Measuring Web Traffic – Matomo – M.Aubry & L.Destailleur

From AWStats to Matomo, with Laurent Destailleur and Matthieu Aubry

Walid : here is a new episode of Projets Libres!. Today, in this episode, we’re going to talk about a subject which is the measurement of traffic. And to do this, I have two guests to talk about it with me. You’ll see, it’s going to be very interesting, their two different points of view. The first is Laurent Destailleur who has already intervened. In fact, he was the first guest on the podcast and with him, we talked at the time about AWStats and Dolibarr. We will certainly talk about AWStats again today. The second guest is Matthieu Aubry who is the CEO of Matomo which I will let afterwards introduce myself. You will see why these two people are with me today. So Laurent and Matthieu, welcome to the Free Projects podcast! And thank you very much for taking your time to be able to talk about this subject with me.

Laurent : It’s a pleasure.

Matthieu : Thank you for inviting us, Walid. Nice to meet you.

Presentation of the guests

Walid : Well, there you go, nice to meet you too. To begin, I’m going to ask you both to introduce yourselves. Laurent, this is going to be a fairly quick presentation, since for people who are interested, I suggest you listen to episode 1 of season 1. I’ll let you, Laurent, introduce yourself very succinctly, please.

Laurent : To be successful, let’s say that I’m a serial open sourcer. That is to say, I develop open source solutions and I make a living from them today. Among the most well-known is AWStats Log Analyzer, which we’re going to talk about today. Today, I focus more on Dolibarr ERP CRM, for which I am the project manager. And that’s it, but if not, there are other projects, Sell your SaaS, DoliDroid, Dolimed, there are a lot of other things. And I am the founder of the company Dolicloud, which is my professional activity. which is a Dolibarr Cloud hosting solution.

Walid : That’s it, thank you Laurent. Matthieu, can you please introduce yourself?

Matthieu : It’s Matthieu Aubry. I was born in Poitou, from a peasant family. Very early on, I was interested in technology. I didn’t know at the time, but I think I was an entrepreneur very early on, although it wasn’t really in business, but I was undertaking a lot of projects with friends, projects on the side of the school. Over the years, it has evolved into the creation of several open source projects, including the one that currently makes me a living and whose company I manage. I currently live in New Zealand, where I moved about 13 years ago, after meeting my partner who is New Zealand. We met in Scotland when I was working for Amazon. I am lucky enough to return to France every year. I was able to meet Laurent this year, we spent a very good evening together, a summit with only founders, co-founders of open source companies. So it was very nice to be together with one of our tribes there.

The different traffic analysis technologies

Walid : For listeners who want to know more about this Open Source Founders Summit, I suggest you listen to the episode with Emily Omier, who is one of the organizers. There you go, we’re going back to that a bit, so there you go. It’s great because it was one of my questions which was why you’re in New Zealand, so you answered. To get to the heart of the matter, the first thing I’d like us to do is to take a little step back and explain to us what the different traffic analysis technologies are, how do we do it historically and now, how does it look to analyze traffic on sites?

Laurent : I’ll start with the oldest one, the one that’s in effect in AWStats Log Analyzer. So it’s the log analysis technology, so it’s simple, it’s a small program that will read the server logs. So it could be a web server, but it could be a mail server, an FTP server, anything. And then we’re going to try to get as much information as possible and then synthesize it all to represent it in the form of a graph. So that’s it, let’s say the ancestral technology, the oldest, the one on which we had the historical traffic analysis tools, so AWStats, but we also knew more, very much at the time, Analog, Webalizer, and then many other solutions, also proprietary, that could work in this mode. And then, another way of doing things came along, another technology, which I’ll let Matthieu talk about, which is, let’s say, the most common one today, in the insertion of Javascript tags.

Matthieu : That’s right, in fact, with the evolution of the web and the evolution of javascript in particular, it has been possible to do audience measurement in another way which has certain advantages over log analysis. So it’s a technique that is the technique used by everyone now where in fact you just have to insert a small javascript code in all the html pages of the site and this javascript code will run in the browser when you visit the website. It will collect information and send it via an HTTPS request to a service that will then store this data in a database to then calculate reports and show dashboards. The advantage of this so-called tag-based audience measurement solution is that you don’t need to have access to the server log, which can be quite complicated to set up. It’s easier to copy and paste a code into a site. And also, another advantage is that with Javascript code, we are more likely to track real people, humans, who visit the website, whereas the server logs will include all kinds of data, including bots, so robots that browse the site. And it can be quite hard, sometimes, to exclude robots, differentiated from humans.

Laurent : We can say that there are other advantages as well, which is that we are also able to extract information that is browser-specific information such as screen resolution, this kind of thing that we can’t do with log analysis. On the other hand, there is also a disadvantage, which is that there are indeed certain types of traffic that we will not see, in particular worms, a kind of robots or viruses that can remain invisible by the analysis of tags since they will not execute the web page as such, they will just make raw requests, Something that the log analyzer will see. And then there’s the hybrid mode, for example, a la AWStats, trying to join the two worlds a little bit by putting a little bit of tags to try to collect information that tag technology can’t collect. So here, we’re going to be more of a mixed solution.

Walid : From what I understand, it would potentially still be relevant to have server logs. Yet, now, everyone uses Javascript tags instead.

Laurent : yes, because by the way, today… The needs of analysis are now mainly transferred to users, administrators in the business sense, rather than administrators in the system sense. And so, for these administrators in the business sense, the log analysis information, the added value that the log analysis had, these are things that, for them, don’t speak. And so, there is no real meaning for these people anymore. On the other hand, indeed, for those who really want a more system-oriented vision, log analysis is the most interesting, but we will see that there are other solutions, this interest is mainly oriented in relation to the virus, that kind of thing, and there we will have tools that are a little more generic, more security-oriented, which will do this kind of reporting for system administrators. And so it’s true that today log analysis, on the server side, is falling into disuse due to the evolution of today’s needs and user profits.

Matthieu : When Laurent talks about business experts, there are several categories of people who use audience measurement tools. There are marketers in particular, these are people whose job it is to bring audiences to sites and qualified people who are interested in the service. Marketers are a large part of the users, but not only, there are also data experts, people who answer any data question from their colleagues, from different teams and who need a tool. flexible that allows you to formulate queries on the data and to have dashboards, etc. Another aspect is what we used to call webmasters, who have now become more product experts. And so, we have titles like the Product Manager, the Product Owner. And so, these are the kind of people who really use audience measurement tools.

The genesis of Matomo

Walid : We’re not going to write the history of AWStats because we’ve already done it. On the other hand, I would be very interested if we started the story of Matomo. So since its genesis, Matthieu, could you explain to us a little bit why you started working on this project and also what was there at the time and why you decided to create a new product in fact?

Matthieu : Actually, until I was 16, I was a classic teenager. I liked school, I loved learning things, reading, etc. But hey, I played hard, I video, I play sports, etc. And then, I had the chance to enter a rather special high school called the Lycée Pilote Innovant, which is now called the Lycée Pilote Innovant International. And really, it was a chance and a big tipping point in my life. In fact, it was a high school that accepted people on motivation. You had to write a cover letter, you still had to be a little academic, etc. And so, there, I found myself with really many people who had a kind of passion in them, who wanted to do things. In particular, I lived… In fact, we were at boarding school, so we did. And at the boarding school, I was in the room with Johan Mathe, who is now co-founder and CTO of a company that makes artificial intelligence for weather prediction. And he worked at Google for years, Google X, etc., where he worked on atmospheric balloons that transferred the Internet. In short, it turns out that at 16, I was in the same situation as him. And in fact, he introduced me. He was more advanced, even more precocious. And as a result, he had already started programming.

And on top of that, in high school, we already had programming classes, you can imagine. That was in 2000. And we learned Pascal, two hours a week. We even had our computer teacher who was a seasoned Linux administrator, a seasoned librist. So really, it was an incredible opportunity. And in fact, I was introduced to Pascal like that, but really, I didn’t like it at all. It was like, I couldn’t connect to Pascal, we were doing drawings and everything. But on the other hand, as soon as I discovered the web, it really unblocked. And so, Johan introduced me to PHP and later to MySQL.

Matthieu Aubry

And actually, what happened was that at the time, in high school, you have to make presentations, things like that. Well, the classic method is to make PowerPoint. But we, in fact, made websites instead. Instead of making a PowerPoint, we made a website. And so, it was the time, it was super fun to make sites. Pixel Perfect Design, if you remember back then, it was a different style. But in fact, after spending so much time making these sites, these presentations, instead of letting them die on our computers and never see them again, we enjoyed putting them on the Internet and sharing our work like that. And then after doing this several times, it’s very natural to say to yourself “are there people who are going to see these sites?”. And so, that’s how it started. And actually, of course, I didn’t have the idea directly to create a tool to do that. First, I looked at the existing tools. So, there was of course AWStats which required to set up with the log servers, but it’s a bit complicated to set up. The alternative at the time was Xiti. This is for those who remember. Now it’s called Piano Analytics. They were bought by Americans, but it’s a company based in Bordeaux. And in fact, it was a very good service. When we used Xiti, we had to put this famous JavaScript code in the pages. And actually, what was happening was that it had an icon and a logo of Xiti, which at the time was yellow and actually very ugly as well. In our Pixel Perfect website designs, we really put some love into it and everything, we didn’t want to spoil it with this yellow logo. So actually, that’s really where it started. Since I was starting to take PHP and the database, in fact, I quickly connected that it was a fun project to build, just like that, in my free time, at the boarding school, in the evening, and that’s how we started. And at the beginning, I didn’t pretend to make a free project, although, in fact, that’s where I also entered the field of free software at the same time. In fact, when you’re curious and you want to learn things quickly, well, you’re really grateful to the people before you, who, after doing all this learning, decided to share it with the community. And really very quickly, I was fascinated by open source and free software. And then I discovered Richard Stallman, a bit like the philosopher, the creator of the free software movement. I was still very idealistic – no, I still am, but much more moderate with age and learning – . But it’s true that it spoke to me a lot directly. And even at the time, I had done an essay on free software to explore it in a little more detail, to understand it well. When I discovered, on the one hand, computers, the Internet, and on the other hand, open source software, and having this need to measure the audience of sites, a little naturally, I said to myself, “that’s perfect, we have to make an open source tool for that”. And in fact, it turns out that it didn’t exist at the time. It was really a chance to be there.

Walid : I asked you, at the time, weren’t there tools like that, more on the client side, that made it possible to measure audiences?

Matthieu : I remember, no, it didn’t exist. There were tools like AWStats, Analog, etc. But there were no open source tools. It was really the beginning and as a result, it was commercial tools like Xiti. So that, I often say when people ask me, when I tell the story, it’s really, of course, it’s a lot of work, focus, trials, learning, but it’s also the chance to have had the timing, to have had the idea at the right time and to have focused on this idea, to have been the first.

Walid : Knowing that Laurent, he was French, did you, in one way or another, try to contact him?

Matthieu : I was just looking through the emails, but…

Laurent : So I have proof, be careful.

Matthieu : From 2012, did you find the email from 2012?

Laurent : I found an email from you dated June 18, 2012.

Matthieu : Yes, that’s it. I found him too. But suddenly, it was already ten years after what we are talking about here. So, it was quite late, it came.

Laurent : So there, you were no longer called phpMyVisites, you were called Piwik.

Matthieu : That’s it. So, the story is a bit long and there are quite a few stages. But it’s true that when I contacted Laurent, we had the Piwik project, which is now called Matomo, which basically uses Javascript technology for audience measurement. But it turns out that in 2012, so it was a few years after the creation of the project, there was a community member who had a need to analyze server logs and who participated in the open source project and contributed to an additional module that allows the analysis of server logs. That’s where we came into competition with AWStats. As you say, since Laurent was French, and between open source projects, we have respect, I contacted Laurent and I told him “let’s launch this. It’s true that now, we’re going to be competing.” I just wanted us to make contact. I don’t know if we spoke at the time.

Laurent : Honestly, I don’t remember in the history of the exchanges. We had planned a video call, but I don’t remember the discussions we must have had.

Matthieu : That’s it, it’s going up.

Walid : So at the very beginning, when you created the project, it was called phpMyVisites.

Matthieu : yes, so actually, when I was in high school, I created this phpMyVisites project. So at the time, the name was terrible, let’s say, but it came from phpMyAdmin, which was the tool to administer MySQL databases in PHP. And well, I was maybe 17 years old, I wasn’t too inspired on the names, so I had copied the name. phpMyVisites. In fact, that was my first open source project. And in fact, it was really a huge learning curve. In fact, I made a lot of various technical mistakes, also in the way I managed the community. In fact, I was a little too closed in on myself. I hadn’t yet figured out how the community worked and how to communicate to the world so that the mission, the vision was clear and people wanted to join. But in fact, it wasn’t clear enough. And so, it was kind of my project: I was doing my thing. Yes, there were some contributions. In particular, it was translated into 25 languages, I believe, etc. But it wasn’t a developer success. And then afterwards, I went to engineering school. So, I still had in prep on top of that, it was very hard for me to prep. I didn’t have time to work on it anymore. But I still had this project that worked. And then, always, there had been no other alternative that had been created since. So, I had my project, I saw all the problems there were and I didn’t like that project anymore.

In fact, when I was in my third year of engineering school, I was lucky enough to have been noticed thanks to this phpMyVisites project. There was an American who had noticed me and in fact, who offered me to come for an internship for four months because I had to do a 3-4 month internship in the third year. And so, I was invited to do an internship in London in a company called OpenAds at the time, which was in fact an open source advertising server, so a kind of decentralized alternative to Google Ads, which allows everyone to manage their ads on their site, to have control over it. And they, in fact, thought that of course, analytics was important in this concept. And then, it didn’t cost them much to invite me for an internship for 3-4 months. They had nothing to lose. And for me, it was a huge opportunity. In fact, I arrived in London. I was 20, 21, really full of energy. And as a result, I had carte blanche, in fact. And really, I remember being on the desk in London. It was a big table. I had a lot of blank sheets, for once. And in fact, I had to redesign the project with all the learnings I had, the mistakes I had made before so as not to do them again.

So, at all levels, anything at the level of the database, the design or the code, make a modular code or the website, have a very clear mission, engage the community, be welcoming, etc. So that’s where, in London, I created the project. So we had to find a name. And there, I had the inspiration. In various regions, the word wiki had inspired me, the word kiwi too. It’s funny because that was before I met my New Zealand partner, who is called the kiwis. So maybe it was a sign of fate too. But in any case, I called the Piwik project. After about six months, I had an alpha ready. And with this alpha, I then recruited friends from engineering school to help me refine a little, do code. And so, together, we created the beta. And actually, I remember that when we were launching the beta, I was doing an internship at Amazon, my other boyfriend was doing an internship at Google and my other friend was doing an internship at Intel. And in fact, we wrote that on the site. And then, it was seen by a big site at the time called ReadWriteWeb and which had done an article about us. And so, that’s where I think it helped with development as well, in addition to all the other articles and all the people who joined the project. And that’s where the adventure really began.

There are quite a few people, especially many Germans, who came because the Germans have a particular history with the post-war period when there was the Stasi in place, where people spied on each other. There was really a terrible atmosphere of espionage and denunciation, I think we call it, where you have to denounce others and everything. And in fact, they’ve been, as a nation, traumatized by this, so they’re very advanced on privacy, data control, and sovereignty. And it just so happened that the message I had proposed, the mission to create an open source alternative to Google Analytics, had a lot with the Germans. And really, we were lucky to have young engineers for the most part, or students again, who were very good, who joined the project and contributed large parts of the project. So new features, new modules. And that’s when the project really started to gain momentum.

The structure in the early days of Piwik

Walid: When you started creating Piwik, was it just a collective of people at the beginning or did you set up a company around it from the start?

Matthieu: Actually, as I said, I was quite idealistic in this concept. There was the fact that I struggled with the idea of making money. I realized it afterwards. And that limited me a lot. And in fact, I believed in the fact that we could do a community project and so what did we do? We received donations, we also had sponsors on the site, people who paid us to put their links, logos, etc. Of course, we had people asking us, can you develop this? Can you expand on that? So we did it, we invoiced for developments. We tried a little bit of everything too, so we crowdfunded very early on. In fact, we were really precursors of crowdfunding. But hey, it allows you to raise 10,000, 15,000 euros but it runs out quickly. Afterwards, I did this internship in London. After that, I worked for Amazon. But after six months, I was tired of being in a huge corporate where we had very little impact. Everyone is very good. And so, what do we do? Every day, we saw the relationships with the number of millions that the team generated. But it didn’t speak to me. Although I learned a lot technically in those six months, especially about quality, automated testing, etc., which was something I didn’t know at all when I started. And after six months, I quit. And in fact, I had an offer, suddenly, from this company OpenAds who had made me a job offer where they told me, in fact, you can work full-time on the project and we, maybe, will use it, maybe not. So, it was an offer I couldn’t refuse since I didn’t yet have the idea of making money from a business. I wasn’t ready, actually, to do that because just doing the product and the development, it actually took all my energy. So that’s all this to say that the first step of the project, the first five, six years, was a mix of donations, sponsors, freelancing, crowdfunding and being employed full-time by a company to pay me to develop it.

But of course, that only pays me, so I had to pay the others afterwards. And that’s where in fact, one of the people who was part of the team, I started to pay him finally to say, we have to work on this. So I had my salary, I paid it a little with it. It gets complicated very quickly, in the sense that salaries are still very high. And to raise 3,000 euros in salary or something each month, you still have to be very good at business. And as a result, I was really not ready. And in fact, suddenly, a little bit overnight, I realized, now, my bank account, it’s going to be at zero in four months, if I continue like this. And we absolutely have to do something. And that’s when I resigned myself a little to the idea of setting up a company at the time.

On the other hand, I had to find a co-founder who was experienced in the business so that I could continue to focus on the product, etc. And so, in the Matomo team, there was a Pole who, as a result, had already set up a company with 40 employees, I think, at the time. He was doing consulting, but suddenly, he had proven that he could do it. And so, we started business together. And we called the business Piwik Pro, so the pro version of Piwik. And this adventure lasted about two and a half years, where we set up the business, we found what is called the Product Market Fit.

Matthieu Aubry

I don’t know if there’s a French translation, but it’s really a very important concept where in fact the product reaches a stage where it meets the needs of the market and suddenly the market appropriates it, needs it, comes to you to look for it. And it’s very hard to get to that. So it took us a lot of tries, etc. But we were lucky once again to be among the first. It was still easier. And in fact, this adventure was really a huge series of learnings for me. One of the mistakes I made, the same, I realized later, was that I didn’t have a mentor, I didn’t have a coach, more experienced people who could have guided me. I was really in fashion, I was all alone in front of the world. Of course, we had our community, etc. But in terms of business, I had no advice. I hadn’t even read books at the time, I only read books about other things. I’ve probably had my fill of mistakes, they too, my ex-partners have also made mistakes and it turns out that after two and a half years, in fact, they put me in a very complicated situation and which is also not in line with my values, let’s say, and that’s one of the reasons why we separated. In fact, they threatened me and said, well, “there you go, if we don’t raise money… We’re going to sink the company or it’s not going to work” when at the time we already had a million euros in income and in fact they were hiring more and more people when frankly we didn’t need them clearly and they were hiring it to maintain financial pressure on me to have to raise funds. And when we raise funds, since I had more than 50% but I didn’t have 90% either because I had made them an honest and generous offer too, and well I would have lost control over the business. And actually venture capitalists, when they invest in open source software, in general, they want control over the open source project, at least indirectly, through a foundation that has a limited charter, etc. And I, in fact, didn’t trust my teammates. Already, since I saw what was happening, and as a result, I wanted to protect the mission of the open source project at all costs and I knew that it would have been threatened by these investors. Although, I have to say they were good investors. I would have met, I talked to people they invested in, including the creator of MariaDB, the creator of Nginx, who spoke well of them. So, that’s not the problem. They were in their role and everything was fine, but it was the fact that I was laid off like that with the obligation to accept a term that I didn’t want, in which case I would lose everything. And in fact, I decided to start my own business. So, it was pretty special. And so, for me, it was a pretty difficult time. I was burnt out, let’s say it. My brain was, I often say, about 20% of capacity. So, it was like, I wasn’t myself anymore. I couldn’t code, I couldn’t think deeply, I couldn’t concentrate. And my body was really tired. stressed, etc.

The license used at the beginning of the project

Walid: Wait, I have a question before we go any further, because I think it’s going to be interesting for the future. I didn’t ask you, when you started, what license did you put on Piwik?

Matthieu: Oh yes, so Piwik…

me, since I was idealistic, and at the time, there was no AGPL yet, I actually loved the GNU GPL. I thought it was the perfect license and I was really a follower of Richard Stallman. I drank in his words. Afterwards, I know that he had criticism later on about his behavior, perhaps, etc. But hey, for me, he’s still a hero, really. I was lucky enough to meet him in person in India as well.

Matthieu Aubry

In fact, just before I created Piwik. I did an exchange year in India, my fourth year, and it was really exceptional, it opened my eyes to so many things. India is really a shock on all levels, and you have to let go to take what India sends to your body, to your senses, etc. And it just so happened that while I was in India, it was the time when Richard Stallman was going around the world to present the latest version of the GPL license, where they had fixed the problems that had been there for years, etc. So, I met him in person in Bangalore, I took the night train with a friend. Anyway, so that day I had met one of my heroes. And a few months later, that’s where I sat at this desk with the blank sheet of paper and created the project. So I think that motivated me too. But then, for me, it was the GNU GPL, so the flagship license, let’s say, of free software, which also guaranteed the fact that if someone wanted to take the product and change it, improve it or whatever, they would have been forced to redistribute the source. So in fact, it was to maintain this community aspect and sharing.

Piwiki, Piwik Pro and the trademark owner

Walid: At the time, Piwik and Piwik Pro were the same thing?

Matthew:

let’s say that I had seen it coming a bit and I had kept control over Piwik, so the domain name, the trademark that was registered in my name. And I had given control of Piwik Pro, the brand, and the domain name Piwik Pro to the company. So when we were no longer in agreement with the old partners and they kind of forced me to leave, so I left, and I still had control over the open source project, the Piwik brand, etc. So then, he started a rather complicated period where in fact, we had the Piwik project. They were Piwik Pro, but we were completely separated and on bad terms, in the end. I even had to sue them because they hadn’t paid.

Matthieu Aubry

Anyway, it was quite a thing, to make a lawsuit in Poland from New Zealand. But really, I was very lucky, in fact, is that at that time, while I was burned out, in fact, one of the two people I was paying with my money that was running out, his name was Thomas Ster. And so, it’s… Thomas became my co-founder. He’s a real genius, really, I would never be here without him. And it turns out that while I was tired, I couldn’t take it anymore, it took me a year to get back to, let’s say, my abilities, he was in full force and he had the energy, the motivation to recreate a business finally, from Piwik. But we still had to redo everything. The whole business, we had to recreate paid products, a service etc. So he was the one who became the lead engineer and product and I was there, I did everything else marketing, sales, papers etc.

Laurent: Thomas, is that the one you introduced me to 6 months ago?

Matthew: That’s right.

Laurent: Okay. So, he’s still on the boat.

Matthew: That’s right. He is the company’s technical director. After that, we’re in the process of changing the role because we’re in the present now. But he, like me, wants to become much more strategic and work less in the business doing appointments, reviews, etc. It takes a lot of time and energy. As a result, we are in the process of changing our positions to be much more strategic and to be able to perpetuate the Matomo project, we will talk about it again, but there are a lot of occurrences now, etc.

The beginnings of Matomo

Walid: You’re going with Piwik and Thomas, he’s all in. He is starting to work on this new project which is not yet called Matomo.

Matthieu: He works on Piwik and in fact, after one or two years, I think after a year.

Walid: Wait, but he doesn’t get paid. Finally, you’re leaving….

Matthieu: So then, how did we pay ourselves? Oh yes, I think I was paying for it a little bit with the money I had made from the sale. I don’t really remember, but I think we really paid ourselves a small salary to live. After that, we just needed coffee, in the end. Coffee, water, the meal. And we were focused for two years. We were also lucky to have had a free office in town, in a company in Wellington, New Zealand, which is pro-open source, in fact. They do open source consulting and they helped us a lot. They gave us the office for free. So, we saw each other every day. And we started to build paid options for the Piwik software, while still developing the open source platform, of course, but we had to monetize and make money. And now, I was starting to understand the business, let’s say, after the 2-3 years of experience I had behind me. But in fact, there was always this really embarrassing problem where we shared the name with people we didn’t like anymore, who had nothing to do with us anymore, etc. So then, we decided that we had to change the name. It’s going to be really difficult. The project has been called Piwik since, I think, at the time, it had been 10 years maybe, or maybe a little less, 8 years. And then, we decided to change the name to stop constantly thinking about others or being put in touch with others.

So the process of changing the name is something. I remember the checklist, we spent weeks planning it and the checklist was more than 100 steps. There were more than 100 things to change and we did it over several months. You don’t have to do everything at once, you have to do it little by little. There are some critical steps such as trademark registration, which is very important, also managing Google’s SEO, managing the redirection well and following best practices, etc. But it’s true that for us, it went well, the name change. And how did we look for the name? Being two geeks that we were, we wrote a name generator. In fact, we identified the sounds we liked and then we generated all the names that combined those sounds. And I think there were maybe 2000 of them, something like that. And then, we read everything one by one and we said to ourselves, “oh yes, this one is not bad”. And I think there were about fifty that we had like that, that we liked.

Laurent: in Matomo, there is Matthew’s ‘Mat’, and there is Thomas’ ‘Tomo’

Matthieu: yes, it’s crazy, but that’s the bias of that. We realized that afterwards, that in fact, we had chosen the name that had our two names in it. But we didn’t realize it at the time. Indeed, we happen to like the sound of our names

The Matomo Community

Walid: And when you choose Matomo and you start making all the changes, it’s never easy to recreate your branding. So you already had your community, it was already there, but it’s still a pretty critical step to recreate a brand. It’s a long-term, long-term thing.

Matthieu: yes, I think the process is still in place to create a brand image. It’s really a continuous work. And I think we have to do better, of course. After that, it’s true that we didn’t necessarily do any specific work. We tried to change the name everywhere, in a structured way. And it went well, people understood. I know that after 2-3 years, more or less, the change was finally effective in people’s heads. It was clear. And then, what we did was that we also kept the old estate. So when people searched for the old name, they found a very clear page that says, “Oh there you go, the name has changed.” So, there you have it, it made it possible to reach people who were still using the old name. So, there are different techniques to put in place to make it clear. But as long as we explain to people, everything goes well, I think.

Walid: If we look now, because we have started to talk a little about community, if we look at your community a little now, how does it make up the community? Are there regions of the world that stand out more than others? Is it very diverse? Who makes it up a little?

Matthieu: Yes, it’s very diverse.

When you look at the country dashboards, for example, Matomo is used in almost every country in the world, at least more than 200 countries. So that’s for sure. After that, we have a lot of popularity in some countries, especially in Europe. The countries that use Matomo the most are Germany. There are about 12% of websites in Germany that use Matomo, or in Austria, the same. In Switzerland, I think it’s about 7%. In France, it’s about 7% too. So these are the countries where Matomo is used the most.

Matthieu Aubry

The community, we believe… Matomo is known to be used on a million websites. which represents about 1.5-2% of the market. Because there are about 150 million active websites. And of these 1 million sites, it is estimated that there are about 250,000 people who will look at the dashboards in Matomo every month. But in fact, we don’t know, since Matomo is completely decentralized and we’re obsessed with respecting people’s data. We know how many Matomos there are, it’s about 250,000, but we don’t know anything about what people do in them. So that’s why we estimate that about 250,000 people use it every month. So then, the community, there are different types of communities. For example, there are users, the ones we just talked about. There are also translators. So there, Matomo is translated into more than 50 languages. There is also the community of contributors. Contributors, actually, I think that’s something that happens in a lot of projects. Not all of them, but a lot. Where in fact, at the beginning, there are a lot of contributors since everything has to be built. So in fact, people, they use the product, they say “Ah, but I need this and it doesn’t exist. So, I’m going to build it. I’m going to create it, I’m going to contribute it.” But once you become a mature product that does most of the things that people need, there’s less stuff to contribute that is interesting like that or really that people need a lot because you’ve already built it and on top of that it gets harder and harder to contribute as the code becomes more complex, You have to write tests, so it requires a more advanced level of development that actually limits contributions to people who are professionals, in the end, in code. And people who are coding professionals, they already have a job, in the evening they want to spend time with their family, so that’s it… We realized that the developer community on the platform itself has really shrunk, and there are very few contributions. So in fact, what we did was that we created side projects to allow people to contribute in a simpler way. For example, one of the projects we created is called Device Detector. It’s an open source library that detects the hardware used by people in Javascript, precisely, from the User Agent. And in fact, this project is a big success. There are a lot of projects that use it and there are competitors of ours who use it too. And that receives a lot of contributions. Then, what we also did from the beginning, almost from the beginning, was to create Matomo, so at the time Piwik, as a platform. A platform means that there is a common base, which is the open source platform, but it is modular. It allows people to create plugins, like WordPress, it’s the biggest example out there. We really followed this model. We were inspired a lot by WordPress too. As a result, people, when they want to contribute, so as not to limit them to what we would accept in the platform, we allow them to innovate on these plugins. And we created an open marketplace where people can publish their plugins. There are more than 100 plugins currently in the marketplace. It allows the community to express themselves, to contribute, without having to go through all the doors necessary for their code to arrive in the Matomo platform.

Matomo’s business model

Walid: It makes a transition for me, because I had asked if there were paid plugins, but it makes me a good transition to your business model. And how does it work? What is your business model? And by the way, before, is your license still a GPL or did you change the license?

Matthieu: It’s still GPL. In fact, we happen to have three business models. So it was done step by step. So the first model we launched was precisely this open source platform that had a lot of users. We decided to create paid modules for this platform. So our paid modules are not under the GPL license since in fact we distribute them via our marketplace and therefore we are not obliged to distribute them under the GPL and it was also a way to protect this part of the business that was becoming our business as a result. So we created paid modules. Currently, we have 16, but it happened little by little, one by one. So it’s modules such as video analysis on a site to know which videos are watched, until when people watch. We have a module for form analysis, to find out how people interact with forms. We have an A/B testing module, where we can compare two versions of a page in a scientific way and measure which one performs better, etc. So, there are like that 16 paid modules. And so that’s called an Open Core business model, where in fact, the core of the platform is open, but the modules are not. So, although people, our customers can see the code, but they don’t have the right to modify it or redistribute it.

So we started like that and in fact, frankly, we thought we would continue like this. The idea at the beginning was to run a small business in the end, not to worry too much, to have few employees, if possible to make big profits and not to bother too much by selling software and licenses like that. We did this for 2-3 years. At the same time, we still had companies that were using the Matomo platform on their infrastructure. And when you’re a big company and you use open source software like that, very often, they need technical support. And they want to have someone to talk to when it doesn’t work. And I, very quickly, wanted us to rise to a position where we would give this guarantee to our customers. I wanted Matomo to be seen as a truly professional project, that companies could trust. And I saw it as a bit of a test. I tell myself, if we support companies, at least we’ll be sure that it works and we can tell customers, it will help us sell more plugins, etc. So we also launched that, the support offer. At the beginning, I was the one who did the support for several years, I created guides, etc. And I liked it actually, it showed me that the project we’re doing is used by CAC 40 groups, etc. The big companies were ultimately a validation of all the work we had done.

So we had these two business models, so support and open core, we sold licenses. But I, in fact, I don’t know, for quite a long time, I was still, how do they say, enthusiastic and I really wanted to, I don’t really know why, but I wanted to run a SaaS company. I thought, it’s so cool to have servers that handle data 24 hours a day and manage that. It’s something that excited me at the time. And it actually happens that I don’t really like doing that, managing servers or whatever. So, of course, we can do with employees who are very qualified in this.

But I had this intuition that I wanted to get into it. But it turns out that Thomas, my founder, didn’t really want to. He wanted to stay on the simpler version, etc. And so, that’s where we didn’t agree, actually. It’s a bit like the first time. And that’s where we hired a coach. that helped us resolve this conflict. It so happens that I managed to convince Thomas to launch the business model called SaaS, Software as a Service, or Cloud, and that’s when we launched Matomo Cloud, the service.

Matthieu Aubry

Oh yes, I’m saying that I don’t really know why I wanted to do this, but in fact, I do: it’s because the future, let’s say that now, we see it, a lot of services are SaaS, simply because it’s simpler. to use, to register, to set up. And I saw it coming. People, sometimes, on calls, would say to me, “Do you do SaaS?” People were asking for it, at the very least.

So, we embarked on that, in our third business model, which was Matomo Cloud. And in fact, today, it represents 75% of our revenues. So, we launched it last, but the growth was really much faster than the rest. Our company has 75% SaaS. 20% license sales and 5% technical support.

Matthieu Aubry

Walid: You have all the possible models to make money around free software.

Matthieu: I think we tried everything. Between the first step, donation, crowdfunding, being hired by another company, freelancing, now having these three business models.

I think we tested a lot before coming across this mixture that works well for us. In fact, currently, we also have a clear message to our community, it’s yes, we’re fully into open source, on-premise, you can be sovereign over your data, we have zero access to your data, on-premise and we support it. Or you like our values, the product, what we stand for, but you don’t want to go so far as to install everything yourself, use the cloud version.

Matthieu Aubry

And in fact, we’re not pushing one or the other, we really want people to choose the one that suits them. And it turns out that for most people, it’s the cloud version, since it’s simpler.

What are the criteria for putting a feature in the Open Source or proprietary version?

Walid: Before we move on to the advantages of Matomo compared to the competition, one last question. How do you decide if a feature goes into the open source core or if it’s going to be a proprietary module? When we interviewed Emily Omier last time, she said, “Actually, if you want your community, your customers, to understand what you’re doing, the message has to be simple enough for them to understand. If I take the open source version, I have this. If I take the proprietary version, I have this.”

Matthew: Absolutely. It’s really difficult. There is indeed an analysis that I don’t know if I would be able to explain like that, but really, it’s more of an intuition. In fact, you have to respect the community. In fact, I’m the first user. We are the users, so we understand when we download an open source product, we want it to work, to meet the needs. And then, it’s about knowing what percentage of needs or people go… directly be delighted to use this product.

And in fact, oh yes, maybe one of the criteria we use is to know if this feature is useful to a freelancer who puts this on his personal site or his blog, or is it only useful to a company in the end, to a commercial site?

Matthieu Aubry

So if the answer is that, to do A/B testing for example, frankly, there are only companies that will do that, and in addition to companies that have a lot of traffic, because it takes a lot of traffic to do A/B testing, and then it’s clear that that, limit, There is no point in making it open source. Well, yes, there’s interest, but I mean, it’s clear that it’s a paid feature.

That’s how we reasoned, more or less, for these 16 features. But now, since we launched 16 of them, I don’t think that this past year, we haven’t launched any, or even the last two years.

So now, we invest more proportionately in the maintenance of the existing products we have created and in the maintenance of open source as well. So we’ve created so many products that now we don’t want to create too many of them, but rather improve on what we’ve done and provide big value to our existing customers and the community.

Matthieu Aubry

It’s true that it’s not always easy, but an intuition to say to yourself…

Walid: It changes from one product to another actually. From one product to another, the answers are different depending on the founders, depending on the business and all that.

Matthew: That’s it. But what we want is for the product to be functional in an open source version. In fact, if you look at the product, when you have activated the 15 or 16 features, suddenly, in fact, the menu of the application is about twice as big. So, half of the things are open source, half are not. Afterwards, in terms of code, I think there is more open source than private, but that’s pretty much the idea. In visual functionality, there are half that are Premium and the other half that is not. But the half that is not premium, that is free, is really the basis. These are the really essential things.

The Matomo Company

Walid: Another question that comes to me. The company is a New Zealand company. How many employees are you and where are your employees?

Matthieu: We have just reached 50 employees. There are about 30 in New Zealand and 20 overseas. In fact, we are 100% remote. At first, I said that we were in this office with Thomas. But after about two or three years, right after we hired our first remote employee, we said to ourselves “we don’t need to come to the office every day anymore”. We know each other so well. We’re going to talk on the Internet instead. And so, we both started working from home. And from the beginning, anyway, we didn’t want an office with employees. In fact, we also wanted to have access to a larger pool of talent. So, it was clear from the beginning, it would be full remote. So there you have it, 50 employees. There are roughly, I think, 7 in France, 5 in Germany, 4 in Canada, 2 in the United States, one in Australia, 1 in India, and one in the Philippines.

Walid: It’s good, time zones to do support.

Matthieu: Yes, I think we’re in 16 by 5. And we are trying, little by little, to make it 24-6.

Walid: well, the clock is ticking. So, I have two more important parts that I would like us to address, unless Laurent has questions.

Laurent: No, no, no. Look, he’s doing well, frankly.

Matomo vs. the competition

Walid: I wanted to talk about how Matomo compares to the competition. You said earlier that there was a lot of competition. In fact, can you explain to us a little bit what the competition is and how you stand out from the competition?

Matthieu: So that’s really my current battle.

We have a lot of competitors, and in fact, we have dozens of competitors, but we have about maybe 20 who are really serious and very good. And out of the 20, there are 5 that are companies worth more than a billion.

Matthieu Aubry

So, there is Google with Google Analytics, which is used by 80% of websites, and which is free. So, we are competing with Google and a free tool from Google. In addition, we are in competition with Adobe Analytics, Yandex Analytics, the Russians, Content Square too, and Amplitude, the Americans.

These five companies are worth more than a billion, they have hundreds of really highly qualified employees, etc. It’s complicated. And on top of that, in recent years, especially since the GDPR, the data protection law, there have been a lot of other players who have come out, who have also been inspired by us, it must be said, and by others, and who have launched competitors. And as a result, we have a lot of tools like Plausible, Fathom, etc., which start as simple, they are positioned as simple analytics tools, but over time, they also evolve.

And so, it’s true that now, we’re in an incredibly competitive market, and it’s getting very hard. And so, that’s where we have to be able to stand out, to play to our strengths.

Our strengths are the fact that we are a very established, very complete tool. So on the one hand, we are very complete, while remaining easy to use and pleasant to use. And we’re working on it to improve that. So we hire designers, etc. Then there is the aspect of respect for privacy and ethics. We want to be ethical and we are ethical. So we have to do what’s right, what’s the right thing to do in any situation. There are no shortcuts when you are ethical, right? That’s a strength. Most of our employees also come because they reasoned with this message. Our employees are ethical. We are working, we insist on it.

And to conclude, the last aspect that is unique to us is that we are open source and we are an open source that is mature. This is called “enterprise proven”. So, Matomo has been tested on the battlefield for years, we’ve tracked billions of hits, it works, there are hundreds of possible configurations. So it’s ready for the company at any level.

Matthieu Aubry

Laurent: But it must be all the harder because the open source argument, which used to be a real strong argument, today also has to face open source competitors, not just proprietary ones.

Matthew: Exactly. Although the open source competitors, after I don’t know all of them, but the ones that are not very simple, the very simple tools, yes, they are open source. But there are some who call themselves open source, but they are not as committed to open source as we are, in the sense that there are several who at the beginning were open source, and after that, they stopped developing open source and it went commercial.

But they leave open source like that to say: “Ah, you see, we have a little bit open source” or “We were open source, but now we’re not.” So, they play a bit like that on things, while we’re really purists, let’s say, open source, although we do open core. So, some will say that it is not purist.

But then, the competitors, if they are really entirely open source, for me, it’s not necessarily a competition. Rather, they also help our mission. Our mission is precisely to offer an open source alternative to giants like Google to control this data. So if there are others who do the same thing as us, that’s great. We would like to encourage them too. It’s rather the others that are not necessarily open source that are also very good.

Matomo and institutions

Walid: There’s one last part that I’d like us to discuss and you’ve started to talk about it a little bit. We told you about the GDPR, and in fact, what I’d like to know is a bit of the relationship with the different institutions. Do these institutions finance you? Do they help you in any way? I think that the GDPR has been a big plus. In fact, basically, a little bit, do you have any relations with them? Do they use your products?

Matthieu: When the GDPR came out, it was a huge boost for us. In fact, I remember, the GDPR was just out or just before it came out, and we had to read it, actually. We read the law and tried to understand. It’s not easy because we’re not lawyers, but hey, we managed to understand. And what we did very quickly, we implemented changes in the product so that it was really compatible with the GDPR and that it made life easy for people who had to make sure that their tools and data were compatible. So that helped us a lot, indeed, to win new customers, new users.

Walid: In France, you had articles on this that said: “yes, but wait, Google Analytics is not GDPR, we should rather switch to solutions like Matomo”, etc. It also boosted our notoriety.

Matthew: Exactly. In fact, what happened was even for a period of time, Google Analytics became illegal. Especially in France and a few other European countries, where the institutions said: “It’s illegal to use Google Analytics because Google hadn’t done the job and refused to work with the institutions.” And for us, it’s true that it benefited us a lot, but it didn’t last very long. I think it was a year. After that, we were also helped by the CNIL.

The CNIL is really incredible people. The work they’ve been doing since, I think, the ’70s, when it was created, it’s really a very important job. And they have set up a kind of labeling that allows certain tools, if configured in a certain way, to be free of consent.

And this is one of the strengths of Matomo that I forgot to mention, is that in France, Spain, Italy, in some European countries like that, where the CNIL’s recommendations apply, it is possible to use Matomo to measure visitors without asking for users’ consent. Data is measured in a way where it is impossible to correlate the measurements with an individual person, etc.

Matthieu Aubry

Laurent: including on the IP address. That’s what I liked when I was looking for a solution using tag technology. There is a clear feature in Matomo. We can, let’s say, anonymize IP addresses with a simple checkbox and, from there, the last piece of data that can be used to identify or trace someone back is no longer stored. Indeed, we can get out of the consent process.

Matthew: Exactly. That’s one of the criteria to put in place, and that helped us a lot. Afterwards, to say that we have relations, certainly, but they don’t finance us. After that, there are some who are customers. We have quite a few cities or regions in Europe that buy Matomo Cloud at 20 euros per month. But we have never received funding or asked for anything else, we don’t need it. Afterwards, yes, the CNIL was really instrumental for us.

Walid: And in the United States, for example?

Matthieu: It’s our third most popular country after France and Germany, especially in terms of income. So yes, there is indeed a certain popularity of Matomo. It may also be a lot because American companies, which are international, must be compatible with the GDPR, as long as they serve European customers. So that helped us too.

But generally speaking, in the United States, states are starting to launch their privacy legislation. And in fact, Matomo is an ideal tool to be compatible with all kinds of legislation, since we have a lot of controls and parameters that allow us to control all this. But I think that in the United States, there is a lot to do. And we, since we are small, with a limited budget, etc., we try to focus on Europe for the moment. But the United States is really our second market.

Walid: And Oceania?

Matthew: Oceania, no. Even though we are based here, it represents 1 or 2%. Not even.

The announcement of the halt in the development of AWStats

Walid: Laurent, do you have any questions before we move on to the conclusion?

Laurent: no, we’ll be able to move on to the conclusion. Indeed, all I can say is that indeed, I was a former user of AWStats, of course. Like many, I was seduced by the tagging technology, which still brings a plus in terms of information that can be reported. I was seduced by the simplicity, the screens, the monitoring offered by the Matomo software, with its dynamic consulting. When all software today has a log analysis, it’s really its openness side, the clear GDPR.

Because, when we take over Google Analytics, they try to be GDPR compliant, but will find the option that allows them to be. And you have to be a big guy to understand what the options are to select to be able to be so. So on Matomo, it’s clear. I mean, you see the data that’s being collected. You know that in the end, the most critical is going to be the IP address, since you can go back to the person’s home with it, by rebound. And Matomo, that’s it: you have your checkbox, my IP address, I anonymize it, ultimately sacrificing the end of the IP address, for example.

And so this simplicity of choice, a box to be ticked to be able to comply with the GDPR, is something that I really liked and that makes me focus today on my projects that need follow-up by tag analysis, on Matomo in particular. And there you have it, I encourage everyone to do it as well.

Laurent Destailleur

And then, I don’t know if we can announce from the outset what I wanted to announce in this podcast as well.

Walid: Go ahead.

Laurent: We introduced differences in technology, log analysis, tag analysis. We can see that with the evolution of IT, AWStats is a software that is now 25 years old. Its technology, today, is no longer up to date, even if it still provides services to some in terms of choosing the need.

You should know that AWStats, for a while, was nearly 20 million websites at a time when there were maybe only 80 million. I don’t know if it was 20 or 30, it was an American study that was done. So in terms of market share, we were at 25-30%. So it was a colossal success. But today, we are at a level that is close to zero. And so, it’s true that the investment in this software, for me, is no longer as interesting. I prefer to devote myself to other software. And so that’s why I made the decision to end my adventure with AWStats.

Laurent Destailleur

I’m going to make a final version with all the latest bugfixes, patches, pull requests that are still on stand-by, even if there may be few. There are very few contributions, which also shows that now the public, the users, have already started to make the switch. There is really very little activity on this project today.

And so I’m going to finish this with a version 8 which will be the last version of AWStats that I’m going to make. I don’t know yet if it’s at the end of the year or if it will be in May 2025. Why May 2025? Because the project was born in May 2000, it will be 25 years old. To mark the occasion, it’s not bad to say that it makes him have a good end of life, to die on his 25th birthday.

Laurent Destailleur

Of course, it’s an open source project, so other people may be able to take it up and bring it to life under another name, AWStats New Generation, eventually, if I wouldn’t oppose it, quite the contrary.

On the other hand, the adventure of the AWStats project, which is still a project that will have marked the history of open source, will end in the coming weeks, if not months. And so, I encourage everyone to switch to a more modern solution. We have mentioned a few of them, but here it is. I chose Matomo. Everyone is free to choose their own.

Walid: I’m going to shed a tear.

Laurent: It allows me to focus on other projects that are just as important, if not even more important today, that you know well.

Walid: Well yes, which we’ve already talked about. Yes, that’s perfect. I will leave each of you an open forum to say a final word. Matthieu, do you have an open forum? Do you have a message to pass on before we leave?

Matthew:

A message for the younger ones who are listening to us, to really enjoy it when you’re young. We have a lot of energy, a lot of time and it’s time to find people who think and do things together, to learn to collaborate, to create things together, whatever it is. It can be creative or athletic or whatever, but it’s really important to learn to be in a group and to do things together, because to do great things, you can’t do it alone. So, you have to learn very early on to work like that with other people who are sometimes different from you, even if it’s easier to work with people like you, it only takes you to a certain level.

Matthieu Aubry

Not only to work with others, but also to read. It’s true that I think that reading is very important for young people, to acquire knowledge, and not only to read, but also to put things into action. Maybe my advice that I would give to my daughter and that I give to the young people who are listening to us.

Laurent: It makes me laugh because for me, it’s a young person who says that.

Walid: You’re surprising me.

Matthieu: a young man almost forty years old in three months.

Walid: I’ve always liked to tell the stories on the podcast and tell the story actually. And what I think is great is to have two generations of tools in two different periods. I think that’s really great. That’s why I found it very interesting to have you both at the same time.

Matthieu: thank you Laurent for all your work. Of course, you were an inspiration to me too. I looked at AWStats and seeing your work, it motivated me to try to do better in the end, or differently, or as you said, more modern with it. Congratulations for everything you have done. And then you followed up with another huge project that is a big success, so it’s really impressive. Multi-entrepreneur, free software player.

Laurent: I’ve stayed young, you know, I still have the motivation.

Conclusion

Walid: perfect, we’ve reached the end, we’ve been recording for 1h20. Thank you both very much. Thank you Matthieu for recording with us. I was very curious to have your testimony about your whole story and the whole history of Matomo, etc. It was really exciting. I hope that the listeners of the podcast have learned things, who will try Matomo, who will adopt it, who will migrate from AWStats for the older ones to Matomo.

And then, as usual for listeners, I ask you to share this episode, to talk about it around you. I’m going to share it with my colleagues who don’t use Matomo. Maybe it will give them some ideas. And then we’ll see you soon for other episodes. Thank you both. Good evening Laurent and good day Matthieu and then see you next one I hope.

Matthieu: thank you Walid, thank you Laurent. And yes, see you soon at FOSDEM, not this year, but for sure one of the next years I’ll go. In any case, I can’t wait to meet you.

Laurent: come on, hi Matthieu!

Walid: it’s working, hello!

This episode was recorded on November 19, 2024.

License

This podcast is released under the CC BY-SA 4.0 license or later.