During the week of the 23rd to the 27th of April 2018, I went to my very first conference. And I didn’t know what I will do/learn there. This conference was the Web Conference 2018 (WWW2018) in Lyon. More than 2,000 people were at this conference from more than 60 countries. I have to admit that I expected to see only researchers. However, that was not the case. Indeed, many people from the industry were here. To cite a few: Wikimedia Foundation, Amazon, Google, Facebook, etc. And Lyon is a beautiful city, a lot more green than I expected. So, if you happen to be/go in France, you should definitely visit this city. However, that’s not the point of this post. Indeed, I want to tell you what I learned during this fantastic week.

WWW2018 banner

So, I’ve been invited to this conference by the Data Science Laboratory at EPFL, the dlab, because we wrote an article with Professor Bob West. The whole story about this article will be portrayed in a future blog post. So, my primary goal at this conference was to present this paper. I also wanted to go to some talk, but that was not the most important thing for me because I’m doing my PhD in another field. So, what did I learn? I think I can summarize it in two points that I will elaborate a bit later:

  • Most of the conference happens outside of the conference.
  • Researchers have no idea how to present their work.

Let me first elaborate on the second point, just because it may be more interesting for you if you read this post thanks to the subtitle. At this conference, I have seen around 20 talks about different papers on many different subjects. And on all of these talks, only 2-3 were good. And, I, personally, think that this is a shame. What are the main issues with these talks? I will summarize them in two bullets points that I will cover in details.

  • Speaking very fast with a strong accent
  • Only showing technical details

We, researchers, are from all over the world. It is therefore reasonable that we don’t speak English perfectly. However, I’ve seen some people with quite a strong accent speaking at a slow pace, and even slower when detailing some complicated parts of their paper. And I could understand what they were saying. I may not get all the technical details, but I could follow their talk. However, when someone talks with a strong accent very quickly, my brain wasn’t able to translate what I was hearing. I know that improving our accent takes years of training. I, for example, still have a French accent. And I’m pretty sure I will never lose it. However, I try to speak slowly and to make some breaks when I’m presenting something. Why do I do this? People need some time to process what you’re saying. Indeed, speaking slowly is the best thing you can do. First, you will feel less stressed. Don’t you trust me? Try it next time you’ll present something! And making some breaks is not only good for your audience but also for you. Indeed, it will give you a bit of time to think about what you will say next. When I explain this to my students, they often tell me that if they do this, they won’t have time to present all of their slides. I usually have a straightforward answer to give them: Remove some slides! Indeed, the golden rule is the following one: You have to spend one minute per slide. In other words, for a talk of n minutes, you should have around n slides. So, next time you’ll make a presentation, stick to this rule and try to speak slowly. In the end, we always end up speaking faster than we think. So, you will never talk too slowly. In the end, you will see that people will most probably be able to follow your presentation this time.

Now, let’s move to the second point: Only showing technical details. Let me tell you one thing. You should never do this! If people are interested in your research, they will read your paper. You cannot give a lot of details in a 20 minutes presentation. Thus, it is much more critical to catch the attention of your audience about your research. And it is pretty hard to achieve this. Fortunately for you, dear reader, I’ve made the same mistakes. So, let me take you through the journey of creating my presentation. But first, let me write a few words about my research. (A future blog post will follow on the whole path from the beginning until this final article.) I did my Master Thesis with Professor Bob West in the dlab at EPFL. My goal was to study thoroughly two datasets about beer reviews. In the end, we decided to write a paper about a proper methodology on how to study herding effects: When Sheep Shop: Measuring Herding Effects in Product Ratings with Natural Experiments.

Presentation v1, v2, v3, …

Let’s go back to the creation of the presentation. The first presentation was following to the paper very tightly. I was trying to give most of the details about the methodology. And as you can see, it contains all of these details. But the main issue was that it was missing the motivation about the problem we are trying to solve. Indeed, I was not explaining why herding effects are challenging to study and why we need such methodology. These were the comments I received after presenting it to the dlab. Therefore, I decided to start again. For the second version I took a complete 180 turn. As you can see, I started to create this story about the bakery to introduce the concept of herding. It was Bob’s idea. And it is a brilliant idea. Indeed, it is known that the first 2-3 minutes are the most important in a talk. Within this time frame, you’ll either lose everybody, and they will open their laptops and start answering their emails. (Too many emails!) However, if you can catch their attention at the beginning of your talk, you will most probably keep them until the end. Yet, this presentation was far from being perfect. Indeed, if you go through this presentation, you can see that I’m jumping back and forth between different topics. For example, I’m already talking about the way to study herding effects on slide 5, and I’m talking about this again later in the slides. Thus, I lacked a clear guideline that the presentation had to follow. (It was closer to a roller coaster than a line for this presentation.) So, after the second presentation to the dlab, I decided that I will write down, using bullet points, the topic of each slide. Then, to make it even clearer in my mind, I decided to write the text I wanted to say. It didn’t matter if I would follow it during my talk. I just needed to do this to avoid this roller coaster behavior. Thus, the third presentation was much better (and funnier) than the second. You can easily follow the presentation using the presenter note which can be interesting for you. And if you have some time and read the paper, you will see that this presentation is completely different than the article. Indeed, I do not give any details. I’m here to tell you that there is a problem and I know a way to study it.

Some advice

Let me give you some advice (that I previously received from Bob) to help you create an amazing presentation:

  • Do not present all the details of your paper. You’re here to explain your problem. Your final goal is to push people to read your article.
  • Write down some bullet points to develop the structure of your presentation. It can also be useful to write the text of your presentation even if you won’t read it.
  • Use one minute per slide. You don’t have to rush. Take your time and breathe.
  • have fun!

The fun part

The second thing I learned during this conference is the fact that Most of the conference happens outside of the conference. To be honest, one Austrian researcher told me this at the very beginning of the conference. And I have to admit that I was amazed how true this statement is. In the conference, I didn’t talk to many researchers, mostly because most of them wanted to see some bad presentations all the time. However, it’s when you take the time to discuss research and different topics with a beer (or a non-alcoholic beverage) that you can share your thought with inspiring people. I’ve talked to many researchers this way, and it was terrific because I learned so much from these people. Thus, I only have one advice: If you go to a conference, enjoy the social moments as much as possible. You will definitely learn much more than going to all these bad presentations.