Florian Fournier (Payfit): how to invent its own language and develop its own low-code platform

We all know Payfit. In just a few years, this start-up has become a scale-up with an impressive growth. Since its 2015 creation by all three co-founders Florian, Firmin and Ghislain, the company has raised €100 million, employs over 550 employees, and has offices in several European cities including Paris, Barcelona, Berlin, and London.

Payfit is a member of Next 40 and has won numerous awards. This is quite an impressive track record for a company that is only 5 years old. We all know Payfit for its dynamics, but less so for its origin and its technical base, which make it quite unique. That’s what we’re going to talk about here with Florian, CPO and one of Payfit’s co-founder. 

You graduated from an engineering school. This is not necessarily the best road for entrepreneurs to follow. How did you get into this adventure?

Since I was 16, I’ve really had this desire to try the entrepreneurial adventure. It comes from my family. I was lucky enough to have examples such as my grandparents or my uncles who set up successful ventures. I’ve always been influenced by this state of mind, and it has always inspired me. So, I wanted to try it.

As soon as I started engineering school, I set up my first company with some very good friends from prep school. It didn’t have Payfit’s ambition, but we still had a lot of fun making and selling gifts for military regiments and prep schools.

A few years later, at the end of my studies, I joined forces with Firmin and Ghislain to launch Payfit. I had met Firmin in prep school, he was in economics when I was in engineering. We kept in touch, and it was he who introduced me to Ghislain. They were childhood friends and grew up together in Laval.

As of today, what is your function, and what are Payfit’s current challenges?

I am in charge of the product, design and engineering teams at Payfit. They represent more than 160 people. We are organized in a unique way, with both global and local teams. The global teams represent about a hundred people and are built around a classic organizational pattern with software engineers, product managers and designers. Alongside this global team, about fifty people make up the local teams, which we call product builders or JetLang (our language) masters. They have a great deal of responsibility as their mission is to implement the best payroll solution in their respective countries.

As far as our challenges are concerned, I see two main ones. The first is to strengthen the reliability of our payroll and reporting products in each country and to simplify user experience. Our ambition is to make them easier to learn and use, and thus to encourage customer autonomy. We want them to be as user-friendly as possible. The second challenge is to offer a wider experience regarding human resources and enable our customers to manage the lives of their employees, from on-boarding to off-boarding, in a very simple way. The idea is to go beyond payroll management. We really want to develop this HR component so that, in addition to being a payroll solution, we can be a real Core HR solution.

As a CPO, what is your typical day?

They’re very eclectic !

There is a multitude of things to manage because I wear both a CPO and a co-founder hat. A large portion of a typical day as CPO is devoted to working with the engineering, product and design VPs, whose teams represent a large proportion of the workforce. I work with them on team structure, possible improvements, friction points, etc. This is the real management side.

As for my second function, as co-founder, it has more to do with strategy and, amongst other things, planning future investments. We offer several products in several countries, each with different needs. So, we ask ourselves a lot of questions about the future, especially on the product side. We try to both arbitrate between choices as accurately as possible and build the best road maps for the different teams to better serve our customers, month after month.

It’s really about these two things: the day-to-day management of, and care for, the well-being of the teams on the one hand and, on the other hand, taking a broader perspective with preparations for the future and the definition of our objectives. That represents a lot of discussions within the product team, but also with other departments. The team is growing fast, and it is becoming less easy and less intuitive to harmonize the lot of them. So, we spend a lot of time on it. We try to both gather their needs and align the teams (sales, customer…) with each other, in all countries. It’s very exciting!

Remaining agile is also a great challenge, especially when you’re growing so fast! How is it going on that front?

Yes, it’s a very important subject. We want to avoid becoming a big organisation with such rigid processes that we would lose the ability to innovate. I think it is thanks to innovation and bets we made, particularly in terms of architecture and product, that we are both where we are today and able to serve so many customers throughout Europe. And we’re really afraid, as we grow, of losing this capacity to innovate.

As a consequence, we are providing very concrete answers at the organizational level, particularly with this idea of start-ups within Payfit. We really want to build an organization with a very high degree of autonomy in all countries, in terms of both business and product. We want to give enough direction to the teams upstream, for them to be able to evolve freely and make the best decisions by themselves later on. Likewise, we try to ensure that each team has as few constraints as possible and that they take full ownership of their project. Trying to give autonomy while retaining a degree of control and guaranteeing efficiency is a very complex thing to orchestrate.

When you got onboard of the Payfit adventure, did you imagine the project would reach this scale with this kind of success?

It would be arrogant to say that I was confident we would get there. However, I remember our first discussions with Firmin and Ghislain, when we were really, concretely talking about Payfit. We had a huge ambition from the start. That led to us making some radical choices, especially with the creation of JetLang, our metalanguage. We were convinced that it could go very far, but we had no guarantee that we would get there. Above all, we were ready to take the risks inherent in these very strong choices. Today, I think we are very lucky to be where we are. And it has been possible thanks to the great team we’ve managed to put together from the start. That’s really the main reason for our success.

How did you come to work on payslips?

It’s based on a relatively simple observation. In 2015, nobody was satisfied with their payroll management solution, at least in France. We started with this observation and were inspired by what was being done in other countries, particularly in the United States. In France and Europe, payroll management is much more complicated than in Anglo-Saxon markets. We were convinced there was a huge unmet need and that if we could solve it, with very high-quality products and services, we would have a real impact.

However, in order to convince customers, we needed to develop a product and a service of sufficient quality. We wanted people to be happy using their payroll service. That was our mission from the start. The challenge was also to create a better solution than the ones created by our competitors. So, we spent the whole of 2015 creating a specific language. We knew that by not doing what everyone else was doing, we would inevitably achieve a better experience. So, we invested early and heavily in the product and the architecture. This allowed us to gain our first clients in April 2016.

And none of you had ever worked in the fields of HR or social law?

Exactly. And I think it was an advantage not to be aware of this complexity in 2015, when we set out on this adventure. It helped us. Because if we had been aware of the full extent of labor law and the reporting processes in France and various countries, we would have asked ourselves a lot more questions.

Let’s try to dissect the reasons behind your success. We can start with the creation of your own programming language! Why did you make this choice?

In 2015, we pondered how to simplify payroll management, social declarations, and the whole French administrative labyrinth. And we realized that, even with the best developers at hand, it would be impossible to code labor law.

First, because these developers would not be payroll experts and they would have to work with people who had the expertise. Secondly, because it is so complex that we would end up building something very complex and convoluted. We wanted to give payroll experts the ability to create end-to-end interfaces. This was achieved through the creation of a metalanguage called Jetlang. It can be seen as a low-code platform, i.e. giving non-developer profiles the ability to enter rules and define calculations that will be carried out on the basis of these rules. Specific local teams, the product builders, are in charge of this Jetlang project. They are experts in payroll and declarations.

Can you tell us a bit more about the organization of the teams?

Our product contains a first part that is common to all countries and does not depend on the rules specific to each geographical area. This relates to everything that does not depend directly on payroll, such as holiday management, absence, the calendar, the employee area, etc. This part is managed by a global team which is organized in the classic way with software engineers, product managers and designers. They are the ones who build the common product for all countries.

The product is then made up of a second part that is unique to each country due to administrative specificities. It concerns payroll and social declarations. This is built by the product builders, who are divided up by country. These local teams, payroll and declaration experts, are the ones who use Jetlang. They have complete autonomy and do not depend on anyone to make changes to the products. This is a huge advantage. In this way, there is no need to arbitrate and prioritize between the needs of each country.

Had you seen this kind of low-code model working elsewhere, or did you conceptualize it yourself?

We started without being aware of the concept of a low-code platform. Today, our teams are very competent on Jetlang. They have a great knowledge of the platform and are familiar with best practices. But at the beginning, we groped our way through it.

At first, we approached the problem in an iterative way. To avoid just one of us taking care of the product, we all took control of the product to enrich it and make it evolve. But the more we progressed, the more this theory of having local teams who are Jetlang experts became a reality. When we started the iterations on this metalanguage, we didn’t realize that it would represent such a big competitive advantage and that we would push it as far as we have done in recent years. We coded Jetlang in JavaScript, and we coded the rules in Jetlang. So, we ended up creating our own low-code platform.

Where does the name Jetlang come from?

We didn’t theorize too much. The first thing was that it sounded good. And then Jet for rocket and Lang for language.

How long did it take you to set up this platform, and what were the main stages?

We spent all of 2015 on it. The first testers used the platform in January 2016 and the first clients in April 2016. But you have to know that we have been improving Jetlang constantly since 2015. Today, a team of nearly 25 people is working solely on Jetlang to make it a more powerful language while also enriching our local teams.

Our goal is to simplify payroll management even more. To maintain the level of competence around this language, we have created an internal training course. An advanced learning process for product builders with a side on the technical part common to all countries and another side specific to each country to understand the architecture and use of Jetlang in that particular country.

We also have a large team because we still have a long way to go if we want to reach the kind of experience we want to offer our customers. We want to go further, and the tools developed in Jetlang need a lot of improvement to get the experience just right. That’s why we continue to invest in its evolution. It’s important to us. 

Do you know of any other company that has launched or tackled a complex problem with this approach?

The only company with a similar approach in recent years was Business Objects, which, if I’m not mistaken, was acquired by SAP. They had a language like Jetlang to manipulate objects within an industry-dedicated application. There must be other companies, but I’m not aware of any other.

Did you think that through this approach you could potentially both rethink other complex subjects and uberise other sectors?

Jetlang could be made more generic and more “payroll agnostic”. As of today, it could already be used to make a product on a completely different subject. Using Jetlang or a metalanguage is especially valuable when you’re dealing with a sector with a lot of rules. The value of Jetlang is mainly related to payroll and to simplifying the administrative labyrinth. So, it could be used, even today, to make a product that is not linked to payroll and even more so if it were in a sector with a lot of complex rules. So, yes, it’s not at all excluded that, in years to come, Jetlang could be used for other sectors, at Payfit or on an Open Source of some kind. Nothing has been decided yet.

What is the distribution of roles between the CPO and the CTO?

Today, our organization is special. The engineering organization is managed by the VP of engineering, the product organization by the VP products and the design organization by the VP of designing. We apply this triforce concept at VP level and I manage the three VPs. Then we also have a triforce at director level in each of the tribes on the major subjects. For example, the Time topic is managed by a triforce of directors with an engineering director, a design director and a product director. These directors then manage squads. These squads are headed by an engineering manager, a product manager and a design manager. The triforce concept is very strong at Payfit and is very effective in global teams.

Between 2015 and 2021, how did you set up R&D and the technical management at Payfit?

The first phase (2015-2016) was devoted to Jetlang and creating a payroll solution using it. From mid-2016, we decided to create the employee space, which is the first HRIS brick built in and not linked to payroll. It allows employees to post holidays and absences, to see their team’s calendar, and to record expense reports. In the years that followed, we developed this HRIS with workflows, document management, on-boarding, etc. At the beginning, we really concentrated on payroll, then we gradually widened it to become a tool, not just for payroll, but for the simplification of all HR processes for SMEs. And that’s really our ambition, even though we still have a lot of work to do to simplify these HR processes and allow our customers to easily manage the professional lives of their employees from start to finish with Payfit. So, from mid-2016, we started to build this HRIS part.

Then we decided to open up in Spain. This was quite a milestone because it was the first time we used Jetlang abroad. At that time, we had to see if we could adapt it to other countries. This internationalization was a great challenge for the product and for engineering. Then, there was a big focus on our technical stack and our desire to split it up into some kind of multiservice. This is a problem that many hypergrowth companies face and which we experienced first-hand. We made many improvements to the product. This has led to an increasing complexity in terms of technical staffing. At the same time, we realized that it was becoming more and more difficult for each of the engineers to have a good perspective on the whole of the technical staff.

Each new development was becoming increasingly complicated and understanding what needed to be changed and the impact of that change, was becoming increasingly time-consuming. In addition, we noticed that it could cause production errors for our customers with unexpected product behaviors. That’s when we decided to divide this monolith into multi-services so that each of its parts could be completely controlled by the teams. And with this independence, each team was able to work on its microservice with a better understanding and faster.

That’s one of the architectural bets we made with Jetlang. But there are others: how to create an agnostic payroll metalanguage? How to remove all the rules to be recoded in Jetlang in relation to employment contracts and social declarations? How to extract this complexity from Jetlang? How to manage this problem and an entity system which, from a technical point of view, would be completely independent of payroll? It’s a technical vision that results in a lot of successive iterations, to which we can devote a lot of time!

What are the organization’s current strengths allowing you to preserve this culture and agility on the technical side?

Our first strength is to have autonomous teams in each country devoted to the implementation of the rules of these countries. This is a huge advantage! Our product teams have the necessary independence to improve their payroll product according to the specificities of their country and to code changes in local labor law.

Our second strength is that we have divided our product organization (software engineer, product manager, designers) into tribes with very specific areas and responsibilities for very specific products. These tribes are also very autonomous. This autonomy is achieved not only through clear direction, but also by feeding these teams and giving them the means to explore a huge number of subjects independently. This was really challenging to set up. We spent a long time thinking about how to feed each of these people with real, prioritized information in terms of both existing users and business needs.

To achieve this, we created a highly regulated process. We try to understand precisely why new and existing customers are lost. As soon as a customer is unhappy or dissatisfied, we categorize it. This allows us to have both qualitative information on the precise reasons but also quantitative information because we are able to measure in each of the countries where there is friction. Having both quantitative and qualitative data on our users’ needs allows us to make decisions with a very high level of awareness in each of the teams and to give them relevant missions.

Recruitment seems to be an essential area to ensure the smooth running of your organization in terms of both hard and soft skills.

Absolutely! The team we have managed to build is the main reason for our success. We owe this to the level of excellence of the profiles we recruited and those they recruited themselves, but also to the pleasure we have taken in working together since the beginning. The culture established with these first profiles and its continuity through our rapid growth are essential. We enjoy working together on a daily basis, even in difficult times, even during huge rushes. It’s exceptional! Our teams are not only very competent, but they also enjoy working together. That’s why we’re moving fast and can look forward to the future with such confidence. Thanks to this strength, we will continue to move fast and fulfil our mission which is SMEs’ digitalization while continuously having an impact on our existing clients.

What were the right decisions or the major elements of your organization that allowed this team to have this spirit and be kept so united and committed?

The first thing that may seem anecdotal is the barbecue test. During our first recruitments, before making an official offer, we invited candidates to meet the team informally over a drink. At that time, we really encouraged them to talk to the different departments to see if they enjoyed spending time with them. And we also asked employees to talk to the candidates and give us feedback on whether they enjoyed the conversation and whether they saw themselves spending time with them. Right from the start, we gave real importance to this culture, and it was immediately very strong. We made our employees responsible, regardless of their place in the hierarchy, and we listened to them. Based on their feedback, we sometimes refused profiles that were excellent from a technical point of view. This culture is essential and non-negotiable! And it’s quite unique. It allows us to have a very strong cohesion.

Then, the second element is the difficult decisions that were taken.

The third thing is something we used to do at the beginning, but no longer do now. For a long time, we organized weekends every three months with the whole team. Together, we felt more like a bunch of friends than a group of colleagues. Today, we no longer do it in this format because it is very intrusive and logistically complex, but we keep on organizing happenings to meet up and spend special moments together that are not directly linked to the day-to-day work. I think it’s very appreciated and valuable to create memories together. And we think that it is even more important today, because many new operators come on board every month in each country and many of them will only get to know each other through these moments! It’s a big cultural challenge, especially with each country’s specificities. We want each employee to feel at home, no matter which office they are in.

In terms of recruitment, can you tell me what skills and mindset you are looking for?

It depends a lot on both the team and the department. But in terms of mindset, there are common traits to all employees and all departments. The first is benevolence, being curious and really interested in people. It doesn’t matter whether you are a VP or a trainee! The second is humility. It’s not necessarily the top performers who are put forward the most. The third trait is ambition. A huge ambition for Payfit regarding the experience and the impact we want to have on our users. We really want people to come and help us evolve, propose new ideas and new experiences. Another thing shared between our employees is participation, the fact that they want to live intensely together not only at work but also during more relaxed moments.

What are the stages of recruitment at Payfit?

The first step is to get in touch with our Talent Acquisition team to talk about what the person is looking for, to be sure to direct them to the right team, to present Payfit and our business and to help the candidate better understand the environment and the areas in which he or she could develop.

Afterward, candidates will have several discussions with people from their future team or department. On average, 5 interviews take place on different days to have time, potentially, to debrief between exchanges and for it not to be too heavy on the candidate.

For most of the positions, a technical test is also administered. It requires the candidate to prepare for it. This test is extremely important! Because the more you grow, the higher the level of excellence and the more essential it is to have this filter in place at the entrance gate.

Then, as soon as the competence part is validated, we move on to the culture part. This allows both the candidate and the employees to see whether it’s a fit or not. The candidate will be able to talk informally with the employees and better understand our culture.

And finally, to date, all candidates have had an interview with me or one of the other co-founders during which we answer their questions about the adventure, the vision, and the future or even the values.

In this adventure, what has been the most challenging for you?

The most complicated thing for me, personally, is the fact that I have to adapt and learn a new job every six months. With the growth we have, the organization changes enormously and regularly. So much so that daily life changes all the time. This is very challenging and requires you to be very humble. In six months’ time, you may no longer be the best person at your job because expectations, operations and organization will have changed. It requires both a particular mindset and continuous training. So, I read a lot of books and articles, I listen to podcasts about people who have the same responsibilities as I do. I also exchange a lot with these profiles (CPO, CTO…), mentors or people who have experienced the same growth and were in the same positions. And I have been able to meet some very interesting people! But the more you grow, the harder it is to find matching profiles. That’s when it’s valuable to have investors who have companies of this type in their portfolio and who can put you in touch with them, especially across the Atlantic.

Are there things you could have avoided or understood a little too late?

There are so many things we could have done differently and so many different focuses we could have had. For example, on internationalization, we should have done it later to be able to focus more on our bases, especially in France but also on Jetlang. With more time, we could have done more solid things.

The second thing is investments and foundation choices. On the technical side, there are parts you build knowing they are not ideal while also being aware that, in time, their modification will not add to the updating cost. So, there will be no cost impact of redoing this less-than-ideal part. But there are other parts, where each passing month adds 10% to modification and management costs. These choices, often made to meet customer needs in a context of hypergrowth are difficult to make, and different ones could have been made.

There are also areas of focus where we could have been more precise. On specific customer segments, for example. I’m thinking particularly about the size of our customers. Today we are able to serve all customers who have one to several hundred employees. But very quickly, we accepted clients with several hundred employees. Perhaps we went too fast. We could have both chosen not to mobilize a team specifically for this and refused this sizeable client for a time, even if it brought in a lot of revenue. We could have taken it on six months or a year later. We could have avoided a very strong focus that ultimately did a disservice to the majority of the other clients. In each case, at the time, these choices were legitimate. So, it’s difficult to say today that we could have done better. And I think we will continue to make a lot of mistakes. But the main things are to be ready to accept one’s mistakes while remaining agile and able to make quick decisions.

What are the big tech and R&D challenges ahead of Payfit?

First, we’re going to focus on strengthening and making the payroll product more reliable. That still requires significant investment. Then we want to offer a complete HR solution by providing a complete employee management suite from on-boarding to off-boarding. Today, we offer a powerful suite, but tomorrow we want to offer a complete suite allowing SMEs in Europe to manage all HR processes with Payfit in an efficient, simple, and intuitive way without having to use other tools. Finally, we want to propose integration possibilities to our clients, giving them the choice of either using Payfit solutions or other solutions.