Open Food Facts – episode 1 – Manon Corneille

Interview with Manon Corneille from Open Food Facts

Walid : Welcome to this new episode of Projet Libre. Today, we’re going to talk about a rather fascinating project, I must admit, called Open Food Facts. I first heard about this project because I saw a lecture on a Perl conference day in Paris at La Villette in 2015. There was a conference on Open Food Facts and I must admit that afterwards, I didn’t necessarily follow much. And I was very happy when Manon, who is with me tonight, contacted me, because it allowed me to be able to get back into it a little bit and to be able to prepare, in fact, a series of two episodes. In this first episode we’re going to talk with Manon Corneille today, it’s going to be an episode for the general public. We’re not going to talk technical.
And a second episode behind it will be a more technical episode in which I can address all my geek questions. But the idea tonight is that we have an episode that we can play to people who don’t have any particular technical skills. So as I said today, I’m very happy to be with Manon Corneille, who works on the Open Food Facts project and who contacted me.

Presentation of Manon Corneille

Walid: Manon, listen, welcome to the podcast, thank you so much for being here and taking the time to talk with us about Open Food Facts. First of all, I hope you’re doing well.

Manon : Very good, thank you very much. Thank you Walid for your welcome. We’re excited to be able to talk a little bit about the project. So to introduce myself, my name is Manon Corneille, as you said, I’ve been in charge of partnerships at Open Food Facts for the past two years. I’m working on developing partnerships with food manufacturers and I’m also working on the internationalization part of the project, so we’ll have the opportunity to talk about it maybe a little later. So it’s a pleasure to be here tonight and thank you again for the welcome.

Walid : So the first thing is that I’d like you to explain to us how you discovered free software and the digital commons? How did you get into it actually?

Manon : So a little bit by chance since I don’t have a technical profile at all, these are subjects that are often a little difficult to grasp when you don’t have technical skills. But suddenly, I found myself, well, I rather entered through the door of environmental impact and I was also very interested in everything that has a systemic impact on food. So it’s rather this door that made me discover Open Food Facts and through the Open Food Facts door, I then discovered the world of open data, the world of free software. And by the way, I find it fascinating.

Recently, in the fall, we went to a forum organized by NGI in Belgium. And I was able to meet a lot of free software players. And it really opened my eyes to this super powerful ecosystem. It allows a lot of things. I wasn’t necessarily familiar with that world. And it opened my eyes to all that open data allows and the values it carries in terms of transparency, independence, accessibility, the fact that everyone can contribute, share knowledge and build together.

And there you have it, I think it’s great. So kudos for hosting a podcast about free projects because we need more visibility.

Walid : Absolutely. There’s something you didn’t specify, I think, is when, when did you join Open Food Facts?

Manon : I joined Open Food Facts two years ago, almost two years to be exact.

Walid : Okay.

Manon : That’s it. Before that, I worked for an environmental impact consultancy for cities. So these are urban heat island diagnostics. That’s why I have more of an environmental impact background at the base.

The genesis of Open Food Facts

Walid : Very, very interesting. If we start with the first part, which is the genesis of the project, I would like to understand a little bit how the idea of Open Food Facts was born?

Manon : Good question. Open Food Facts was born in 2012, a little over ten years ago now. So this association was founded by Stéphane Gigandet who is a fascinating person, who leads a lot of projects. At the time, he was running a blog called recettes.de, which is still active by the way, with a fairly active community. On this blog, he lists recipes and in these recipes, he was interested in being able to enrich them with nutritional information as well. So he looked for a database of nutritional information, he couldn’t find it, he decided to create it.
And so little by little, that’s how Open Food Facts was born. Then, gradually, he managed to surround himself with people who are interested in the subject of nutrition and who have contributed to enriching the database as well.
So there you have it, the project was a citizens’ project that lived on about 1000 euros of annual budget until 2017, if I’m not talking nonsense. Over the last 4 or 5 years, the project has really become more professional and there has been the recruitment of employees. Today, there are 8 of us in the team and we are surrounded by a lot of volunteers. So it’s super exciting.

Walid : Did he create it on his own or did other people come quite quickly? When he created the project, did he immediately create the form, which is an association?

Manon : So no, in fact at the beginning he created the project on his own, he created the database on his own. But the association… April 11, 2014. So yes, the association wasn’t created right away. Stéphane started to get his foot in the door in 2012. In fact, after that, he met Pierre Slamich, who is the co-founder of the Open Food Facts association. And so they decided to formalize all this a little bit by creating an association.
And for the record, the association was launched on May 19, 2014, which is the day of the Food Revolution Day, with the mission of trying to make transparency about the food system.

The different sub-projects carried out by the Open Food Facts association

Walid : Before we get into the structure of the project, to give a first overview, what does the Open Food Facts project contain as a sub-project? It’s not just the food database?

Manon : Yes, absolutely, it’s a question that interests more and more people because we’re starting to take an interest, we’ve already been interested for a few years now, in the environmental impact of both food products but also beauty products, everyday products. Consumers are asking themselves more and more questions and to meet this demand, Open Food Facts has a lot of ideas, a lot of verticals. So we have Open Product Facts, which is the database of everyday products. We’re going to be able to find a plastic duck, a plate, a microwave, everything we can buy as a physical object. We will be able to offer consumption advice: for example, we will be able to store instructions for the use of certain objects. Advice can be given if the person is looking to throw away their item or give it away or recycle it. We will try to guide the user so that he can extend the life of his projects. We’re super excited because we’re going to work on it in 2024. It is a project that was funded by the AFNIC Foundation. We have the means on it, so we can’t wait to be at the end of the year to see how it goes.

Then there’s the Open Beauty Facts app, which deals with everything from cosmetics and beauty products to sunscreens, shower gel, etc.

Then, very recently, as a result of inflation and everything that has happened on the issue of prices, we have a small group of members of the Open Food Facts community who have come together, who have decided to create Open Prices, the first open price database. So here, it’s very time-consuming and it’s a project that is extremely large because there are also a lot of stores so it’s a gigantic number of data points. But in just 15 days, we already have data on more than 5000 products, so price data, and that’s only for food at the moment, but who knows, depending on the success of Open Product Facts and Open Beauty Facts, maybe we’ll extend it to other verticals.

The structure of the project / its legal form

Walid : Great. So now let’s talk a little bit about the structure of the project. We mentioned this in the introduction. So the structure that carries the project is an association under the law of 1901.

Manon : Yes.

Walid : I would like to understand at the time why this choice was made to make an association and not, for example, I don’t know, a foundation or a company. What made the founders Stéphane and Pierre choose this legal form?

Manon :

given that it was a participatory citizen project, a little bit on the model of Wikipedia or OpenStreetMap, it made sense to choose an associative status. Then at the time, Stéphane and Pierre were very close to Christian Quest, who is the founder of OpenStreetMap (Editor’s note: France). So they took a little bit of inspiration from how they worked.

Manon Corneille

OpenStreetMap is an association, and that’s how things happened quite naturally, in the end.

Walid : Okay, it’s a French association. Are the employees all French employees?

Manon : Yes, today, anyway. We are all based in France, working from home, and we meet once a month in Paris. And we’re starting to have a new type of role, which is the ambassador role. So we have an ambassador in Germany, an ambassador in Belgium and an ambassador in Croatia. They do not have the status of employees. These are all people who work on a voluntary basis. And then, we have about 300 or 400 volunteers who help us every day to work on the quality of the data, to develop the mobile application, to develop the website, to help us translate all the content we can have, many other things.

Sources of funding for the project

Walid : It works, we’ll come back to it in any case, I’m very interested to see how it all comes together. We have seen the legal structure that supports the project, so at the beginning we understand that in the first years there were no employees, that afterwards there was funding, so then there are employees who arrived, now that there are people like you who are paid by the project. If we’re talking about paying for the project, it leads right to the next chapter, which is a bit of a big chunk, which is the economic model. My first question is, what are your sources of income? That was one of the first questions I asked myself, it’s wow, you do a lot of things but how is it all financed?

Manon : That’s a very good question, especially in projects that are free. I have the impression that this is a topic that comes up a lot, that is difficult to fund. People are very motivated, but there is not necessarily a lot of money. So for us, we are now about 30% funded by subsidies from the public. We will work for example with Santé Publique France, with ADEME, and therefore we will make projects for them. This is funding that is still earmarked. For example, we are currently working with Santé Publique France on the support of the Nutri-Score, the new version of the Nutri-Score. We have a little bit of education, we discuss with manufacturers to make adoption easier.

Then there is another part that comes from philanthropic organizations, so foundations, such as the Google.org Foundation. Two or three years ago we won the Google Impact Challenge, with the sum of one million euros. So it allowed us to put a lot of butter in the spinach for a few years.

We were also supported by the Mozilla Foundation, or players like OVH, which has been funding our servers since the beginning. We also have the Free Foundation, which is on the infrastructure side.
So 30% of the public, about 30% of philanthropic foundations. Then there is another big third that comes from European projects.
We are trying to forge partnerships with different actors in Europe to be involved and participate in these projects. And then we also have a lot of donations from the public. So we do donation campaigns every year to try to raise a little bit of money. That’s pretty much how it’s organized.

Walid : What I understand is that, in fact, you have both funding that is for specific themes, funding that is prizes, etc. which you have as you wish, a funding cushion of about 10% that is from the public. So in the end, you don’t really have any problems being independent and financing yourself.

Manon : For the moment, we have enough funds to see the next two years, but the difficulty, I think, that is present with many free software projects, is that we don’t necessarily have trouble getting funding, but that the projects are always earmarked. So that means we have to deliver, we have deadlines to meet, we have specifications to respect, and that puts a lot of pressure on the whole team. For the year 2024, we’re going to try to prioritize a little bit the search for unearmarked funding, to be able to finance all the maintenance of the common in the end, because we still spend a lot of time interacting with the community, responding to contributors on GitHub who have done some work and that’s great, but it requires a lot of support.
Same on the manufacturer part, we may talk about it again but I spend almost a third of my time responding to requests from manufacturers, finally doing support. All the maintenance and upkeep of this common that we create together is not always financed to the extent we would like.

Walid : Do you also have funding from manufacturers or partners, and do you have other players who help you finance this database?

Manon :

For the time being, we really have a red line, it’s even written in the statutes, we can’t accept funding at all from agri-food players since we’re totally independent, so no one from the agri-food industry on the board of directors or in the finances.

Manon Corneille

The relationship with the Next Generation Internet Fund (NGI) and the NLnet Foundation

Walid : We mentioned it earlier and that’s a subject that interests me a lot and on which I will normally be able to do episodes in the coming months. We talked a little bit about NGI, Next Generation Internet and a little bit about its armed wing, which is the Dutch foundation NLNet. I wanted to know what your report was, did they also fund you or are funding you? Because they fund a lot of free projects. And I’ve already had the opportunity to interview the people from Perturbe, the people from Castopod and more interviews to come as well, people who are getting funded by NLnet. So I wanted to know if that was the case for you too.

Manon : Yes, absolutely. We are very happy that this organization exists and also to see that the resources at home will increase year after year. It’s great to see that Europe is putting the resources behind all this. We started our relationship with NLNet, which we knew well, so there you have it, as they are people who have been funding open source for a long time. So over the last few years, we have benefited from three types of funding. One recently, there, via the NGI program, which will allow us to boost our research on Open Food Facts.

Walid : I’ll interrupt you for a second. So NGI, Next Generation Internet, is one of the funds, one of the European programs for financing free software. It has a new version called NGI Common, but we’ll certainly talk about it later. Excuse me, it was just to clarify.

Manon : Yes, you’re doing well, you’re right.

So, a project was funded last year that we’re going to do this year, which is the development of our research tool. So we’re going to be able to search more easily in Open Food Facts, we’re going to be able to put a lot of filters on the panel on the left. We’re going to be able to have a much more ergonomic search. And then, it’s true that thanks to NLNet and NGI, we’ve also been able to finance two other projects in the past, so we’re really happy that this fund exists. And now we’re also preparing a file for the month of March: another NGI call (Editor’s note: call for projects), this time more focused on Europe-US collaborations, so we’re keeping our fingers crossed, but we’re really happy that this program exists.

What is a digital commons?

Walid : I had just mentioned it before, there is a new program that has been announced, which is called NGI Common and which is a new fund of about twenty million euros I think to finance the digital commons. So my question was, what is your definition of a digital commons, do you consider yourself to be a digital commons and also are you going to be a beneficiary of this fund?

Manon : So are we a digital commons? For me yes, I admit I’m not necessarily an absolute specialist. I consider that, as the Open Food Facts project is open-source, the database is open-source, all the code, all the algorithms are open source. This benefits a lot of players. Today, there are just over 300 mobile apps that reuse our database. There are more than 3 million unique visitors every month who watch Open Food Facts, who consult us for their knowledge. Over the past ten years, there have been more than 600 scientific papers using Open Food Facts. So in that sense, I consider Open Food Facts to be a digital commons. We are regularly solicited by students who are doing PhDs, by scientists, by all kinds of people. They see a lot of value in this database. So my answer is yes, but it’s up to you to tell me if it’s the other way around.

Walid : Listen, I have the impression that when you ask the different players of the digital common headset, it’s not very simple. I had the opportunity at FOSDEM to ask the people at NLnet and the answer was not very obvious either.

Manon : Yes, that’s right.

Walid : Well, the definition is not necessarily very clear. So we can imagine that in the coming years or months, you will benefit from these funds on the digital commons?

Manon : We hope, in any case, that we will follow them closely. We will apply if the stars are aligned and a project can make sense.

Open Food Facts and Citizens

Walid : On this subject of financing, I think we’ve done quite a bit.
I wanted to talk about another one of my favorite topics on the podcast, which is communities. And so, I’d like us to talk about the communities around Open Food Facts. I have seen a certain number, I was going to say, of indifferent personas, and the first is the citizen. I wanted to know how citizens take ownership of the platform, what do they have at their disposal? Can you explain to us how citizens interact with Open Food Facts and how?

Manon : With pleasure. So, we have the classic case, the most common case of the use of Open Food Facts by a citizen. It’s the consumer who goes to the supermarket and asks questions about the products. So at this point, he’s going to be able to use our mobile app that you can download on Android, Apple Store, many others for that matter. And so it’s going to scan the barcode of the food product and from there, on the app, it’s going to see all the information related to the product. So certain information is already displayed on the product such as the list of ingredients, the nutritional tables, but he will be able to see them digitally, which will also allow him to compare several products together in a quick and easy way, which is a little more difficult to do in real life. So we’ve got this app that’s the majority interaction.
Then we also have the website. In this case, it’s more of a home use. If I’m going to ask myself questions about which cereals would be best suited to my diet, I can enter my food preferences in Open Food Facts. So I explain that for me, salt is very important, sugar is very important, I absolutely don’t need sugar in my diet. And after that, I’ll be able to search by category and see the products that are most suitable for my diet, since we also calculate an Open Food Facts score, which is the score of correspondence with my diet. So we’re going to have match percentages. This makes it possible to do research outside the supermarket as well.
The information that consumers are most interested in today is going to be anything that revolves around nutritional information: whether they have specific diets. More and more, we see the case that people are interested in the environmental impact, and therefore in the calculation of ecoscores.

And also, we hear a lot about ultra-processing, additives, emulsifiers. And so here it is rather the Nova index that will interest the consumer.

The Nova Index

Walid : What is the Nova index? Because we’re going to talk a little bit about the Nutri-score later, but I don’t know too much about the Nova index.

Manon. : then it’s the Nova group, more precisely. It’s an index that ranges from 1 to 4. So Nova 1 is raw products, not processed at all. If we take the example of an apple, it’s going to be an apple.
Nova 2, for example, will be an applesauce, so a little more processed, either by the manufacturing process or by the addition of additional ingredients.
Nova 3 is a little more transformed. And Nova 4, even more so: for example, in our case of apples, it could be apple candy, where there are lots of additives, preservatives, etc.

Walid : Earlier, we talked about the fact that Open Food Facts was a bit inspired by OpenStreetMap. I was wondering if there are user groups that do a bit like Open Street Map and go to a place and map. I was wondering if there are groups of citizens who will volunteer to map a certain number of products etc without necessarily going shopping?

Manon : yes it definitely happens so we’re not aware of everything that’s going on, but it’s what we call “scan parties“. So we regularly organize them with the Parisian community, but yes yes it’s done. We get together and we scan and enter data all together. But it’s true that it’s much more fun to do it with others.
There are quite a few researchers who organize this on their own as well, because they are particularly motivated to have the data and to enrich the database. So they surround themselves with their friends who are researchers or not, and they do scan parties.

Walid : The next question is, how many articles are in the database, for example Open FoodFact, and how much of the information entered by citizens does it represent? Is that quantified?

Manon : So today we have a little more than 3 million products on the base. The citizens’ contribution represents about 80%. That’s still huge. And so the remaining 20% is either going to be data that we’re going to recover through win-win partnerships that we’re doing with reusing mobile applications. So they feed on the Open Food Facts database, but in return, they send us fresh data, updated by their users.

And we also work with manufacturers who send us their data by various means.

Walid : That’s a whopping 80%!

Manon : Without the citizens, the project would not exist, that’s clear. They also help us a lot with everything that is data correction, not only do they enter, but in addition all the updates, it’s a painstaking job that is precious.

Walid : I also think that because we touch on food, which is quite fundamental. I don’t know if on cosmetics or on other bases, is the rate the same? Or is food something that really interests everyone?

Manon : I pretty much agree on that. It’s true that we eat every day. Over the course of a year, we have 1000 meals. It’s a number that we like to highlight.

So 1000 opportunities to make better decisions. But it’s true that the subject of food touches a lot. And there you have it, the fact that each of us, individually, we all have specificities, particular needs, specific diets, gluten intolerances. There are a lot of profiles. Pregnant women, for example, people who eat halal, people who have allergies. It’s true that food is a subject that speaks a lot and brings people together, that’s for sure.

Projects or applications that are based on Open Food Facts data

Walid : The second type of actor is projects and applications that reuse your data. So the first thing I was wondering was, what are the types of non-citizen actors that actually interact with the Open FoodFact database?

Manon : We were talking about it a little bit earlier. So there are these famous reusable apps, as they are called. So actually, these are mobile apps that are developed by people like you and me, who are starting their project, a start-up. And so, in order to run their project, they need a food database. So there, there are all kinds of apps, apps to track calories, for example, so more slimming-oriented, apps for pregnant women, which we were talking about earlier, or for people who eat halal, apps that are more fitness-oriented, for sports audiences, apps of all kinds.
Among the most well-known, there are, for example, ScanUp, Foodvisor, Y’a quoi dedans? , which is the application of SystemeU, My Health.

Few people know this too, it’s true that I forgot to mention it, but for example Yuka, which is one of the most well-known applications, was launched thanks to Open Food Facts a few years ago now.
We are delighted to see all these projects see the light of day. We don’t see them as competition at all, because that’s often the reflection we get, doesn’t it bother you to see all these new projects that are maybe sometimes better known than you?
Not at all.
Our mission is to make all this food information accessible to a greater number of people. In fact, these mobile applications are considered to be impact multipliers rather than competitors.

Walid : So these applications have access to the entire database in the end?

Manon : Yes, absolutely. They have access to the entire database and they can also access via our API

Walid : which allows you to interact remotely with the database

Manon : Absolutely.

Walid : If I’m making my own application, what do I need to meet in order to use OpenFootFact data?

Manon :

So, it’s unfortunately difficult for us to keep up with all the projects that reuse OpenFootFact, but in the best of all possible worlds, the rule in any case is to respect the ODBL license, which stands for Open Database License, which stipulates several rules, including attributing the source of the data.

Manon Corneille

So this means that in all the pages where Open Food Facts data is mentioned, they must display the mention Open Food Facts, possibly our logo, a link to our site. Then, the second obligation is that they must share the data as it is, i.e. raw, without having retouched or altered it. So that happens a lot, people download the whole Open Food Facts database and do data cleansing and then republish. So that’s not a problem. This is not compliant with the ODBL license. And then there’s also one last rule, which is that they can’t mix an open database with a database that wouldn’t be under the same license. It is often said that licensing at ODBL is contagious. If we mix these two databases, it will mean that everything must be released under an open license. That’s what you have to respect. After that, we try to hunt a little bit and call people to order when we see applications that don’t comply. But it’s still a lot of work.

Walid : I guess it’s quite time-consuming because you have to detect and then you have to make contact…

Manon : It goes without saying, but I think you can imagine that these projects are not funded, so it’s a matter of maintaining the common and respecting the constraints.

How apps consume and contribute to Open Food Facts

Walid : If I’m an application that uses your data, can I enrich the data? What data can I enrich? We don’t do it like that, I guess we still have to talk to you. What happens if I have an application where I want to enrich your data?

Manon : Yes, there is indeed a contact. So we have this exchange with about twenty applications today. So we set up exchanges, data transfers. It’s still a bit technical and as you can see, I’m not the one who takes care of it. So I don’t have a lot of technical details, but maybe Alex (Editor’s note: Garel) can give you some in the next episode.
But there you have it, so it’s an agreement where the application commits to sending data in return. And it’s on any type of field. So all the information that we can have on the OpenFoodFacts product sheet, reusers may or may not send us data, or not for that matter.

They decide how much and what type of data they want to send back to us. We take everything, in any case.

Walid : Doesn’t that cause moderation problems?

Manon : So, we consider data that comes from a reusing application in the same way as data that comes from a consumer. Therefore, the most recent data is considered to be more important. But it’s interesting that you point this out because, for example, manufacturers have a kind of shield that makes their data more legitimate than data that comes from a mobile application or a consumer. And it’s going to be protected for a while. But that’s not necessarily the case for mobile apps.

Open Food Facts and Researchers

Walid : I wanted to talk about it later, because I watched one of your conferences for professionals and that’s one of the things that I had spotted, that I found very interesting: these were the tools that were made available to them. But we’ll talk about that later.
To finish on this part, we talked about the applications that use your data, but apart from the applications, what other organizations use your data?

Manon : We’ve talked about it a little bit, there’s the whole community of scientists, whether it’s on nutrition or health subjects. There, for example, there are quite a few debates around additives, emulsifiers, etc. There is a paper that was published by Mathilde Touvier from EREN, which makes the link between these additives and their impact on health, particularly on forms of cancer. So there you have it, a particularly interesting case that EREN is conducting.

Walid : Can you explain what EREN is?

Manon : So EREN is the nutritional epidemiology research team. That’s right, excuse me for the shortcut. But suddenly, which is directed by Mathilde Touvier, who is the director. She appears on TV from time to time, she has a little bit of a reputation on all these subjects.
A few years ago, they launched a cohort. In fact they are following, I don’t know the number but I think it’s around 1000 people, they are monitoring their health and they have been following everything they eat for a little over 10 years. And so the idea, it’s something that’s never been done until now, it’s the cohort that’s called NutriNet Health, if you want to go and see, and so they track everything that people eat and their impact on their health. And they regularly use the Open Food Facts database to get information about the nutritional tables, the ingredients of what people are eating as part of the study anyway.

Walid : So that’s scientific, researcher, basically?

Manon : That’s right. We also have a little bit in the spotlight among the cases, there is what was done during the creation of the Nutri-Score. So the research team that worked on the Nutri-Score in 2016 used Open Food Facts to challenge the Nutri-Score formula they had made and therefore check that there are not 95% of the products that come out with Nutri-Score D. So checking that their formula allowed them to have results that were a little bit homogeneous and that really allow you to make a difference within a product category, to see a little bit the different scores.

Open Food Facts and French State Structures

Manon : That’s the scientific part. After that, there are quite a few state agencies that reuse our data. Last year, we did a major project with ADEME which consisted of collecting data on packaging materials, and therefore the packaging data of food products. Thanks to them, we were able to collect data on more than 12,000 products. This new database, which is unique in its kind, ADEME was able to rely on the data collected on packaging to refine their environmental impact index.

So that’s another subject, but as a result, ADEME is currently working on the future official French eco-score which will make it possible to calculate the environmental impact of food products. And among the components of this environmental impact index, there will be a focus on the impact of packaging. And so they worked with the data that we were able to collect to refine their criteria a little bit.

Manon Corneille

Open Food Facts and consumer associations

Walid : Are there any other players?

Manon : So yes, it’s true that Open Food Facts, in the end, we’re in the middle of this whole ecosystem. We have a lot of discussions with consumer associations as well, whether it is theUFC QueChoisir, recently we’ve been working a lot with it, we’ve been talking a lot with FoodWatch Also, so there’s going to be a little bit more about consumer protection and there in particular it’s also about inflation, all the topics related to the shrinkflation (Editor’s note: reduflation) too. Shrinkflation is the fact that manufacturers put less product in the packaging or they have degraded the quality of the products to save money and maximize margins. So Foodwatch is following this closely and is therefore supporting the Open Prices base to conduct the investigation and try to defend consumers as best as possible.

Database data moderation

Walid : I think we’ve done quite a bit of a tour of the different actors. I had questions that had to do with the moderation of the base actually. Because we have a lot of data here, this data is entered by actors who are different. I was wondering, and this is the case for Wikipedia or many others, about the moderation of this data. Who does the moderation? Who’s looking at the errors? Who checks? What have you put in place over the years for that?

Manon : A question that comes up a lot too. We have several levers: the first is going to be the community. That’s really the heart of the project. It’s all about community, we don’t say that often enough. But as a result, we have a team of about fifty people in our Data Quality team, so it’s a team that is dedicated to data quality, who every day will go and correct sheets, make the comparison between photos and data entered verbatim, so there you have it, update products, enter them, correct them.

A second lever that we have and that we have been working on quite intensively for the past few years, well for the last two years, is everything that is going to be related to machine learning and artificial intelligence. So we have developed tools that allow us to detect errors automatically, with also tools that allow us to read the information that is present on the photos, OCR to be exact, and to compare it to the text values. So these are tools that complement and are super useful to the community to then go and make corrections.

After a third lever that we have, it’s going to be our pro platform. So in 2019, it’s been five years now, we developed a platform dedicated to professionals to allow them to import data in bulk into OpenFoodFacts, something that is not possible at all on the public platform and that allows them to do other things, maybe we’ll talk about it later. But as a result, this professional platform makes it possible to make the data more reliable since we have a contribution that comes from the professionals who are at the source of the data. So in general, the data is pretty clean. And then, on top of that, we set up a lot of what we call data quality checks, so data verification points.

So there you have it, check for example on a nutritional table the information is indicated per 100 grams. And so if I add up the amount of all the nutrients per 100 grams and that number is greater than 100, then there’s a problem. And so this kind of logic test, we have a little over 180 of them today. And it really allows you to be able to bring out the mistakes and then correct them. So we have all these tools that exist today. But we must not forget that out of 3 million products, you can’t be perfect.

Moderation remains a big challenge. It’s never going to be exhaustive, but we’re still doing our best.

The work around the Nutri-score

Walid : In the end, it’s the same moderation issues as data moderation everywhere, but it’s interesting to see what you’ve put in place.
One of the little asides I wanted to make, it was what I think is that it’s a little bit about it but I wanted to see if you had any more things to add, it was on the Nutri-Score, because I think it’s one of the most visible things, at least I don’t know very well about Open Food Facts, it was a little bit of a review of the work of the Nutri-Score. That work, etc. Did this work make you stand out?

Manon : Yes, definitely. Especially since the Nutri-Score today is calculated in 8 European countries outside France, so it is true that it allowed us to shine even more. So, what I would eventually have to add is the effort we made and the impact we had on the democratization of the Nutri-Score because it’s true that there was a first collaboration on the challenge part of the formula based on data.

Walid : What year was the first work?

Manon : So that was in 2016. It’s going to be almost ten years, it’s passing. And so after that, in fact, we integrated the Nutri-Score formula into Open Food Facts, which makes it whether the manufacturers like it or not, we calculate the Nutri-Score and whether it is visible or not on the package, we make it accessible to consumers through our application and our website. We are convinced that this had an impact on democratization since our application was already used quite a bit and it allowed users to familiarize themselves with the index and to appropriate it little by little.

Walid : Have you seen any direct contributions since this work with Nutri-Score?

Manon : So we’re aware that there are some, I don’t have the figures, but there is a certain part of the community, well, users in any case, who use OpenFoodFacts only for the calculation of the Nutri-Score. So it’s definitely important for the development of OpenFoodFacts. By the way, there is often confusion that OpenFoodFacts created the Nutri-Score, which is not the case, and I want to make it clear. We have been more of a vector of democratization, but we are not at the origin of the Nutri-Score.

But we’re very proud of it and it’s useful for the development of Open Food Facts, that’s for sure.

Open Food Facts and Manufacturers

Walid : The next big piece is what I’ve called Open Food Facts and its environment, and the first one is the industry. What I’m really interested in is understanding what the industries are interested in working with you. In 2012, the base was created, in 2014, there was the association. Initially, you were largely unknown. What actually made the industry work with you? What’s the point of it?

Manon : Actually, at the beginning, it was born in our interest, since we were interested in their data. And so, in order to convince them, we had to develop a tool that was useful enough for them so that they didn’t have to worry about whether or not to join Open Food Facts. In fact, given that there was already a lot of data present on food products, so data that was entered by consumers, the interest for the manufacturer is already to correct this data that already exists, because in general they don’t like it too much when they see that their product has been photographed by a consumer with a light that is certainly not ideal, with the photo of the hand holding the product.
So it’s not really marketing. So the primary motivation will be to correct the data and add photos that are clean.
Another lever, and still it’s a little minimal, but it’s the fact of highlighting this transparency approach. CSR (Editor’s note: Corporate Social Responsibility) matters more and more, brands like to highlight when they do things for the planet, for good, for society. So there are some companies that are surfing on that a little bit. In general, it’s quite easy to convince manufacturers, given that it’s free to join the platform. It allows them to keep up to date and make sure they have their own data. And it’s an approach that is for the common good. So in general, they put in the resources and they contribute.

Walid : If I’m a professional, in addition to correcting my data, what else do I get out of it?

Manon : yes, so good question. We have this tool that allows you to import data in bulk, so manufacturers can import an Excel file for example, with all the data in columns in a raw file. But beyond that, we have tools that allow manufacturers to have opportunities to reformulate their products. For example, we will give them advice on how to improve their Nutri-Score. So we’re going to be able to identify by analyzing the list of ingredients and the nutritional tables, we’re going to push them all the products for which, by making a minimal adjustment on the amount of salt, fat, sugar, all these products for which a minimal adjustment will allow them to change the Nutri-Score. So switch from Nutri-Score B to Nutri-Score A. And so that’s an analysis that they’re often fond of.

Walid : The question I ask myself is, is it useful for big industrialists? Because I imagined, like that, that it could be used by small manufacturers potentially. But is it also useful for big manufacturers?

Manon : Oh yes, yes, yes! It’s useful for big guys too, yes yes. Then in general they have tools on their side as well that are certainly pushed, but I have examples where really manufacturers of all sizes can use these tools. And we also have a tool that allows them to compare the nutritional qualities of their products with all the other products we have in the base, so products in the same category.

For example, if I have chocolate cookies in my product portfolio. We’re going to compare my chocolate cookie with all the chocolate cookies from Open Food Facts. And I’m going to be able to say, Manon, your cookie, there, it’s 15% sweeter than the average cookie on Open Food Facts.

Manon Corneille

Walid : Okay. It’s quite interesting. While doing some research, I came across a platform, but which seems to no longer necessarily exist, which was called Numalim. And actually, the question I was asking myself was, do you have competitors? I feel like Open Food Facts is unstoppable. That is to say, it is an association that has no bias. Being a competitor with Open Food Facts seems complicated.

Manon : Actually, no. In any case, I have never come across a similar project, which is an association and which is collaborative, where consumers can participate and enrich. We haven’t seen any others, no.
Afterwards, we met a type of actor, but in this case, it’s private actors. For example, there is NIQ Brandbank. I don’t know if you’ve heard of it, maybe not, because it’s a bit niche. But basically, it’s a player whose job, in fact, is to digitize the information of food products. So, the manufacturers will send them the products and after them they have a very professional process which consists of making very nice photos of the product, entering all the product information, marketing descriptions too.

So their job is to showcase their customers’ products and so they have a bit of a foothold because they’ve been around for a long time. So they also have a big database that sells very expensively and in fact we can’t even consider them as competitors, since we don’t bring the same value to the world at all. In quotation marks, they are very customer-oriented. What they do is high-quality data, but it’s not the same price.

Open Food Facts around the world

Walid : The question I’m asking myself, because we haven’t touched on it too much, is how many countries do you have data in? In how many countries? Because there, we talked a lot about France, a little about Europe, but in how many countries?

Manon : Today, we are present in nearly 200 countries, so almost all countries. Let’s say that we have the beginnings of a database in nearly 200 countries. After that, if we look at the number of countries in which we have more than 50,000 products, I think we must be around 30 countries. So we’re mainly in Europe, the US, South America a little bit too.

Walid : It’s crazy.

Manon : But we’re always surprised, it’s quite incredible. Sometimes we receive emails from people in Venezuela who explain that they wanted to enter data but that they are blocked, they have a bug. They want to contribute and we’re always amazed by the scope of the project, it’s quite impressive.

Open Food Facts and the State

Walid : Great, I’m picking up my thread. The second actor in the Open Food Facts environment is the State. Initially, how did the state look at you and now how do you collaborate with it?

Manon : I don’t have all the history, I’ve only been here for two years, but I know anyway and I notice that we are seen more and more as a credible player and almost unavoidable in fact, because we are being asked more and more about European projects. We have established fairly solid relationships with Santé Public France, which has been supporting us for five years now, with whom we are working in particular on the evolution of the Nutri-Score, and ADEME, with whom we worked last year. We also talk a lot with the General Directorate of Health and the Ministry of Agriculture. There is a sense that there is growing confidence in all these state agencies, which I hope is right. We’ve proven it with the number of products we’ve been able to build and the whole community we have today. But we feel that they are increasingly recognising the usefulness of the Open Food Facts database and in particular the tools of the digital commons.
We were talking about that a little bit earlier. But I think there’s a real awareness at their level. It’s nice that they also communicate the fact that they use these kinds of tools. yes, that’s great. But it’s true that it’s something that has really been done over the years and that hasn’t been done in a snap of the fingers.

Is it an asset to be a French association?

Walid : Does the fact that it’s a French association play a role in one way or another? Maybe if it had been done elsewhere, it wouldn’t have been the same legal form and it wouldn’t necessarily have been the same independence, I don’t know. Does the fact that it’s in France play a role, so already certainly with relations with the state, I suppose, but does it bring additional things?
That’s a question I’m asking myself when I talk about it with you.

Manon : Yes, so it’s true that on a European scale, I think that Open Food Facts was finally very well placed by being in France because already on the issue of food, the French are more involved than average, it seems to me in any case.

When it comes to food, the French cook a lot, are interested in their health, more than for example a British public where they eat a lot of ultra-processed products and there is a little less debate around scores, etc. It’s also the country where the Nutri-Score was born, so obviously I think it plays a role. And a little more recently, on the issue of environmental impact, France is really a country that is a leader in Europe on all these subjects. So for sure I think it’s fertile ground to be able to develop this initiative. We are delighted to have had the outstretched hand of all these actors of the State and we hope to replicate this success in other countries. We are in discussions with some ministries of health in Spain, Germany and Ireland. We hope to replicate that a little bit and be able to bring them the value that we bring to Santé Public France and establish ourselves there too.

Solicitations for other databases?

Walid : Do you have requests to make databases out of food?

Manon : It’s a little bit like what happened with the project on food packaging with ADEME, it came from them. Otherwise, no more than that, apart from the case of Open Product Facts, which is a hot topic and was funded by the AFNIC Foundation.

And otherwise, things happened quite naturally for Open Beauty Facts. And I forgot, we also have Open Pet Food Facts for animal feed.

Relations with Europe

Walid : My last question on this subject is relations with Europe. We’ve just talked a little bit about your relations in Europe. Who were you in a relationship with? What were the reactions of other countries to Open Food Facts?

Manon : In fact, not all countries react in the same way when it comes to food. For example, in England, there’s a lot of talk about everything that’s ultra-processed at the moment and consumers are very interested in that. In Italy, for example, they are putting on the brakes compared to the Nutri-Score. So as soon as they understand that we are working with the Nutri-Score, the discussion is usually complicated. Because the Nutri-Score, for the record, doesn’t highlight all their local products, such as olive oil, ham, cheese, etc. So Italy has a little bit of a difficult relationship with this index. But suddenly, our relations, we discuss with a lot of actors, we try to replicate a little what made the success of Open Food Fact in France. We are talking to government departments, researchers and manufacturers as well. We try to do as many trade shows as possible to meet the manufacturers.

We were in Italy last year, and in Germany as well. The difficulty is to have a database size large enough to be credible, both with consumers who will find the information they want, because if you don’t have a lot of products, people won’t use the application. If you don’t have a lot of users, manufacturers won’t want to send their data. So there’s actually this question of the threshold of sufficient products to be met to make Open Food Facts take off in other countries. For international development, at least, we talk to all possible actors, so manufacturers, researchers, ministries, consumers, and reusable applications as well. For the moment, the priority, at least for this year, is rather on the countries of Germany, Italy, Spain and the United Kingdom, so the neighbouring products, where it is easiest to trade. But Paris wasn’t made for another day. And Open Food Facts from all countries won’t be done another day either. We are also counting on the development of the ambassador role to have local actors who can engage the community and develop the project like this.

Future challenges for the association

Walid : We’re going to move on to the last part, which is a bit of a conclusion. We’re going to talk a little bit about the challenges ahead. So, I had noted some technical challenges, but we’ll talk about that in episode 2. I wanted to know what the relationship challenges were in terms of organization, etc. ?

Manon : A challenge that is non-technical is going to be the commitment of the community, to be able to structure the effort of all these volunteers who want to get involved in the project. We tried to do some work on team organization. There’s the communication team, the partnership team, so with profiles that tend to want to meet people through email reminders, etc.

Financing teams, a data quality team and more technical teams as well. So we tried to divide the effort into different poles and for each team. We have a person employed by Open Food Facts who manages all of this and who tries to animate the community of volunteers. We have a few volunteers in each of these teams. It’s still difficult and it still takes a lot of time. It is an effort that is not funded today. So we talked about that at the beginning of the episode, but it’s still a challenge. Keep contributors motivated.

Walid : It’s always a challenge in free projects to keep contributors because we’ve trained them, they have the experience and we don’t want them to leave.

Manon : yes, it’s also difficult to estimate the progress of a project because you don’t know how long people will be involved, you don’t know. It’s hard to plan when you’re volunteering because you can’t demand to do things either. And to also be able to keep the documentation up to date because there are a lot of things, a lot of projects, even within Open Food Facts. And so for all the newcomers who are new to the project, in general, they pull their hair out a little bit because it’s quite bushy. There’s also something you can count into the challenges.

Final Words

Walid : Yes, very interesting. I think that if we talk again next time, we will talk about this relationship between volunteers and employees. I think there’s a lot to say, but unfortunately, we’ll have to take more time. Listen, we’re coming to the end. I would like to leave you with a final word. Do you want to convey a message to the listeners of Projet Libre?

Manon : My message is to be careful what you eat. Take care of yourself. Let us not forget that we are all on this earth together. So all the projects that involve the community, that contribute to the common good, we must not lose sight of them, even if we are all caught up in our daily lives and by our smartphones and all these notifications etc. Take care of yourself, engage in open source projects, that’s what will save us.

Walid : Thank you very much. It’s a nice conclusion. Maybe a little ambitious, but… When you look at how far we’ve come since the creation of Open Food Facts, I’m not sure Stéphane and Pierre thought that one day it would do all this. So you also have to dream, I think, it’s important.
I’m also going to thank you for contacting me. I think I’m also going to thank Pouhiou from Framasoft who talked about me.

Manon : Definitely.

Walid : He’s one of my volunteer sponsors! There you go. Listen, thank you very much. It was exciting.

So we’re going to meet again soon for a second episode which will be more technical in which I think I’ll be able to ask all the questions that I didn’t have time to ask here. Thank you very much Manon. And then, listen, see you soon. I hope we’ll have the opportunity to talk again about all your projects, you’re in them, etc. in the future.

Manon : With great pleasure! A big thank you to you Walid, it was super nice.

Good luck for the future and see you soon!

This episode was recorded on February 14, 2024.

To go further around OpenFoodFacts

• OpenFoodFacts, the wikipedia of food products, Anca Luca, FOSDEM 2023
• Presentation of the Open Food Facts project and its prospects Pro Days 2023
• Join the association as a volunteer
• Support the association
• The Open Food Facts Blog
• Open Food Facts’ Slack
• Open Food Facts’ LinkedIn page

License

This podcast is published under the dual license Art Libre 1.3 or later – CC BY-SA 2.0 or later.