Three lessons from Parrot’s saga

Photo Henri Seydoux / Chairman and CEO, Parrot / September 28th, 2016


L’Ecole de Paris

Parrot's adventure began in 1994 with voice-controlled organizers. It never became a huge commercial success, but the technology involved was recycled into hands-free car kits, and Parrot quickly became a major player in this market. After that, it started manufacturing communication products for the general public. In 2010, a toy drone controlled by smartphone was launched; this success made Parrot one of the leading manufacturers of consumer drones. Three lessons can be drawn from this experience.

Parrot did not start out of nothing. Its story actually started with my dropping out of the university, back in ’78. I was interested in writing: I became a journalist. In an interview with Roland Moreno, the man who invented the chip card, he told me that he too had worked for a media outlet, before switching to technology. He encouraged me to learn about computing: “It’s very easy, you’ll see. Read the books, and, if you understand arithmetic, you’ll find programming quite easy. In fact, with all these new computers from Apple, programming has become child’s play.

This was in 1979. I bought my first computer and started programming. I discovered that in fact, it wasn’t writing that I was interested in, but writing computer programs! I became a geek before the word even existed, and I quickly realized that I could make good money out of this activity. I had no trouble finding customers, and it often took me only a night to write the program they needed.

In the early 1980s, a friend reported that someone in the Computation Center (Centre de calcul) of École des mines (now Mines ParisTech) was looking for programmers. It was François Mizzi, who was putting together a startup with funding provided by the investment fund of Elf Aquitaine. His project was to create a touch-screen computer that fit in your pocket, a sort of foreshadowing of current smartphones. He hired me. I was 20, and I joined the team, most of whose members came out of top engineering schools such as Mines or Supelec.

In the next 14 years I joined or founded several companies – not all of them in tech. In 1982, I was hired as a software developer by SSCI, before joining Micro Archi. In 1986, I created BSCA, a company that specialized in rendering 3D computer images, and served as CEO until 1990. In 1991, along with three other partners, I founded Christian Louboutin, a luxury shoes company.

Launching Parrot
In 1994, I created my own start-up, Parrot, for which I secured a loan of 2 million francs (€300,000) from Agence nationale de valorisation de la recherche (the French innovation agency), repayable after four years, depending on the results. My ambition was to design a small object that would be a cross between a digital recorder (a concept which, at the time, did not exist) and an agenda, with a voice recognition system to take notes and classify them.

After 18 months, I sent a funding request to Sofinnova, a venture capital fund. The people I spoke to found my project a bit weird, but they liked atypical profiles like mine. In addition, I explained in my file that I wanted to find applications for the general public, not for businesses. As it turned out, Bill Gates had just announced that Microsoft would focus on the mainstream market. As venture capitalists are always on the lookout for new trends, Sofinnova decided to invest 2.5 million francs in my project at first, then the same amount should the prototype be successful.

Everything went well until the market launch, which was a failure. To be honest, the unit wasn’t working as well as expected: I was mainly a software developer. Electronics were not really my specialty.

While talking with a seller from the FNAC, I discovered that the only buyers of this product were blind and partially-sighted people and that they were very excited. So I met members of the Valentin Haüy association, who explained that this organizer changed the lives of all those who felt stranded when it came to reading a phone number. I then advertised the product to similar associations in several countries, and I sold my product everywhere. But it still remained a niche market.

At that time, cell phones were beginning to develop. In France, it was the Bi-Bop, before the appearance of the first version of GSM. A friend of mine told me that the Belgian subsidiary of Ericsson, the world’s leading cell phone manufacturer at the time, was looking for new applications. So I contacted the manager to offer him to integrate my voice recognition software in their phone. He refused because it seemed too complicated to program. However, knowing that a new Swedish law forbade the use of hand-held phones while driving, he was interested in an application for the automotive environment. Now, the number-one enemy of speech recognition is noise: when the ratio between signal and noise is low, it doesn’t give good results. In a car, there’s a lot of noise!

I went on holiday with a stack of numbers in the American journal PNAS (Proceedings of the National Academy of Sciences), and I came across an article in which an expert explained how the human ear extracts speech from a background of noise. Back in Paris, I wrote a new program and requested an appointment with Ericsson. We left by car to test the device on the highway, while rolling deliberately on rumble strips: “If your system is working under these conditions, it’s a win!” Fortunately, the experiment was successful.

This happy event occurred in the nick of time, because my company was on the brink of bankruptcy. I managed to convince investors to dip into their pockets: mobile phones were booming, cars were the second place where people phoned and I had just landed a contract with Ericsson. I got a new funding of 18 million euros .

At Telecom Paris, I found a lab that agreed to help me develop speech recognition systems for car phones. Quickly, we encountered a major obstacle: Nokia, Ericsson and Motorola shared 80 percent of the market, and their devices lacked interoperability. We weren’t able to develop a system compatible with all three brands.

That’s when I received a visit from an engineer from Ericsson’s R&D department: “We just invented a system that connects to wireless phones. The code name is Bluetooth. Do think it could have a future?” It was exactly what we needed!

The development took us a while anyway, because we were only 12 at the time. Our capital was melting down… But we managed to create a Bluetooth system adapted to our speech recognition tool, which now became usable with any phone.

From disaster to success
Finally, the big day arrived: the launch of our Bluetooth hands-free system at an automotive show, in Frankfurt. It was on 9/11/2001… The show was deserted and closed the next day. I found myself with 30 employees and enough cash for three months. Our venture capitalist, who had already reinvested on three occasions, was no longer attending the boards. Parrot was on his blacklist.

Paradoxically, this disastrous situation made things better for us. The three major telecommunications companies abandoned all that was not their core profession, particularly car phones. Four months after 9/11, Parrot became profitable: we were the only company in the world to produce Bluetooth hands-free kits.

In 2002, we achieved a revenue of 4M€, which doubled every year thereafter.

In 2005, our revenue reached 30M€. I wanted to take the next step: sell our hands-free system directly to car manufacturers.

I reindustrialized all of our software and tried during the next three years to find an interested builder or OEM. I never succeeded! However, I met a Japanese manufacturer, Hitachi, who was seeking to develop European technology in Japan in the form of a joint venture. We opened a joint office and found our first customer: a Japanese manufacturer of car radios that was interested in our Bluetooth solution. Our second client was an American who offered equipment for navigation. This experience allowed me to measure the strength of American companies: three weeks after the presentation of our product, the contract was signed. I sent one of our engineers there and, six months later, the product was released!

Having signed the first two contracts for car radio projects, I was contacted by Pioneer, which had tried to design its own Bluetooth system, unsuccessfully. For many years, Parrot produced all of Pioneer’s Bluetooth devices.

Following this initial success, nearly all Japanese and German manufacturers adopted our products. Regretfully, I never managed to convince French brands to do the same. No one is a prophet in his own country…

In 2004, at a time when we realized a revenue of approximately 20M€, I received a call from Sofinnova: “Henri, it’s great, your business is profitable, we’re going to sell it!” I replied that it was out of question, but that I would find a solution, because I understood that Sofinnova needed to reap the fruit of their investment. The Internet bubble had burst, and the context was no longer favorable. It took more than a year to find an investor willing to take over the shares of Sofinnova, provided we allowed the company to go public.

So I took the leap and, contrary to what I had been told, I was extremely happy about it. The IPO, completed in 2006, allowed us to move from 25 million euros of equity to nearly 80 million. Knowing that in the meantime, our revenue had reached 160 million euros, it gave great strength to the company.

Moreover, I was free from venture capitalists, which are invaluable when you start but quickly become very invasive. They can work on everything, including matters beyond their control, and require you to go always faster, even if this puts you in danger. Once you’re listed on the stock market, it’s enough to explain to your shareholders what you want to do and, provided the explanation and application hold up, they’re free to buy and to sell, and everything goes well.

Internal start-ups
Hands-free kits were sold from $15 to $200  apiece. At the peak of activity, we were selling over ten million units per year, for a total of 50 million vehicles produced each year. In principle, this left us with a margin, but I knew Moore’s law and I also knew that what was sold for $20 on a given date would be worth only $10 some time later, then $2, then only a fraction of dollar. Legions of low-cost developers were drafting new software and, sooner or later, Bluetooth devices would earn us peanuts. It was an urgent need to think about new products.

The company had money and was not tied to venture capitalists. I was free to launch new projects and very well aware that it was the right time to do so, just after the IPO. Considering that the starting point for any of Parrot’s activity was the phone and that the sector was extremely dynamic, I decided to try to innovate in three major phone features: photo, music and video games. I have to recall that, at the time, the iPhone did not exist. We were still in the Nokia era. When I announced the project to the members of my board, they stared at me with wide eyes. But as main shareholder, I am the first one concerned by the risks I amdeciding to take… It’s the secret of my freedom.

To explore these possibilities, I created three internal startups. For each one, I chose one or two talented engineers with whom I got along well, and I teamed them with one or two engineers recruited from outside. I installed these three teams in a corner of the office and gave them the task of creating new products with two main prohibitions: no specifications, no market research. These principles are keys to innovation and sometimes still in force, today, at Parrot. A one-page roadmap is enough to glue software features together. We then conduct tests to see if it works. In general, it doesn’t work, and you have to start all over again. Gradually, we accumulate knowledge and sometimes, it ends up working.

Of course, there are many failures. Half of the startups I created within the company had to be closed. Five engineers, for example, worked during three years on a balloon project that had an accelerometer that served as support for video games. All children in my building participated in the testing of prototypes. After three years, we had to give up because it wasn’t successful. At another time, I dreamed of creating a musical instrument that could be controlled by a joystick. I worked during three years with a music professor from a French conservatory on this project, but it didn’t succeed either.

A worldwide success: the civil drone
The project of a toy drone piloted by a smartphone took us five years of work. At Parrot, many were against: “What’s wrong with granpa?” At that time, Parrot employed approximately 500 developers, half of which were very young, recent graduates. They were involved in signal processing and other technologies of this kind and couldn’t understand why their elders were focusing on a drone project that didn’t work. Our business model wasn’t very convincing either. Despite using phones’ “organ banks” as much as we could, the expected sale price reached over 300€. Representatives from Sony and Nintendo to whom I showed our prototypes dismissed the project: “You won’t sell a single unit. In the video games industry, you pretty much have to give away the hardware. It’s the software that generates profits.”

I was a little shaken up, but we stuck to the project until its finalization in January 2010, when we introduced our first drone at the Consumer Electronics Show, in Las Vegas. Our unit was fluttering inside a cage, and its main performance was to stabilize in the air, on a fixed point. It was a huge success! The drone was the main attraction of the show and was the subject of nearly 300 TV reports: we were featured on CNN, CBS, Sky News, ZDF, BBC, TF1, etc.

When I returned to Paris, our CFO asked me how many units I intended to manufacture: 50,000? No: “200,000!We still didn’t have the slightest idea of its true potential but within six months, we sold 120,000 drones. Overnight, this represented revenue of over 20M€. I decided to speed up the development of this activity and expand it to business applications by acquiring companies specializing in drones, including two startups from the Swiss Federal Institute of Technology in Lausanne: senseFly and Pix4D.

Today, our revenue reaches 330M€, including 200M€ in drones and 130M€ in the automotive sector. The automotive sector is in decline because car phones are “commoditized.” Drones, on the other hand, are in a full swing. During a fundraising campaign last year, we were able to raise 300M€, a sum which the Group needed to become a global leader in consumer and professional drones.


I draw three main lessons from this experience. First, it’s perfectly possible to create an industrial company in France. In fact, it’s even easier than ever. There are now investors in all development phases of a start-up. Whether you need 2 million, 5 million, 18 or 25: you can find them. An IPO is also much easier than it was 20 years ago.

The second major lesson is that we shouldn’t delude ourselves: high technology works in cycles, and you can’t expect to sell the same product for decades. When I started in car phones, Nokia was the world leader of the phone industry, the largest European capitalization. It ranked among the six or seven most recognized brands on the planet. Today, the company is gone, even though it had excellent engineers. All high-tech companies are fragile, and Parrot is no exception. Even Apple was on the verge of bankruptcy before the return of Steve Jobs… and nobody can say today where the company will be in ten years. It’s a juggling act: you can be very successful at one time and then be forced to start all over again.

The third lesson is that the software industry is fundamentally oriented towards B-to-C. Industrial customers and big clients are quite flighty. It’s very difficult to build something with them. It’s only with products for the general public that we can actually do business and, especially today, three main factors contribute.

Before, to sell a consumer product, you had to advertise it on television. Now, whatever products you wish to sell, you can communicate via YouTube or Google. Another major change: it’s now easy to manufacture industrial products in large volumes. Whether you need to make five, ten or 500,000 products, you will find your happiness in the Shenzhen area where about 25 percent of all electronic products are made. Finally, distribution has become much more accessible. Previously, it was the private turf of big electronic groups such as Sony or Philips. Now, major retailers offer many products made by startups: these are the ones that sell best and offer the greatest margin to distribution. Besides, you can also sell products online.

More on paristech review

By the author

  • Three lessons from Parrot’s sagaon September 28th, 2016