Breakthroughs – Cleaned Transcript
Episode interview with Michelangelo Rossi (HEC Paris)
Hi everyone and welcome to this latest Breakthroughs podcast. Today I’m pleased to welcome an academic researcher whose work goes from what Oscar nominations do to our expectations… to our behavior when Google Maps adds an extra step in its search mechanism. And his name is…
So my name is Michelangelo Rossi and I'm an Associate Professor of Marketing here at HEC Paris.
YOUTUBE Extract 01
And the Oscar goes to…”
It was an Oscar ceremony watched live by almost 20 million people last year. And, as always, the buildup began weeks before with the announcement of the nominations…
YOUTUBE Extract 02
The 2026 nominations brought an unprecedented media and social media buzz this year. Maybe it’s thanks to the record number of nominations for Sinners, or the attraction of Timothée Chalamet and Wunmi Mosaku that saw over two million social media mentions in the five days after the January 22 ceremony …
Nominations, awards, badges or rankings. At regular intervals, organizations release what researchers call “prestige signals”. But what does this do to our expectations? HEC marketing academic Michelangelo Rossi and Hi! PARIS post-doctorate Felix Schleef ask that precise question. They study countless data and show that these signals can be double-edged swords. They can satisfy our expectations… or they can inflate what we expect beyond what the experience can deliver. 
To find out more, I went to see Michelangelo and his delicious expresso coffee machine in HEC’s W building…
Pl 3
SFX + chitchat…
Audio file 2: 2026 01 21 ROSSI ITW (2).wav
[00:00:02] Michelangelo Rossi: So my name is Michelangelo Rossi and I'm an Associate Professor of Marketing here at HEC Paris.
Daniel Brown: Michelangelo, before we get into the 2025 research paper called Quality Disclosure and Disappointment, Evidence from the Academy Nominations, can you first tell the audience what brought you and your co-author, Felix Schliff, to look at Oscar nominations in the 1st place?
[00:00:28] Michelangelo Rossi: Actually, it started in a personal way. almost, I would say, more than 10 years ago, I invited a girl who will later become my wife to the cinema we decided to watch La La Land And at that time, there was an enormous buzz around the movie. It got nominated, then awarded Everybody was talking about it. So we watched the movie, and then we discovered two things. So the first thing is that, okay, we didn't like that much the movie. But the other thing that's probably more important is that both of us don't like musicals And essentially there I started to think  that, okay, because of the nomination, generally, or the buzz around the movie, You could have two things You could have disappointment Meaning that everybody was telling us that La La Land was the movie to watch, and probably the movie didn't live up to our expectations. But on the other side, there could have been many people like us that don't like musicals, and they will have never gone to watch La La Land in the 1st place without the nomination. So from that moment, I started to think how we could disentangle the two and identify disappointment from selection.
YOUTUBE
Daniel Brown: We're exchanging for Breakthroughs on two research papers, not the most recent to begin with because it's January the 21st and January the 22nd happens to be the moments where the Oscar nominations are made public and it makes all the more relevant your earlier paper On this academy and the disappointments or the quality disclosures that this can engender. So what makes you think nominations are a good setting to study what a quality stamp changes in terms of people's ratings?
Michelangelo Rossi: We could think that many certifications, quality disclosures could play a similar role But here, movies, right, are particularly important because Prices does not change The price of a movie, of a ticket for a movie theater more or less stays the same before and after. And the quality of a movie is fixed So if you think, for example, at another case studied, by the way, like the Michelin star for a restaurant, well, the price of a restaurant can change The effort put in the cook, by the cook, could change And this could make this identification problem more complicated.
Daniel Brown: And indeed there are several sectors in your paper that are briefly mentioned, hospitality, car sales, maybe even business schools. Could your rather counterintuitive conclusions in the movie sector be transposed to these other sectors.
Michelangelo Rossi: I believe so. I believe so. At least the channel could be there. But of course, again, as I said before, to some extent, the specific characteristics of the movie industry and movie ratings make our life a little bit easier to understand the relationship between quality certifications and potential disappointment.
Daniel Brown: Your data focuses on a very long period, but that ends in 2019. Any changes since then, or would you say the conclusions are pretty much the same?
Michelangelo Rossi: So we, I want to refrain talking about post and pre-COVID Because probably after COVID, certain things have changed. But we do observe that the effect is stronger between, I would say, the golden years of the Oscars. So the effect on the audience perception seems to be there more massively for the 90s, the early 2000s So if you think about, I don't know, the Titanic or the big movies that got awarded, I would say that we do observe that in the last decades the effect starts to be a bit less there, also we could claim because the relevance of the Oscars nomination and awards has changed a little bit. 
Daniel Brown: Let's go into the specifics a bit of your research. in this paper you look at, as we've said, Oscar nominations as quality disclosure, a signal that should raise expectations in a sentence. But for you, what's the central result about what happens to consumer satisfaction after nominations in terms of how you track the ratings change after the nominations are announced? And you compare it with similar films that weren't nominated. What do you see?
Michelangelo Rossi: So the first things that we wanted to see, right, that is probably the most important, not even the most important, but the very first part that is, relevant to continue our, investigation is that, okay, if we look at ratings before and after nominations for nominated movies, and let me say similar non-nominated movies, we see that ratings of nominated movies go down, they decrease. This is, I would say, it is a very important starting point, but of course, it's not the end of the story Because again, this decrease in ratings, just comparing nominated versus non-nominated movies, could actually be related to selection or disappointment. So this first exercise is a very important starting point, but we can do an extra step.
JINGLE
[00:07:53] Daniel Brown: Michelangelo Rossi, a big issue is that nominations can change who watches and rates a film, not just how they feel. Can you walk us through the difference between what you call selection effects and the disappointment effect that you want to isolate?
[00:08:09] Michelangelo Rossi: Yeah, exactly. Back to the starting example of me and my wife La La Land. La La Land. La La Land is actually a nice movie But if you don't like musicals, right, you are not going to enjoy much And it's possible that you decide to go to watch La La Land because everybody's talking about that and they got nominated. Now, this is the selection So you could simply have A moviegoer who is a fan of musicals, watching the movie before, gets a very high ratings, but not because he loves La La Land, he loves musicals And then it happens that there is me or my wife, we don't like musicals, and we give a very low ratings, again, not because of, necessarily because of disappointment, Because we don't like them. Then of course, one could say, let us try to remove this, what we call selection story And try to look at expectations and disappointment.
Daniel Brown: Marketing papers are actually followed by quite a wide public. So without a crystal ball, it's been out there for about a year in Thanks to its publication. Without a crystal ball, what impact could your results and your analysis have on the movie industry and the awards like the Oscars?
[00:09:45] Michelangelo Rossi: Again, we don't want to say, right, that certifications and awards are bad If you go to, imagine you go to a new city you have never been, you want to discover what is the best restaurant, probably you are going to the restaurants with a higher ratings or, let's say, that have received a Michelin star. That's good What I think is important across many industries, also the movie industries, if you want, is that, we have to be aware that, for example, if we care about ratings, and usually a lot of firms and companies, industries do care about ratings, we have to expect that sometimes ratings, yes, they tell us something about the product, but they also tell us something about us, about the reviewers And a certification can indeed affect how we Feel, how we think about the product.
Daniel Brown: But strategically, can managers, can people who are in those industries accept this and change their strategy in terms of how they present it to the public so that disappointment and expectations are brought down?
Michelangelo Rossi: We actually observe anecdotically, right, that across different years and decades, the way producers and movie producers, right, decides to release The movies has changed And there are some, papers suggesting that this could be driven by a expectation management If you are releasing the movie massively, right, you are expecting a lot of people watching Whereas if you are only selecting in a few theaters, probably you could have a different type of expectation. So we do believe that this channel could be relevant and it's probably taking into consideration by movie producers. And others.
Daniel Brown: Michelangelo Rossi, can you say if the drop is just because different kinds of people start watching after the nominations or are the same kinds of viewers actually more disappointed?
[00:12:04] Michelangelo Rossi: So What we do, apart from comparing ratings before and after for nominated and non-nominated, we say, Okay, let's hold, let's wait for a second. We could actually focus only on... Nominated movies We have ratings before and after. And then what do we do? We consider only ratings that appear before nominations across different years And we train our recommendation system, something similarly to what we all experience, for example, using Netflix. So for every users, and once I say users, I'm meaning for every raters, a person who rates, posts some ratings on the platform We use, we could create recommendations, specifically predicted ratings, for all the movies they watch after the nominations. And so in this way, we could say, okay, now let's take two users. One has watched the movie as rated the movie before and one has rated the movie after nominations. But these two Users, right, have the same recommendations for all the movies on the platform It is to say, look, probably, me and you, we have different tastes, but I can find someone very, very similar in terms of taste compared to you, because that's the very same recommendations for all the movies on the platform. And probably if I compare you with this twin, actually we can check Disappointment somehow.
[00:13:51] Daniel Brown: And these are taste profiles that you create. And that's really interesting, as you say, very clever. What does it tell you about disappointment once you control for who's showing up for who, these profiles.
[00:14:06] Michelangelo Rossi: Yeah, we actually observed that this drop in ratings that we observed also by simply comparing nominated versus non-nominated, it stays. It actually, it stays in the magnitude. And one could claim, okay, but if you look at simply at the change in the stars rating, we observe it probably is not super large. But if you look at the difference in the average ratings between nominated and not nominated movies, we will expect that nominated movies Have higher quality, so higher ratings This disappointment effect accounts for more or less 7% of this change So if you think that, I don't know, let's say like a movie like Titanic, who got a lot of awards, well, Is better than a movie that is not a similar movie, say, that is not nominated, well, this change in terms of ratings is not small And this disappointment, this short-term disappointment accounts for 7% of it. So it's quite big, I would say.
[00:15:17] Daniel Brown: Michelangelo Rossi, you also introduce experience and you find that the effect is stronger for less experienced users. What does that imply in terms of the impact of disappointment? Why are less experienced users more vulnerable to this?
[00:15:38] Michelangelo Rossi: Like a signal of quality, like the awards or nominations, Impact people differently So if you are a movie critic, probably you already know what is a good movie and what movie is actually not very good And you may decide to watch the movie independently of the Academy nominations or awards You already know everything you have read extensively and so on and so forth. So we could expect that these people, these moviegoers, are actually not very affected by The Academy Awards. Whereas if nothing or imagine you go to the theater only once a year, probably because you heard about Academy nominations and awards, well, in that specific case, probably, you really are affected the most.
[00:16:33] Daniel Brown: Michelangelo, you also show placebo tests. In other words, predicted ratings don't move even as actual ratings do. Which do we understand from this?
[00:16:45] Michelangelo Rossi: Here it's somehow saying we try to use this placebo test to see whether this machinery that we created, this recommendation system, indeed captures somehow well these tastes Because Now, we don't observe any effect, and we should not observe any effect, because if imagine that me and you, we have exactly the same predictions, same recommendations for all the movies Our predicted ratings, if you watch it before or I watch it after, the predicted rating should not change massively. In effect, it doesn't. But our actual rating changes because You watched before, you didn't get disappointed, whereas I watched later and I got disappointed. But the predicted ratings, I should have rated according to this recommendation system, does not change, because indeed we are pretty similar.
[00:17:47] Daniel Brown: Finally, you also reflect on the implications on consumer welfare. Can you sum up your conclusions on that?
[00:17:56] Michelangelo Rossi: So the classical econ 101 story about asymmetry of informations and certifications says, well, if there are, say, used cars and new cars, and don't know the quality of a used card, actually a certification from mechanics is good. It's increasing the user welfare because now we can tell whether there are used cards are lemons or not, right, or very good cars. What we are saying is that that's true, that remains true, but probably the story is not ending here Because a certification could affect our utility and affecting our utility could also affect the welfare of consumers So I will say that this is a cautionary tales about the different possible effects of certifications.

Audio file 3: 2026 01 21 ROSSI ITW (3).wav
[00:00:01] Daniel Brown: Michelangelo, let's try to connect this research link to movie awards with your other paper that has come out quite recently in Marketing Science. It's called, as I said at the top of the program, Is Competition Only One Click Away, The Digital Market Act's Impact on Google Maps. And it's co-signed this time with Louis Daniel Pape. He's an assistant professor at a place, I think you worked a bit at the Telecom Paris Crest. In the Oscars work, there's an institution change, well, sorry, in the Oscars work, an institution changes the information environment by disclosing quality, expectations move and ratings move. In the DMA work, policy changes the choice architecture by removing a one-click path. And I'm really summing up very succinctly. But is it fair to say that both papers ask this question: when you change what's salient or easy in the environment, do people actually change behavior or do they snap back to the default?
[00:01:20] Michelangelo Rossi: That's a nice perspective, I would say. I never thought so much about the link between the two, but I agree with you that the anchoring, the human anchoring, right, plays a role in both cases. On one side, this anchoring channel is related to expectations That's changed our ratings. Whereas I agree that in a fairly different setting and environment, this anchoring part could be related to the default. So if we used always a certain type of service, independently of a small changes in the architectural infrastructure, well, we are going to keep on doing what we have been doing in the past five years or 10.
[00:02:04] Daniel Brown: Do you see other links between the two? Perhaps your approach personally as a researcher, you have a similar approach or not? And perhaps you introduce, I don't know, behavioral psychology in both? I'm just pitching that.
[00:02:19] Michelangelo Rossi: So I would say that there is a methodological similarity in both cases. Because in both works, we focus on a shock. And we try to understand, okay, there is a group of, we could think in this case, moviegoers or countries that receive a treatment Think about it, an experiment. And then there is a control group: Some users or some countries that did not get this change, this shock And we analyze it before and after. So I will say that in both cases, and generally speaking, my research tries to focus on these quasi Experiment to understand something causal about policies or, in the case of the Oscars, like awards. Yeah.
[00:03:10] Daniel Brown: Let's focus on the work that you've done on the DMA and its impact on Google Maps. First of all, for people who aren't following this DMA, can you describe it and why it's interesting in a market context?
[00:03:26] Michelangelo Rossi: Yeah, so you probably remembered in 2024 at the very beginning, you may be, imagine that you've typed something like, HEC Paris And you observe that there is a small little map after you are writing on Google, for example, HEC Paris, but that map, you can't click on it, whereas before you could. It's not you, it's not your computer, there is no bug, no problem, but it's actually, precise policy by the European Commission, right, inside this very large initiative that is called the Digital Markets Act, that wants to try to, how could you say, remove the bundling of two services, Google Search and Google Maps. Essentially, the European Commission said, hey, it's possible that a lot of people are using Google Maps, right, and mapping services are not really contestable, meaning that you cannot really create mapping services because everybody's using Google Maps, and possibly also because since everybody's using Google Search, they directly go to Google Maps. So you know what, we want to remove that channel. I believe that this idea is, it's not necessarily a bad idea
[00:04:46] Daniel Brown: No, it's noble. It's trying to level the playing field.
[00:04:49] Michelangelo Rossi: Absolutely. The question is how you do it and how easy is to make a mapping services contestable It turns out what they try to do or what they implemented By the way, this was how Google, right, received the DMA. There was no one in the European Commission who specifically said, “you have to do this” They did it. They unbundled the two services And the question is, was it successful or at least did this change make the mapping services contestable? The answer is not really. We observe in our paper that essentially massively we observe that in European Union countries, people, users typing for maps, for Google Maps, and whereas this doesn't happen, for example, in other countries like the United States So we observe that we have a lot of search to go back to Google Maps, essentially, and the traffic of Google Maps didn't change. so they our point is that probably people want to are not are open to other alternatives they are in some cases searching for maps, not necessarily for Google Maps, but if you search for maps on Google search, you are very likely to end up in Google Maps That's that somehow creates the difficulty, an obstacle to indeed finding out new services.
YOUTUBE EXTRACT ON DMA
MIKE: Industry experts trying to unpack the EU’s digital act and its effects on tech giants like Google…
[00:06:29] Daniel Brown: Now, Michelangelo, some of our listeners might say, well, this is such a small detail, but what you also show is that the extra 3 seconds that hypothetically it could take for them to click again to go, that step further has a huge cost or quite a cost. You estimate 3.3 million EUR a year are lost or spent as a result. I'm not too sure about exactly what it is. So, one has to weigh the benefits of limiting Google's steering the consumers towards them, towards, is it worth it? Can you elaborate on that?
[00:07:11] Michelangelo Rossi: Yeah, that's a very good point. So we presented this paper on multiple conferences, right, in Europe and in the US, for example. And actually you can see that the receptions of the very same results is very different. So, the US approach It will tell you, look, you have changed something A bit like the GDPR. You have just made our experience of the internet more complicated, more, you degraded the service But It didn't happen anything particularly favorable for the European consumers. This is the US point of view. The European point of view probably is telling you, look, actually you see a lot of new searches for maps, not for Google Maps. That means that actually people want to search So probably at one point, you will end up with a new European mapping service. Now, I don't want, I just want to know that these are the two extremes I don't want to pick a side. I'm not judging. I agree that to some extent both of them are true But let me say that we do observe that at least right now, aside from the calculation, right, that one there could be difficult, different details in the actual measurement Back on the envelope calculations are always complicated. But I would say that clearly right now, European consumers are Bearing the cost of this change Who knows, maybe in the future this could create, new environments, new scenarios, new services. Let's see Who knows, maybe this is just the very first step. But of course, right now the costs are bared by the European consumers.
[00:09:00] Daniel Brown: Just to clarify that, you make this estimate in your paper of 3.3 million EUR costing the European consumer. How does it cost that money? Where does this money, I'm just a bit naive. Where does this money go? Why does it cost the consumers 3.3 million EUR? Just the time when you add it all up.
[00:09:22] Michelangelo Rossi: So what we tried to do, by the way, in the very last version, we decided not to not code in. Yeah, it gets complicated because, so what we did, I tell you what we did. So we are saying, okay, so we have a... We can say that there are new searches Extra searches. Before you were not searching You're searching And you could think, okay, this search, these extra searches Imagine that one search is, say, one second If you sum all them up, right, you have a number in terms of time, the hours, days Months And then you can say, well, look, let us take the hourly wage The average hourly wage The European Union, and I can give you an estimate. Now, why we decided to not actually remove it later on is because, yes, to some extent you could think that these are extra searches Before you were not searching for maps, and now you are searching for maps. But on the other side, okay, imagine before you were searching for HSE Paris, and you were clicking for The map that was there Now, maybe at the beginning, you have to do that. You realize, oh, the clickable map is actually not anymore clickable, so let me now click for maps. But it's possible that after months, Now you stop searching for HSE Paris. You directly go to maps. So actually these are not necessarily extra, extra searches. So that's why we decided to say, hey, yeah, true. We observe that for sure there is, there are, if you stamp all them up, there are probably days of months of extra searches, but we don't want to really tell you a number that probably it's not, it's contestable Yes, exactly.
[00:11:11] Daniel Brown: The policy goal is competition. And in your results, do alternatives meaningfully gain either in search interest or in traffic? Or is the story more or less more friction, same destination? I think you've partly answered that question.
[00:11:26] Michelangelo Rossi: Let me say that we do not observe competitors like Bing Maps increase their traffics So we don't see massively, users in Europe saying, oh, let me go to Bing Maps now that I do not have, these clickable Google Maps. On the other side, I have to say, and this is anecdotal, meaning that we cannot really do a before-after treated control analysis, what we used to call diff-in-diff, but it has to be observed that In July 2024, Apple Maps made a move, meaning that right now and from July 2024, Apple Maps is also available on desktop for everyone, not only, Mac users. They didn't do that before. They did it now. Can I tell you that they did it because of the DMA? I don't know. I cannot. I'm not inside.
[00:12:24] Daniel Brown: Strange coincidence.
[00:12:25] Michelangelo Rossi: There is at least a temporal correlation. And so one could claim that one way or another, right now in 2026, the mapping services market is populated by an extra contender. Is it because of the DMA? I don't know. But this is the reality, it’s what now happened now.
[00:12:46] Daniel Brown: Could have diversified the market. Michelangelo Rossi, you discuss welfare as increased search costs; and you also point to possible unintended side effects. What do you see as the main trade-off that the DMA created in this specific mapping context?
[00:13:06] Michelangelo Rossi: Yeah, well, On one side, we do agree that having essentially A monopolist, right, like a firm that is extremely dominant in a service may not be good for consumers, right, because their incentive to innovate probably are not so high as they could be with a more competitive frame landscape. We do believe that probably, who knows, in the future, we could actually observe these effects. But on the other side, you have users who are now simply searching more. And these cannot be for increasing. So these are the two trade-offs Let's say on one side, you want variability And you want new Content stability, you want new actors to be there, but to have them there, you need to have extra cost, and these extra costs, at least at the beginning, are paid by the consumers, and we are paying them.
[00:14:05] Daniel Brown: Michelangelo, you've done this research and very important. There have been other researchers that have looked into the consumer responses to the DMA, but not with the example of Google Maps, with similar conclusions where, in other words, many report friction in their online experience while confidence and trust in platform privacy remains low in this case. And so basically it says that consumer behavior has hardly shifted on a handful of dominant players such as Google, Meta, and Apple. And I'm quoting Matthias Bauer, Diuti Pandia. But he's calling for basically Europe considering replacing the DMA with a more flexible, case-by-case enforcement or at the very least reform it by removing obligation that neither enhance consumer welfare nor improve market contestability. Do you have an opinion on that, given your research?
Michelangelo Rossi: I believe that even right now the DMA is flexible. And I believe that the European Commission could have a flexible mind, meaning that they could change, right, the type of requirements for different gatekeepers I hope that, these changes will be informed by data and, for example, by our paper. For example, if I want to focus on the case of Google Maps, we could say that, well, right now we are probably not observing what they will have liked to happen Meaning a clear, contestable markets. And, then they could say, well, it's just a question of time. You have to, this has to stay for, years, who knows? Or probably you have to do something more So let's say I do not want to give precise policy-making suggestions, but I do agree that regulation, that is touching multiple aspects of our online behavior and our online life, of course, requires a constant updating.
Daniel Brown: And it's quite a young legislation, the DMA. So we're still, I think, looking at the ramifications and as the developments in the world of internet evolves and behavior response evolves with it, there's probably a lot of things to revise or to update. Michelangelo, if you had to leave listeners with one takeaway that's common to both papers, one about quality disclosure and one about platform design, is there's one come to mind? What could it be?
Michelangelo Rossi: I will say that, sometimes there are side effects. And these side effects are, in some cases, not anticipated by the institution who is changing - for example, or is providing an information design like the award or, a new policy change. And yeah, probably the importance is not necessarily not to make this type of changes, right, but to keep updating our understanding of these shocks and hopefully, yes, changing the policy or in this case, certifications, the timing, if we do believe that something can improve.
[00:17:49] Daniel Brown: In both papers, you do suggest further research. for you, what are the most important open questions that you're suggesting?
[00:18:35] Michelangelo Rossi: If I could try to elaborate more on what we find, our finding related to the DMA: We could think that Also what we are seeing right now with Google search and AI overviews So somehow a connection, a partial bundling between Google Search and Google Gemini, for example. We don't see it here in France because we are, it is prohibited, it's not happening here in France, but in other European countries, this is still, is already here and also in the United States. We could think that, of course, bundling this type of services could create favorable access to certain type of service, certain type of AI service That could change, someone could say distort the competition for AI, for AI access, and I think this is extremely relevant. And again, we are not talking about this, so this is just me talking about the type of mechanism that we investigated, and we believe that this is present for Google Maps, it is to some extent a service that is not entirely changing our consumption of the internet, but this could also be true for other type of services.
Daniel Brown: Michelangelo Rossi: thank you very much.
Michelangelo Rossi: Thank you.
MIKE:
Michelangelo Rossi, who’s Associate Professor of Marketing at HEC Paris. Michelangelo continues to study the intersection of industrial organization, quantitative marketing, and digitization. For the first paper we discussed on movie-goers and expectations, he and Felix Schleef looked at 25 million ratings in a 24-year period… I wonder what the stats for expectation are for Sinners after its 16 nominations. I had none before I saw it… and I loved it.
Well, that rounds up this month’s Breakthroughs podcast. Hope you enjoyed it. If you have any comments, do send them to brownd@hec.fr, that’s brownd@hec.fr. Goodbye.

2


2