2 March 2020

University of Edinburgh Business School doctoral student Alex Christian interviews Corentin Curchod about his recent research.
Laptop screen with code

Corentin Curchod currently works as a Senior Lecturer in Strategic Management and Organisation at the University of Edinburgh Business School. His research focuses mostly on identity work, power and governance in relation to market relationships, and technological change. In this extended interview, he discusses recent research published in Administrative Science Quarterly.

Alex: I really enjoyed reading your article 'Working for an Algorithm: Power Asymmetries and Agency in Online Work Settings'. I think it serves as a nice exemplification of how online work settings tend to move away from formerly dyadic evaluation relationships between superior and employee towards a somewhat triadic relationship in which online customer evaluations play an increasingly powerful role. This message was impressively conveyed by illuminating the dark side of human-algorithm entanglement and how customer evaluations can potentially elicit power asymmetries in online work settings. Your article reveals power asymmetries at two levels—the transactional level and the governance level—as well as the three main mechanisms that invoke power asymmetries between customers and sellers at eBay: online customer evaluations as a new form of employee monitoring; algorithms as mediating and objectifying actor relations; and online evaluations as prompting sellers to find practical ways for retaining individual agency. Can you please tell us more about the context of the study? How and when did the idea for your article develop?

Corentin: The idea for 'Working for an Algorithm' initially emerged when meeting Nicolas Neysen in Belgium in 2010, who was a PhD student researching online platforms at that time, and later became co-author of the article. It was his initial idea to explore how online platforms changed the way of work and especially how intermediaries like eBay changed the relation between sellers and buyers in online settings. So, it all evolved from a general interest in acquiring a deeper understanding of online platforms and how they impact work. Interestingly, the data initially collected was on the staff from eBay, not on the sellers of the platform. This data was eventually not used in the paper, however one thing was striking: eBay staff members noticed that sellers seemed to be unhappy, but they did not really understand why.

Somehow, we felt that there must be something interesting going on, and worthwhile studying in more depth. In this sense, our research was clearly phenomenon-driven. Generally speaking, I am more inclined towards phenomenon-driven research designs rather than theory driven ones, because I see phenomena-driven research to be better suited for tackling practical problems in our society. On the other side, drawbacks of a phenomenon focus are that the theorising process is possibly very time-consuming, you have lots of data, and it is challenging to concentrate on a specific topic.

Alex: You already published an article with eBay as an empirical context in 2014. Did this paper influence the 2019 article, and if yes, how?

Corentin: Yes, it is true, the 'Working for an Algorithm' paper can be seen as a follow-up to the paper 'Categorisation and Identification: The Identity Work of "Business Sellers" on eBay' from 2014, which was also written with Gerardo Patriotta and Nicolas Neysen. We started writing that paper in 2012, and it eventually got published in the journal Human Relations in 2014. When I look back at this paper, I think I did not know how to be a good reviewer of myself at that time. I like to think that I have progressed as a researcher since then.

The initial framing in the earlier paper was about identity work and categorisation. In that sense, the theoretical framing was different from the 'Working for an Algorithm' paper. Although the same database was used in both papers, in that we had the same interviews, we did not focus on the same data in the sense that the findings were derived from different parts of the interviews and different quotes. That is why the second-order coding and theorisation are not the same.

I personally think the decision on what journal to submit to depends strongly on the type of paper you have written.

Alex: Often scholars face the difficult question of which journal they want to publish their work in. How was it in your case? Was it clear from the outset that you wanted to submit to Administrative Science Quarterly? Why did you think your article was a good fit?

Corentin: It was my idea to aim for Administrative Science Quarterly (ASQ). I don't know whether Gerardo would say the same if you interviewed him, but actually Gerardo was unsure about submitting to ASQ. We asked around, including some friends of Gerardo at EGOS, and they all said it would be too risky to try getting into ASQ. We ended up submitting to another journal in the first place. Believe it or not, the paper was desk-rejected. But after having read the feedback on it, it was clear that we disagreed with the decision and needed to do something. We thought the decision was not based on solid argumentation, and we decided to appeal. So, we appealed and the Editor-in-Chief kind of sided with the Senior Editor, which is quite usual apparently. I was really annoyed by that, because I thought it was a great case and a good story. After discussing how to proceed from there, I was the one who said "I wanted ASQ from the beginning, you wanted the other journal, now we go for ASQ."

I think the paper fits well for ASQ because it is phenomenon-driven. ASQ appreciates a good story based on data. I personally think the decision on what journal to submit to depends strongly on the type of paper you have written.

Alex: In your case, as you had already collected the data as part of an earlier paper, how did you go about the new theorising for this paper? Did you try to make the theory fit to the data story in hindsight?

Corentin: The storyline of the data has remained the same since the beginning, yet the coding has evolved. It was an iterative process. The storyline oriented us towards a central theory, firstly sensemaking and power, but then reviewers were not convinced by the sensemaking part. The first feedback from ASQ was: "we love the story behind your data, but we think your theory does not fit the data, especially the sensemaking part". As a result, we re-worked the data, focused on the power-related aspects, and dropped the sensemaking part, which really was not central in the data. So we totally agreed with the reviewers there. During the revision process, the theory, second-order, and third-order coding evolved in parallel. It was a constant back and forth between theory and data.

It is important to note that the first-order coding we did together with Nicolas at the very beginning has never changed. Only the second-order and third-order coding has changed. Metaphorically speaking, the bricks (first-order coding) remained the same, but the way we built the house was different (second-order and third-order codes).

It is rather the combined coherent whole which makes the article worthwhile.

By applying an inductive approach, we started with investigating the phenomenon and then interpreted the phenomenon based on theory. In this way, the theory helped simplify the story and made it more impactful, understandable, and useful for the reader.

In our article, the theoretical story and the empirical story are not very useful on their own; it is rather the combined coherent whole, which makes the article worthwhile. That is how you make a contribution empirically and theoretically.

Alex: Academics often highlight the importance of striving towards delivering an original contribution to a specific stream of literature. In this context, what would you say is the main theoretical contribution of your article?

Corentin: For me, writing the contribution is the most difficult part of the work. It's a whole paper by itself. The contributions were constantly evolving during the writing process, as we always challenged ourselves, and the reviewers challenged us, asking us to go more in depth, to leave nothing in the dark. I think we succeeded in linking evaluations and power and showed that the way people are managed at their workplace changed entirely because of online evaluations. So it is not like a peer-to-peer process any more whereby you are being supervised by your manager. This online evaluation system changes the way you are evaluated and creates a double layer of power in the form of the customer and the platform owner. The 'post-panopticon' is a nice analogy to understand the way it works, but again it did not appear suddenly—it was more an iterative process of writing and rewriting the contributions.

Alex: You said you went constantly back and forth between theory and data. Can you please elaborate on how you eventually decided on what data to focus on?

Corentin: Nicolas and I agreed on a specific research design from the very beginning. I remember we discussed it together in a pub in Brussels. The research design was mainly about what the objective of our research is and what was the best way to collect our data. In our case, the objective was to understand the role of online platforms in the relationships between buyers and sellers, and the way sellers were adapting their work practices to the intermediation of the platform.

eBay managers offered us access to a forum where sellers expressed their discontent with the platform so we could approach these sellers in order to set up interviews. In the end, we did not use the forum to contact sellers. We contacted sellers directly based on pre-defined criteria to avoid potential biases.

I think that having a rigorous research design is more important than choosing a specific method.

It was important that we had certain criteria to select the sellers. As we wanted to study work, we focused on interviewing business sellers only. It is their main job to sell stuff on eBay. Furthermore, we chose sellers with enough evaluations to avoid inexperienced sellers and also focused on sellers within a mix of product categories to avoid interviewing only book sellers, for example.

To sum it up, the research design was pretty rigorously designed from the start. It was more about who to interview, how to select them, and what the objective of the study was. I think that having a rigorous research design is more important than choosing a specific method.

Alex: You published your article as a team of four authors, with Gerardo Patriotta, Laurie Cohen, and Nicolas Neysen. Can you please elaborate a little on the specific roles each co-author played, as well as on the experiences of working with co-authors?

Corentin: Nicolas Neysen was mainly involved in the data collection and first-order coding, whereas Gerardo Patriotta and Laurie Cohen came in at a later point in the process and were not involved in the data collection. The collaboration with Gerardo and Laurie followed up from our work on the Human Relations collaboration. After we finished working on the Human Relations paper, Gerardo one day in 2015 told me "you know, I think there's something else in the data which could be interesting to study". And the interesting bit was "evaluations, work, and power". I thought about it and said "you're right", and then we decided to bring Laurie on board because she is an expert on work and professions.

We turned out to be very complementary.

It turned out to be particularly helpful to have Gerardo looking at the data with fresh eyes and not being already too drowned in the data, like I was. He quickly sensed that there was something about power in our data. With this input, our theorisation certainly got stronger. So, Gerardo took the role of the second co-author, that is, the person who does not write successive drafts (that is the job of the lead author) but who rewrites parts of the different drafts and brings conceptual inputs.

Laurie took on the role of the third author, which means she was contributing less to the writing part compared to me and Gerardo, but challenged our ideas, proposed new ones, and forced us to clarify everything we wrote. She was the reviewer before the official reviewers, so to speak. The different authors brought different bricks into the paper. We turned out to be very complementary.

Alex: What are your top tips for working with multiple co-authors and making the collaborative process as smooth as possible?

Corentin: I would try to be as clear as possible from the very beginning, and agree on rules for becoming a co-author. That can prevent tensions. Furthermore, you need to be aware that lively and tense conversations are a natural part of the process. Therefore it is important to choose co-authors you feel comfortable with. Moreover, it might help to have the meetings in nice surroundings instead of being in a dull office. Location and atmosphere of meetings matter, so you can for instance talk through your disagreements over a beer.

Alex: You said you were planning to make an impact for academia and practitioners. What do you think are the main impacts of your article?

Corentin: When you think of impact, you think about citations and ask yourself how much your paper will be cited. This is where the prestige of the journal plays a big role. Also, the way you frame your paper, the keywords and the title help you to get citations. To be honest, we focused mainly on this kind of impact.

The other kind of impact can be connected to the Research Excellence Framework (REF), the evaluation of research by the UK system. This is a way to prove that you have an impact on the broader community of practitioners. Sadly, we did not really think about writing a REF case from our article, because we were too focused on the publication process. Yet we think that our paper would have been a good fit for that purpose, because it can have an impact for practitioners. For instance, it could make people aware that these online platforms change the nature of work and how you are assessed at work. This in turn can have an impact on the wellbeing of the sellers, because they feel oppressed by something they don't really understand. So it could incite eBay and other online platforms to change the way they design the evaluation system, for example.

Alex: Finally, what would you recommend early career researchers do if they aim to publish a qualitative study in a well-renowned journal?

Try to make the process enjoyable, because it can be a hard and long tiresome process.

Corentin: My first advice would be to focus on the data. In my opinion, the quality of data is paramount. I believe that qualitative research will not be published if you have rubbish data, though you can be published if your data is great but your theoretical framework a little bit shaky. I would compare this with French cuisine: if you have bad ingredients, you can have the best recipe in the world, it would still make an average dish. Whereas if you have really good quality ingredients, then even if you don't have a really good recipe, the dish will still be fine. However, if you also have a good recipe, the dish will be amazing.

Another piece of advice is try to make the process enjoyable, because it can be a hard and long tiresome process. Reviews can be terribly depressing. This is why you should find ways to celebrate the small victories. For us, each time we managed to go through a revision, we celebrated.

My last piece of advice is that place and environment matters. Sometimes I need to be in a nice coffee shop, rather than in a clinical environment like an office. You need to be in a place where you feel comfortable. That makes the process more enjoyable.

Corentin Curchod

Corentin Curchod is a Senior Lecturer in Strategic Management and Organisation.

Read the original article: Working for an Algorithm: Power Asymmetries and Agency in Online Work Settings.