in organizations. There’s too much going on, too many moving
pieces. But without topsight, we have a hard time figuring out how
information circulates, where it gets stuck, and how we can get it
unstuck."

Clay, how does a professor of rhetoric get into studying how people
organize and work together? What experiences or forces led you on the
path to "topsight"?
  

    
Thanks for having me! 

I actually get that question a lot. The bottom line is that I've been
interested in user behavior ever since my undergraduate degree, in
which I majored in computer science and minored in technical writing.
The computer science major taught me a lot about systems; the technical
writing minor forced me to think through how people had to use those
systems. Technical documentation was an essential lifeline back then,
since people hadn't had much chance to develop computer literacy and
interfaces hadn't standardized.

Later, I started my Ph.D. program in rhetoric and professional
communication — essentially focused on how people write in and across
disciplines. And this question of how users behave stuck with me. So I
began conducting workplace studies. I knew how the textbooks said
people should act, but I wanted to see for myself.

It turns out that people often don't act the way the textbooks say
they do. In fact, they often act in ways that seem unusual or even
bizarre to outsiders.

For instance:

- When I studied software developers, I found that one group never
read comments embedded in the code. But they DID use those comments -
as landmarks when they were skimming a big chunk of code.

- When I studied traffic safety workers, I found that many of them
struggled when relating two different information sources (a map and an
interface). But a few didn't. One had zero troubles, mostly because
she used a sticky note to bridge the two information sources. 

- When I studied telecommunications workers, I found that they often
worked more closely with their counterparts at other organizations than
they did with people down the hall. 

These instances can't be explained by looking at a single piece of
writing in isolation. I had to look at the whole sociotechnical system
to figure out how the different components fit together -- and why they
broke down. 

And that's led me through a lot of theory (genre theory, activity
theory, actor-network theory) coming from a lot of different places
(writing studies, educational psychology, human-computer interaction,
computer-supported cooperative work, science and technology studies,
information studies, anthropology, sociology, management science, and
even warfare studies). 

Basically, I read all that stuff so that the readers of Topsight don't
have to. :)
  

    
Sounds like ethnography, is that a fair characterization of what
you've done with this book? I know you've tried to make it more
accessible, less academic. As someone who does a lot of academic
writing, has that been hard to do?
  

    
It has been surprisingly hard to do! Fortunately, I've been teaching
field methods courses since 2000, so I've had 13 years to hone the
materials. And of course I have a lot of war stories to share.

In fact, one thing that I really wanted to capture with this book -
and that I think so often gets leached out of more academic books - is
how *fascinating* it is to see how people work in organizations,
innovate, and work around their problems. People are born problem
solvers, and if you give them leeway, they will come up with the most
inventive solutions. 

Of course, those solutions don't always fit together well. So part of
the fun is to get a 50,000 foot view of the system and see where the
innovations and workarounds cluster. 

Let me get to your question about ethnography. The approach shares
some characteristics with ethnography, but it's more of a qualitative
case study approach. A true ethnography has these characteristics:

- it studies culture
- it involves a lot of time -- the rule of thumb is that if it's less
than six months, it can't be an ethnography.
- ethnographers tend to participate during their observations, then
write notes afterwards

In contrast, my approach

- studies a bounded activity
- involves less time -- weeks, not months
- involves writing detailed notes *during* observations so that you
can get to the second-by-second details of people's work

It doesn't get at culture, which is what ethnographers study, but
that's okay -- ethnography doesn't quite get at the things that
Topsight studies either.
  

    
The obvious question, after noting that the book's written to be more
accessible, is who it's written for? And possibly how is it already
being used?
  

    
Right, I wanted to reach out to three audiences.

First, consultants. Consultants often do this sort of work - entering
an organization, examining information flow, and figuring out where
things are going right and wrong. In fact, I gave a seminar on
workplace research at a consulting firm early last summer, and the
enthusiastic reaction was really the trigger for writing the book.

(I should mention here that I'll be doing two seminars on Topsight
here in Austin for the Information Institute:
https://infoinstitute.utexas.edu/)

Second, undergraduate and graduate students, especially in technical
and professional writing programs. The book is based on materials that
I've been teaching my own students since 2000, and professors in other
writing programs say they're eager to bring it to their own students
too. I'm really excited about this aspect -- Topsight provides concrete
skills that these students can apply to analyzing organizations so
that they can develop smart, best-fit information resources. I tell my
students that understanding the organization is fun, but changing it,
making it work better, is their mission.

Third, people working in these organizations themselves. Everyone
needs a little topsight: they need to be able to step back and see how
their work fits into a larger whole. Once they do, they can often see
why things go wrong upstream, why the manager's new suggestion won't
work well, why the folks in the mail room have so many sticky notes.
And they can make smart local changes that can make things easier for
themselves and others around them. 

The book's still very new, but I'm seeing some uptake from all three
readerships. For instance, the consulting firm I visited last summer is
very excited about the book. I just got back from an academic
conference in Las Vegas, and every professor I talked to wanted to look
at the book. And I've been reaching the third audience through local
TV, a quote in Financial Times -- and of course this interview.
  

    
Speaking of, here's the FT.com link: 

 
http://www.ft.com/intl/cms/s/0/3b793a3e-8cd0-11e2-aed2-00144feabdc0.html?ftcam
p=published_links%2Frss%2Fmanagement%2Ffeed%2F%2Fproduct#axzz2O6V0w4cS

And here's the local TV interview:

 
http://www.myfoxaustin.com/video?autoStart=true&topVideoCatNo=default&clipId=8
325964
  

    
Do you have any stories of specific organizations that have been
impacted by the book, or are beginning to use it to drive changes in
information flow?
  

    
Sure. Here are some examples.

First, the company I describe in my second book, Network. 

 
http://www.amazon.com/Network-Theorizing-Knowledge-Work-Telecommunications/dp/
0521895049/ref=sr_1_1?ie=UTF8&qid=1363873987&sr=8-1&keywords=spinuzzi+netw
ork

This was a telecommunications company I investigated for a research
project, using the same approaches and models I discuss in Topsight.
And they're a good example of a company that needed topsight: They had
expanded rapidly and their training models and information systems
hadn't kept up. Their plan was to train managers by having them work in
every part of the company; I was able to show them that this approach
wouldn't pay dividends because the company was changing too rapidly.
Instead, I suggested ways to network within the company so that they
could stabilize interfaces across teams.

Second, the firm I consulted for. This company was already familiar
with the basics of field research, so they would come back from
clients' sites with tons of data. But they needed Topsight's models in
order to sift through those data and find how the patterns related. 

Third, just for the sake of variation, I think of my undergraduate
students' projects. At the beginning of the course
(https://sites.google.com/a/utexas.edu/spinuzzi-2013s-rhe330c/), I tell
my students to select a small organization with which they have
contacts so that they can conduct their studies. Throughout the
semester, they study these organizations -- doctor's offices, student
organizations, intramural teams, real estate and marketing firms, Greek
organizations, you name it -- and they effect real change that helps
to identify and resolve the systemic problems in the organization. It's
rewarding for the students, as people who are often invested in these
organizations. It's rewarding for the organizations, which end up with
a much better understanding of their own problems. And it's very
rewarding for me: I strongly believe we should be making the world a
better place, even if it's only in these little ways, and the more
people I can train to gain topsight, the more people will have a chance
to improve their own lives.
  

    
Getting back to the book, why did you decide to self-publish? Some do
that because they can't find a publisher, but you're an established
academic author. Was the decision strategic in some way?
  

    
It was definitely strategic.

I've published two books with top-flight publishers: MIT Press and
Cambridge University Press. Both produced high-quality books, and they
allowed me to reach academic audiences, partly through their
outstanding reputations. They also got me promoted -- the first one got
me tenure, and the second was the cornerstone in my case for promotion
to full professor. 

In a sense, I'd made my reputation. And those academic presses gave me
things I couldn't have gotten otherwise. I can tell you for sure that
UT wouldn't have accepted a self-published book for a promotion case!
But once I'd climbed that hill, the idea of self-publishing the third
book became very attractive to me.

Three reasons:

1. The old adage about creative work is: "fast, cheap, good: pick any
two." But Amazon's publishing arm gave me the possibility of achieving
all three — a solid book, published quickly, sold relatively cheaply.
Self-publishing meant that I could put this book out there quickly: It
took only three months to go from a finished manuscript to publication.
(Compare that to the *two* *years* it took for my first book.) And it
allowed me to distribute it widely to lay audiences — undergrads,
consultants, people who want to better understand their organizations.
And without the layers of people working on the book, I could put it
out more cheaply: even though the print-on-demand publishing process is
actually more expensive, I save on labor. And of course I knew it
would be good, because I've been road-testing this material for years
in my classes and seminars.

2. Self-publishing allowed me more control over the tone of my work,
something that I thought was important in this kind of guide.
Traditional publishers don't like some of the things I do in this book:
using contractions, making pop culture references. Something as small
as referencing Scooby Doo can really set the tone for a book, making it
friendly and accessible, and I didn't want to give that up. Beyond
that, I didn't want to give up strategic control over the book and its
contents: it's a distillation of the work that I do, the consulting
that I do, and I want that under my own control.

3. Beyond the above, frankly, I just wanted to try something new. I
went into this project knowing that I might never make a dime off of
this book, and I had to be okay with that. I'm still okay with it,
because I strongly believe in the book and its mission. 

That being said, I'm pleased with its reception so far.  
  

    
Please feel free to share this conversation far and wide using this
short link: http://bit.ly/topsight-talk

If you want to be part of the conversation, but you're not a member of
the WELL, you can still add your comment or question. Just email
inkwell at well.com
  

    
Topsight is not just a book, but a methodology. Have you had a chance
to spend time with an organization actively using it as such? Do you
expect to see consulting practices form around your thinking?
  

    
Give your varied audience, why did you decide to "take on" the tone
you have--that is, with CSI, Scooby Doo, AND Actor-Network Theory
references in the same chapter?
  

    
Becky - Okay, so I have been wrestling with genre theory, activity
theory, and actor-network theory for years. Theory can get complicated
- especially when people are first writing about or formulating it. And
unfortunately, because theory is initially hard to express, it's often
a chore to read through. 

But it shouldn't be. When I'm reading a piece of heavy theory, I tend
to talk myself through it with plain examples. (That's probably due to
my training as a technical writer.) And once I can get my head around
it with some plain examples, why not share those examples with other
people?

And that's a big chunk of what Topsight is trying to do. To
investigate these complex problems, you have to have a solid
methodology. To have a solid methodology that holds together well, you
have to have theory. And to turn lay readers into theorists, you just
have to step through it in plain language. People who read Topsight may
not recognize that they've just learned a big chunk of sociocultural
theory, but -- if I'm successful -- they'll be able to apply its
insights.

Also, I just like Scooby Doo. The original ones. Not the
Laff-a-Lympics or Scrappy Doo stuff. Those were appalling.
  

    
Jon - Unfortunately I haven't been able to spend time with a company
as they actively use the Topsight methodology. I really want to. 

I do know that at least one consulting company is working topsight
into their standard operating procedures, but they're early in the
process, so I haven't gotten to see results yet. 
  

    
I'm glad to hear an academic admit to Scooby Doo love.  I noticed you
left CSI off the list, though.....

How are you avoiding the "one size fits all" idea that comes with a
lot of methods texts?  
  

    
I hope I have avoided that idea! 

There are lots and lots of methods out there to do different things,
from general approaches (ethnography, discourse analysis, experiments)
to more specific design approaches (such as participatory design and
contextual design). These come from different places and have different
theoretical commitments, but they all do important and useful things.

I think of them as tools in a toolbox (not a very original way of
thinking about methods, I know). Each excels at certain things and not
others. You might force one method to do something it really shouldn't
do -- just as you might use a crowbar to pound in nails -- but it works
much better to define the job and pick the right tool for it. 

So: If you want to understand workplace culture, you turn to an
ethnography. If you want to codesign with users, you use a
participatory design approach. If you want to establish causality, you
conduct experiments. 

But what do you do if you want to develop a multileveled understanding
of how people circulate information through an organization? For that
use, these other approaches are limited. Ethnography isn't focused on
this issue, it's focused on culture. Experimentation is hard to pull
off because the inputs aren't well defined -- you don't know what
people are actually doing until you go in and observe them. PD is not
great either, because it tends to pull people away from what they're
doing and focus them on new possibilities instead. Contextual design
comes the closest, but it's focused on designing specific systems and,
in my opinion, tends to overemphasize developing a centralized system -
something that reduces flexibility in the system.

For what *I* want to do - understand information flow in a bounded
organization - topsight works pretty well, I think. But for people who
want to take the next step and design new information sources, I use
the last chapter to point people to some of these other alternatives. 

To your other question: CSI just leaves me cold. I admit it. I threw
that reference in for people who aren't Scooby Doo fans.
  

    
<But what do you do if you want to develop a multileveled
understanding
of how people circulate information through an organization?>

Clay, were there some approaches you started to take and then realized
didn't work? And from that realization, are there some current
business practices that are being taught, but just not beneficial in
the end result?
  

    
Ted, sorry to take a while to answer. 

I remember when I finished my dissertation and then, relieved that I
had some time to relax, picked up Beyer and Holtzblatt's book
Contextual Design -- and thinking, why couldn't I have read this book
before? It was sharp, smart, and well integrated. I really liked it.
And for a few years, I taught it in my classes. 

Reading CD led me to look into participatory design, which led me to
begin folding it into my classes too, as well as writing a few papers
about PD as a movement and about specific PD techniques such as
prototyping.

In the meantime, I was reading everything I could get my hands on
about qualitative research, from case studies to ethnographies to
grounded theory. (I still do -- in fact, I review all of these books at
spinuzzi.blogspot.com , partly because I don't want that reading to go
to waste, partly because I want to remember what I thought about
them.) And I started to apply bits and pieces of all of these to my own
methodology as well, especially when I ran into wrinkles that my usual
toolkit couldn't address.

All of this reading and application really helped me to sensitize me
to the advantages and limits of each approach. In fact, that's a big
part of how I learn: by trying to figure out why Approach A is
different from Approach B. That's basically what my second book Network
is about. (It's a great read, and a great Christmas gift.)

Re business approaches - So let me give you an example based on what
I'm researching right now. I'm interested in how different types of
organizations tend to facilitate different sorts of collaborations,
embrace different sorts of values, and communicate in different styles.
And it turns out that there are a lot of different frameworks for
discussing those differences. For instance:

- Quinn and his collaborators describe a Competing Values Framework,
in which organizations are described as clans, bureaucracies, markets,
or adhocracies.

- Boisot and his collaborators describe I-Space, in which
organizations are describes as clans, markets, fiefs, and adhocracies.

- Snowden and his collaborators describe Cynefin, in which
organizations function within simple, complicated, complex, and chaotic
environments.

- Ronfeldt and his collaborators describe TIMN, in which societies
function as tribes, institutions, markets, and networks.

It's almost always four. Remarkable, right? I thought there was a deep
underlying significance until I read a piece by Snowden in which he
casually noted that these frameworks are usually based on a 2x2 matrix.
Ha ha.

But the point is that there are a lot of different ways to slice this
pizza. Superficially, the temptation is to just map these together and
say that they are talking about roughly the same kinds of
organizations. But if you look closer, you start seeing that they're
after different things: values, information transfer, environment,
society. And you can then start seeing their advantages and limitations
for different questions. 

And then -- and only then -- does it start making sense to pull them
back together and use them in tandem. These sorts of frameworks, like
research methodologies, shouldn't just be shoved together, but rather
carefully fitted together in ways that harness their complementary
strengths. Otherwise it's like taping the + sides of two batteries
together. It looks okay but ends badly.
  

    
PS, I'm sorry for the very long replies. Topsight is much pithier.
  

    
Great reply. That was very helpful in contextualizing the different
approaches being used and where they drill down. 

Complex adaptive systems theory is getting a lot of push these days.
Are you incorporating some of that into your analysis of businesses?
And if so, did it emerge from your computer science background?
  

    
I haven't attempted CAST yet -- not from lack of interest, but from
lack of background. We really are talking about things that could be
characterized as complex systems, I think, but from my limited
understanding, a CAS analysis involves quantitatively analyzing and
modeling the emergent properties of such systems.

And here we run into three problems. 

The first is that, despite my CS degree, I don't have the quantitative
background to keep up with a CAS analysis. I might collaborate with
someone at some point to make the two approaches talk to each other
(and I'd also love to bring in Social Network Analysis for similar
reasons), but at present, I'm sticking with the qualitative.

That brings us to the second problem, which is that interfacing the
two approaches would necessarily involve figuring out what to quantify
and how. Topsight really leverages the strengths of qualitative
research, including figuring out how individuals and teams interpret
local phenomena and how they leverage texts and tools in surprising
ways. For instance, I might find that a local, idiosyncratic innovation
(e.g., a sticky note, annotation style, or way of stacking paperwork)
provides a critical mediator that avoids common disruptions. These
little innovations are hard to discover without qualitative approaches,
so they tend to be opaque to quantitative approaches. To pair the
approaches, we'd have to figure out how to discover them via topsight
or another qualitative approach, then quantitatively model them in such
a way that we could generalize them for CAS. 

And here's the third problem. For Topsight (the book), I worry that a
CAS approach wouldn't be accessible to practitioners in the same way
that qualitative approaches would.
  

    
Clay, do you ever worry that qualitative research/techniques will get
a "bad rap" in industry?  

On another note, how important is a label?
  

    
Becky - a "label" as in a name for a methodology? I think it's helpful
when making the transition from an academic research method to one
that can be used more generally. 

I really sweated over "topsight" and over the names of the models -
"resource maps," "handoff chains," and the rest. All of these were
named something else ("genre ecology models," "communicative event
models") in my academic work, but those ungainly names just didn't do a
good job of conveying what the models did or how they worked within a
larger system.

I do worry that qualitative research, done poorly or haphazardly,
could result in a bad reputation. Of course, qualitative research isn't
the only kind of research that can be misused - we've all seen poor
use of statistics, biased survey questions, and the rest. But
qualitative research relies heavily on the researcher's interpretation,
making it easy to blame the researchrer (and, frankly, easy to be a
bad researcher). 

My approach has been to systematize a lot when developing Topsight.
Yes, the researcher still has to rely on interpretation and judgment.
But I walk my readers through specific steps of interpretation, warning
them about common missteps in interpretation, describing how to tie
together the different models, and insisting on evidence-based work at
all levels -- from data gathering to modeling to reporting. If they do
things right, readers should be able to anticipate and respond to
skeptical questions at any level. 
  

    
How can you be more accessible without sacrificing academic
credibility? I think you should be able to bring any kind of research
to the "real world" and make it relevant, without losing points in the
academic race, but practically speaking, can you?