My personal colour-coded system: An interview with Davide Ravasi

Interview with Davide Ravasi
Professor of Management
Cass Business School
City University London

It is one thing to interview acquaintances and strangers, it is quite another to interview your friends.  This particular interview was a bit weird because Davide and I have been friends for quite a number of years now. When we did the interview my impression was that our conversation sounded stilted and contrived.  Here are things that we have discussed at length on multiple occasions. Davide is part of that inner circle of academic friends to whom I’ll send my drafts for friendly reviews and who will not waste time with niceties or hesitate even one second before tearing apart every sentence. I remember him telling me once, “You mustn’t give anything to your reviewers that they might grip onto to criticize you. Nothing. Keep it totally smooth and to the point.” And this on the revision of a paper I rewrote I don’t know how many times and in which he still found something to fault. As frustrating as these interactions were (and I admit there have been times when I questioned whether he was truly a friend or not!) I learned a tremendous amount, and I can’t deny that my writing in general has improved as a result.  We all need a circle of friends who are willing to be brutally honest with us. We gain resilience from it (something we need a lot of in this field) and our writing gets better.  I decided to interview Davide so that others can benefit from the very good advice he’s given me over the years. As a lot of this advice has come to me via his harsh readings of my papers, at least you get to benefit without the criticism that usually goes along with it!!

I’m going to ask you the same questions that I’ve asked pretty much everybody, and then we can improvise as we go along.  So the first question is:  Where do your ideas for papers come from?

Theoretical ideas come from the stories themselves, that is, from the cases I study.  I shouldn’t let this be known to my reviewers, but I rarely start with a clear theoretical research question in mind.  I find cases that for some reason intrigue me and I try to understand what’s going on. The theoretical implications of what I’m looking at become clear only later.

Where do you find these stories?

Organizations or phenomena that I find fascinating and that I want to know more about.

So you just think of an organization and then approach them?

Pretty much.  But those are mostly for my own studies.  In the last few years, I’ve been asked by colleagues to join their research. And since I found the stuff they were working on interesting, I jumped in. In other cases it was my doctorate students who had contacts with organizations that turned out to be interesting.

At the same time, a lot of the stuff that you’ve published so far does revolve around similar theories or ideas, around identity and culture and things like that.  So there is some underlying theory driving what you’re looking at in the cases, no? Or would you still say that it’s all very much inductive?

There is a perceptual bias for sure. When I look at a story, I tend to notice first the identity implications of it, or the cultural aspects of it. Others may see it as a wonderful capability story, or a routine story, or a governance story. I see identity.  And perhaps the cases that fascinate me the most in the first place are cases where some of these issues are particularly relevant.

So you’ve found a case that you find compelling, then what?  What do you do then?

I talk to people. I try to interview as many people as I can, collect as much archival material as I can, and then I read and reread the transcripts and the materials. And as I’m talking to people, and reading all this material, there will already be some general interpretation starting to form in my mind.  After that, it really becomes a matter of moving back and forth between this early interpretation and the data until I manage to find a correspondence between them, between the idea and the data that supports it.  So, for example, the latest study that I’ve initiated looks at corporate museums. I started out with a whole bunch of questions about who founded the museum, why they founded it, who manages it and so on. Later I came up with other, more indirect, questions that I used to capture broader or deeper understandings of corporate museums as a phenomenon.  So I asked questions like “What for you is the most important piece in this museum and why?” Why does this company need a museum?” “What would the company lose if this museum was closed?”  “Why should this museum be kept open?” None of these questions are theoretically framed. Rather they are aimed at capturing some deeper interpretation, some deeper meaning structures that these people have in terms of how they interpret the museum and its function within the organisation. And the questions can change over the course of the study, depending on what early insight I might have. Later in the process I may start asking questions that are more directly aimed at testing my early interpretations, or that are trying to gather additional support for them, or the opposite, trying to reject them.

Would you say then that you are actually theorizing, collecting data and writing all at the same time?

Absolutely. Even with studies that I join, I need to go through a period of full immersion in the data.  I’ll take two or three days where I read all the interviews, all the archival material, taking notes the whole time.  I don’t use data analysis software like Atlas Ti or NVivo.  I just use pen, paper, and coloured highlighters. That’s my technology, my coding technology.  And I have layers and layers of coding on the same sheet, reflecting all my thoughts at different points in time.

That’s interesting. Why don’t we talk a little bit about your coding practices then? 

Well, I take notes as I read.  And my notes are like tentative codes. Tentative open codes plus some of what purists might call “memos” which are essentially sentences that remind me about some potential insight. So this I’ll do for all the transcribed interviews that I have.  Then the next step would be to take all these segments and move them into word tables.  So I use tables in Word, very simple, where in one column you have the codes and in another column you have the citations. And sometimes I have a third column for second-order codes.

And you’ll cut and paste.

It’s purely cut and paste.  So the second time I go through all the interviews, it’s all cut and paste.  After that, I start writing tentative labels, and tentative definitions of labels.  At this stage, I might already be using different tables for different overarching dimensions or themes that have emerged from the coding process, so it’s not like it’s just this one big table: I’m already trying to separate the data and the dimensions into different tables.  So each table will have three to six or seven second-order codes, and these are possibly subdivided into even more codes. And the hierarchy of these codes may very well change over time, but this would be my first attempt at making sense of the data.  An intermediate step might involve cutting and pasting all the sentences that I think are interesting into a file, and then loosely grouping them into clusters, classifying them into groups with a tentative label.  Then I’ll move them from there to the tables. I also use colour codes.  So if something is highlighted in green, it means this is a very good quote to be used in the paper.

Okay, so you already do that at that stage? You haven’t quite figured out what your paper is going to be about yet, but…

This is a cool quote, you must keep that in mind.  Obviously not all of them will end up in the paper but these are the cool ones.  Blue means “This is interesting, but there isn’t enough there to make a code out of it. It’s tentatively assigned here, but probably should go somewhere else.”  Grey is, “It’s okay, it generally supports this code but since in the end you only have to show two or three of them, this is probably not one we’ll use, so we can delete it.” At the very end, for a specific code, you may have 12 really good quotes that directly support it. You’re not showing all of them because otherwise you’d have codes with 100 quotes or more attached. At this stage we’re talking about tables that are 60 pages long. So you obviously have to reduce this to something more manageable, like ten pages. This means that quite early on you have to already begin labelling some quotes as “delete” before you are even done coding all your data.  Finally, yellow is something that is in need of attention: a definition that needs to be fixed or some codes that need to be reconsidered.  So that’s my system, my personal colour-coded system.

So now you have your tables more or less reduced to some manageable amount, what happens then?

Well, it lets me know whether I have a paper or not. If I have a table, I have a paper.  Having a table means that you have some concepts and you have the evidence to support these concepts.  So you can build a story around it.

But you also have to trace relationships between those different concepts.  When do you do that?

Well, usually I’ll have some idea about how to connect different concepts, and on the back of some scrap of paper, I might draw, freehand, a tentative framework. So I’ll have a tentative way of organizing the concepts into some kind of model. Then of course, I’ll ask myself, “Can I write it up? Do I have a story to tell?”  Because intuitively I may think that I have a paper, but can I explain it? Can I find an explanation that flows logically and is true to what I observed?  And that’s something that I’ll find out only when I write the paper.

Okay, so why don’t we get into that, the actual writing.  How do you get started on that?

I generally start with an introduction that is going to be rewritten 16 times before the end.  The theory section is also a part that will evolve quite a lot throughout the writing of the paper.  When I start writing, I usually only have a general understanding of the literature I’m planning to use. It’s decent, but it’s probably not as deep as it becomes and as it needs to be by the time I reach the final stage of the writing process. So in the beginning, I’ll often write a general review of what seems to be relevant, and then I move on.  It may not be very refined, it will probably be just a general idea of “Okay, what is the relevant literature and how does this study connect to it?” Methods are usually quite self-contained, so I can write that any time.  Then the findings. To me, the findings section is the most important part because it’s in writing up the findings that you really figure out whether you have a story to tell or not. I spend a fair bit of time writing up the findings because I want to make them engaging to read, turn them into a really good story that is rich with details, and quotes, precise references to facts, events and so on.

So showing rather than telling.

Yes. Exactly.

So even before you start fine-tuning the theory, you’ll work on getting the findings right?

Well, you fine-tune the theory by writing the findings.  The findings already must have a theoretical structure, you know.  So if you manage to tell the story you’re planning to tell through the findings and there are no inconsistencies, then that’s a good sign.  If you manage to make it plausible and smooth, then it will be convincing.

Meaning that when you get around to writing your contribution, it should come together, just like that?

Well that’s different because, you see, there are many ways to spell out a paper’s contribution.  Some people expect you to spend ten pages writing about the implications of your observations for future research, to outline five, six, seven avenues for future research which nobody will ever pick up on anyway, and I don’t like that.  I like to use the discussion section to bring out the theoretical underpinnings of the story that I’ve told, to explain my observations, to bring out the theoretical explanation rather than saying, “This is important because… and this is important because…”  As a researcher, I’m able to judge by myself whether something is important or significant for my own research.  When I read the paper, I want the discussion to go deeply into the theoretical explanation of the narrative and the data that have been presented so far.  I don’t want the author to give me some rhetorical statements about why is this important. I may find a paper very important, but not necessarily for the same reasons claimed by the author.

But then, wouldn’t that deep explanation of the narrative come out in the findings section?

Well again, there are different schools of thought on that.  I like to keep the findings section relatively theory-free, so the findings are the findings. I can have a few references just to help frame these findings in theoretical terms, but for the kind of studies that I do, you cannot really observe directly the underlying cognitive mechanisms that explain the observations.  That is what I’ll talk about in the discussion section.  In the findings section, I’ll tell you what I’ve seen, organized theoretically.  In the discussion section, I will try to give you the theoretical rationale behind my interpretation of what I’ve seen.  That is how I like to write papers, okay?  This doesn’t mean that I always manage to write them like that.  In some cases, the editor or reviewers will ask you to write discussions differently. So sometimes I’ll first talk about my findings, organized theoretically, then I’ll summarize the emerging theoretical framework, which I’ll describe in purely theoretical terms. And then I’ll have another section where I discuss the implications of this, or how this contributes to improve our understanding of… whatever. But to me, the contribution to previous literature bit is more of a rhetorical exercise.

So obviously, with experience, you’ve developed your own idea about how these papers should be written?

With qualitative work it’s difficult.  I have developed my own idea about how I like to write papers but there are many ways, and depending on the preferences of the editor and the reviewers you may end up having to organize your data, interpretations, discussion, implications, and so on in different ways.

And do you find that that’s okay, that it’s part of the collective writing process?

Well, I know how I like to write a paper; I don’t know if my way is the right way. And sometimes you write a paper in a certain way because it’s worked before, but eventually you find reviewers that say, “No, you can’t do that.” And often they’re right.  I mean it may be that, depending on the type of data you have and the type of theoretical claims you make based on those data, there are different templates which may be more or less useful.

So you’re willing to be flexible?

I am.  And, I’m not saying this because this interview is going to be public, but reviewers are often – not always mind you – but often right.  If something is bugging them, then more often than not, there is something that isn’t right.  Meaning that there’s probably something in the way you presented things or wrote things that just doesn’t work.

Any writing habits? 

Well. It depends on what I’m writing. If I have to write up a whole new discussion, I probably need to re-familiarize myself with the literature, go through some of the main articles again, and so on. I cannot just open the computer and write a few lines.  When it’s about restructuring or revising the findings, or clarifying the theory, that’s probably something I can do like in 15, 20 minutes. I can work on that, stop, and start again later more easily. For more important revisions, I really have to get my mind into it – so thinking about it offline, what to do, how to justify certain things.  I have to get into that frame of mind.

How many papers might you work on at the same time?

It’s very difficult to work at this level of engagement on more than one paper at a time.  Of course it’s not always this level of engagement for all papers. But when you get down to actually writing or revising a paper, I cannot just switch from one paper to another. Some people can. I mean some of my colleagues are so prolific that they obviously manage to do it; but I can’t.

So finding the time to give all that focus is hard, even today after all these years of doing it?  Does it get easier though?  

Well, experience means that you know the tricks and you’re more aware of certain conventions.  You are able to produce more solid and more sophisticated arguments perhaps, or you get better at detecting the gaps and inconsistencies in your arguments, let’s put it that way.  Does it get easier?  I don’t know.  It’s still just as hard to get published.

Has your experience as editor changed the way you write? 

Maybe. I’m not sure.  If it has, it has been more at an unconscious level.  It’s more like the practice of evaluating other people’s work and looking at how reviewers evaluate other people’s work that has probably helped refine my general understanding of the whole process.  What I did learn is a lot of things about being an editor.

Why don’t we jump into that then, the review process?  How do you know when a paper is ready to be submitted? 

When I get nauseous just looking at the title. Usually, before I send out a paper, I’ll ask two, three, four friends to have a look at it. Friendly reviewers, people that are competent enough and that I’m on sufficiently good terms with that I know they will take it seriously.  I’ll incorporate their comments, look at the paper again, and then I’ll have a gut feeling that tells me, “You’ve done what you can at this stage, get it out.”

So now you’ve just got an email in your inbox which you know is a decision letter.  Now what? I open it right away. What’s your reaction?  What if it’s a reject?

If it’s a reject, I try to understand what happened. Then it depends. If it’s a paper that I like a lot and I really wanted it to work, then I will go back to it, fix what I can about it, and send it out again. Otherwise you just let it go and move on. Over the years I’ve had many rejections, so they don’t bother me that much anymore. In many cases, I’m even expecting it. I’ve become pretty good at predicting whether a paper has good chances or not.

I’d like to pick up on what you just said – on the one hand you think you know when a paper’s ready to go but at the same time you’re saying that you have a good feeling for whether a paper’s likely to get rejected or not.  How do you reconcile those two things?

If it’s a paper to AMJ, you’re prepared for the fact that they accept only 4%, 5%, 6% of submitted papers.  So sometimes we get published, sometimes we get rejected. It’s part of the game. I mean, you always have hopes, but you also have to be realistic. Not all the papers that you send to these journals will make it, no matter how good you think they are.

Sounds like you’ve become quite philosophical about the whole thing.

Let’s just say that it hasn’t happened to me recently that a paper that I had serious hopes  for was rejected.  So that may be the reason.  All the rejections I’ve had in the last few years were kind of expected. I mean, I would be a bit less philosophical if it was a paper that I really cherished and cared about that got rejected.  On the other hand, if I’m not convinced about why I got rejected, I may drop a line to the editor, not necessarily to complain, but to make a point or ask for advice.  What’s important is that you don’t get upset with the people involved.  This is a game and it’s part of the game, you know.  I might have rejected people I know well and I might have been rejected by people I know well and am friends with. If you can’t accept that, then you’ll have a problem doing this job.  Rejection is part of the game.  Unfair treatment is different, but I don’t think I’ve ever been treated unfairly.

So you trust the process.

I don’t know if I trust the process.  The process is what it is and you have to live with it.  And it’s flawed like any human process. My experience has been that most of the time it has helped improve my papers. Sometimes, something happens and you have the impression that you have been treated unfairly, but unfortunately this is part of the process.  And no process is perfect.  So you have to learn to live with that.

What about R&Rs then. How do you deal with those?

I read the letter and I try to get a sense of the general message.  Although you have to keep in mind that sometimes first impressions are wrong.  I remember once with a paper, I read the letter and thought, “Oh my God! We’ll never make it, we’re doomed, there’s no way we’re going get this paper published”.  And in the end, we did.

So don’t trust your first reaction to a letter.

When I get around to actually working on the paper, I’ll print the letter out and read it carefully, taking notes about what to do, what not to do.  Sometimes it’s like okay, “Do this, do that.”  Sometimes what the reviewers are asking is simple, it’s reasonable, so you just do it. Other times I’ll put a question mark next to a comment and ask myself, “How on earth do we do that?” or maybe a note that says, “Okay, we can do this, but bring in this literature or do this table or something.” And then I’ll meet with my co-authors and we’ll decide who does what. Then we’ll start working on it. Usually the first thing I do is work on the data.  So if there are problems with the data, like we need to provide additional data, or revisit the coding, I’ll work on that first. I want to make sure that all the problems that have to do with the empirical part of the paper are fixed.  I want to be sure on what ground I stand, what can I really claim based on the data that I have. Once I’ve done that, I’ll start writing the response letter.

You’ll start writing the response letter before you start rewriting the paper?

Yes, I prepare the response letter by copying the editor’s letter and the reviewers’ comments into a file. I then break up the comments into chunks that correspond to the list and numbering of the reviewers or even smaller chunks if I think that there are more issues to be addressed.  I make some space between each comment for where my answer is going to be and then I start answering every single point, one by one.  Sometimes some of the answers I write will eventually be cut and pasted into the paper. So for instance, in the paper that I’m working on now, I started writing a response to one of the reviewers that was asking about how context influenced the conditions of our observations and I wrote a couple of pages on that in the letter, but eventually I ended up putting it straight into the discussion section of the actual paper. And it’s only once the response letter is like 85% written that I’ll start rewriting the paper. Sometimes working on the paper will make me change my responses, because I realize that the problem was maybe a bit different from what I had understood it to be at first, so it becomes quite iterative after that.  Back and forth: letter, paper, letter, paper…

You mentioned that sometimes you might call an editor.  When would you do that?

When there’s a serious contradiction between what one reviewer wants and what another wants and the editor hasn’t taken a position in the letter. In those cases, you may want to drop a line to the editor, and say, “Look, reviewer 1 wants me to do this, and reviewer 2 wants me to do the opposite. What do you want?”  If you want to contact an editor, contact them with a very precise question that has to do with how to interpret reviewers’ comments or if there is some ambiguity or uncertainty about how to proceed or about what the editor wants exactly.

What about working with co-authors?  A lot of your papers are co-authored and you’ve co-authored with lots of different people.  What would you say is your recipe for success in co-authorship?

In my experience, everybody has to bring something different to the table, so somebody may have a better understanding of the data, somebody else may have a better understanding of the theory.  For me the most difficult co-authorships have been those where you have people that are too similar. Also, I usually work with people I like.  That helps.  In the end you’re going to have to work side by side with this person and make decisions, sometimes resolve disagreements. It’s easier if you work with people that you generally like and there’s a good relationship.

Any advice for aspiring writers seeking co-authors?

Work with people you like and make sure that everybody has a role. Be clear about who is doing what.

Do you ever get writers’ block?

Not really. I may end up procrastinating but that’s because I have other things to do. It’s not because I’m blocked, as in I don’t know what to write.  You always know what to write.  Like the method and the findings, you’re describing what you did, so what’s the problem there? And the discussion – if  you don’t know what to write in the discussion, just go back and reread all the pieces you cite in the theory section and see how your work is different from that. This doesn’t mean that everything I write, I keep; I trash a lot.  But at least I write.  If I get blocked on anything, it has more to do with making sense of the data.  For some of the papers I’ve written, I’ve produced three, four, five different visual representations of the model; three, four, five different versions of the findings.  It’s not that I’m like, “Oh! my God, I don’t know what to write.” It’s more like, “I don’t know how to write this well, so that when I reread it, I can say, “Okay, now it flows,” and I can move on. It’s more about that.

So in those cases you just keep writing until you get it right? You have faith that if you just keep at it, it will come together at some point?

And if it doesn’t, then maybe it means that there’s something wrong with your research design.

Has that ever happened, rethinking your design?

It might be a matter of collecting different data. It’s not like you trash what you have.  But it might mean finding other sources or reframing your claims.

What advice would you give to people who are starting out?

They should ask themselves, “Are you really sure you want do this?”  Don’t do what I did.  So advice number 1:  Focus. Focus on one paradigm, one theory, one problem. In the short term it pays.  If you do not, every paper is a whole new start.  You have to familiarize yourself with a whole body of literature and you don’t really associate your name to anything specific: you’re just churning out a bunch of cool stories.  So first, focus. Second, become a member of a community. In this job, alone we die; together we have a chance to survive.

Anything else?

If you’re a qualitative researcher, try to find intriguing, fascinating, unusual, different settings that will catch people’s attention and will somehow make them well disposed towards your paper because of that. If you have a fascinating setting, it may intrigue reviewers enough that regardless of your initial framing, they may help you out, and give you smart, insightful suggestions. The editor may give you a chance just because your data is so rare, so unusual, so interesting, and eventually you can take advantage of the help of all these smart people and thus increase your chances of getting published.

Interviewed and edited by Charlotte Cloutier