Building a startup: from darkness to light (may – september 2013)


Last 5 months were so busy that I didn’t find enough time to write on this blog. As you understand, I put almost all my energy in Cozy Cloud. During that time we get our first traction: we were selected in startup events, to be a part of an accelerator, we were pressed and most of all our community get significantly bigger. Here is the story.

Startup competitions

We were selected to compete in several startup contests: Web2Day, LeWeb, ICT… That was a good thing: it offered us feedback on how to build our pitch for investors. The drawback was that it kept us from improving our product or making our user/customer research.
NB: Another good thing is that when you are looking for collaborators of any kind, it brings you credibility. People pay more attention and are more likely to do things with you.


After a publication in LinuxFR, some Mozilla community members spotted us and encouraged us to send our candidature to the Mozilla WebFWD accelerator. So we did! And guess what? we were selected! Since two months they have provided us great advice about how to reach our potential customer, how to understand and please them, how to build a great pitch and how to manage finance. With their insights, we decided to pivot a little bit to reach our audience quicker: we are now building partnerships with hardware manufacturers and hosting companies.

Press and community

We were featured in several news plaforms (Wired, LeMonde, LeFigaro…) and blogs (among them the famous Korben). Previously we talked about credibility, this is another way to get some!  This way people know more about us, we found new supporters.

The other cool thing is that it brought thousands of visitors to our website. People subscribed to our waiting-list and started to use and install our software, some of them even wrote new blog posts about it. Our IRC channel get crowded and we received several contributions on the source code (bug reports and pull requests).

New applications

Last but not least, two new applications joined the gang:
Kyou  A tool to build analytic from data stored in your Cozy. It’s a very interesting usage because it’s the first app that takes advantages of data from other application.
Cozic A friendly audio player (let’s say it’s a winamp/itunes in your cloud).


These last five months made us found a firm ground in the start-up ecosystem. We drew the first shape of our community. We learnt to understand our customer better and improved the platform a lot. It was really exhausting but the experience was awesome.  Now we still have to improve the product (the platform and its website are ok but we still have to improve the applications that run on top of Cozy).

NB : I have tried to apply the 37 signals principles. They argue that working too much is a little bit acting like a hero: you think you can always solve more problem by building new things and work more. This could become an issue because it leads to extra work for the other team members (who are already quite busy) and it prevents you to find the more elegant solution (when you are in hero mode you use brute force instead of thinking). It looks like I failed deeply at avoiding the hero mode. I worked a huge amount of hours and produced a lot of stuff. If you ask why? I think that I simply didn’t know how to handle outside requirements: startup events, community growth, press publication, clients and a productive team. When someone ask you something you can say no. But when an unique opportunity happened or when people around you accelerate the process very quickly, it’s very hard to say no. Getting lean looks easy on the paper but trust me, it’s definitely harder when things become real.


My LSM (RMLL) 2013 Talks Videos: Personal data, Newebe, Cozy Cloud (Talk Tips Included)

Here are the videos of three talks I gave to RMLL (the one I mentioned in my previous post):

I hope you will enjoy it. An by the way here are some tips I could give to you by seeing my mistakes:

  1. Configure your screen to have the same output on your laptop and on the projector. Having to turn back to see what is displayed is annoying and not good to keep your audience attention.
  2. Don’t expect to have enough room to move. I read advice telling you to “fill” the stage, but most of the times, the organizers require that you stay at the same place (technical reasons).
  3. Find a way to add more interactions with the audience in the middle of the talk. Two times I refused questions that was asked before the end. I made a bad decision, that could have bring more attention to what I said.
  4. Be aware of what your arms do. I’m not telling you to work hard on your body language, I’m just saying that you should avoid nervous moves like the one I did on The Cozy Cloud presentation (see my right arm!)
  5. Speak loud, it’s easier to modulate your voice ton and make your talk less monotonous. This is a point I really need to work on.
  6. Identify on which part of your talk you hesitate the most and work on them instead of repeating the ones you already say well.
  7. Repeat as much as you can.
  8. Smile 🙂

About what I’m satisified :

  • I can’t explain why but my talk duration is always ok: it always fit with what is required by the organizers. I think it comes from the fact that I divide all my talks in three parts nevertheless I still don’t know why it works well every time.
  • Slide designs are ok: simple, easy to read with good illustration.
  • Talk structure: I think the message is well given. What I do for that:
  1. Problem/context explanation before the main subject description
  2. Simple ideas
  3. I repeat the main points all along the talk.
  4. I avoid some aspects of the subject (you can’t tell everything in a talk).

That’s all for this post. I would be glad to hear your feedback about my talk and ideas about how to improve them. Feel free to comment this post!

Talk @ LSM 2013 (RMLL)

In two days I will take part of the Libre Software Meeting 2013. I will give three talks on monday and tuesday. I hope to meet you there and have fun time discussing about the future of the web! Here are the subjects and schedules:

What happens when you are selected for LeWeb Startup Competition?

One month ago, Cozy Cloud was selected for LeWeb London Startup Competition semifinals. That was quite a good surprise and we were not ready yet to compete at such kind of event. So last month was mainly dedicated to be ready to reach the final stage. Being selected there allowed us to pitch in front of big VCs and have a stand to showcase our product. That was a great opportunity to make our funding search really easier. The aim of this blog post is to share with you how things went and what changed for our company.

Before the competition

Being selected is already a great achievement: only 16 start-ups are selected among 400. So after more than one year of work, that was really good to be considered by our peers as a competitive start-up. Moreover, that was a great motivator  for the whole team.

Social Proof
Once you are selected to LeWeb, a lot of people from the start-up ecosystem don’t consider you the same way. When you speak to them, they listen more carefully than before. That’s a little bit sad but it seems that the way things work.

We wanted to reach our community and have a little bit traction before pitching. So we made announcements about our product on two Hacker websites: LinuxFr (close to our users) and Hackernews (less focused but wider audience). They both worked pretty well (1450 clicks for the first one and 59 up-votes / 3000 visitors for the second one) and our main Github repository reached the 100 stars. Unfortunately, that was not enough to put it in a pitch, but that brought new users to Cozy Cloud. So during two weeks, we helped people on our IRC channel to setup Cozy Cloud on their home servers. With that feedback we improved our installation script, created a trouble shooting section and learned what to improve next. Then we changed the home page of our website to make it more appealing and send a newsletter. Being selected urged us to get closer to our users.

Following this announcements, someone from Mozilla France endorsed us for being candidate to WebFWD, the Mozilla accelerator. Their values perfectly match ours and they have great things to teach about community management and product marketing. That was quite a good news for us.

Pitch Building
At LeWeb the three best startups are mostly selected on their pitch. So we had to prepare ours. We were not ready yet for that. So, to practice our pitch, the best thing to do was to meet… investors! That’s why we started our fund raising search. We also participated to a smaller Startup Competition called Web2day. For this one we were not ready at all but we learned that we were too much focused on the product presentation and were not ready for Q&A about our financial projections. Fortunately our meeting with investors encouraged us to perfect our executive summary and becoming fully ready!

The competition


The competition was in London. We had a lot of pressure on us but The City was very sunny and we were glad to discover the Center Hall Westminster, a marvelous place. The entry tickets were offered to competitors, that was a good and cheap way to meet people that would be unreachable usually.

NB : One very annoying thing is that we were forced to wear an unmovable bracelet that no one checked. The user experience for a 1400 pounds ticket was probably awful for other participants. Except of that, the organization was close to perfect.

Once there, things became serious. The first day was a mix of stress and networking. LeWeb offered small stands for competitors so we pitched the whole day to everyone who came to us: bloggers, students, journalists and VCs. Moreover, LeWeb organized 15 minutes face-to-face meetings with VCs or start-up experts during the two days. We take advantage of that to meet several investors (between them the famous David Hornik of August Capital).

About the competition, we had to be present early in the morning to listen to the introduction. The principle was simple. Every start-ups were pitching during the whole day. The best three pitches went to finals. So we waited with a growing pressure until our turn. We chose to not listen to other competitors and prefered to stay at our stand. At noon, we had to check the stage to see if our computer and our mobile will be properly displayed. At 2pm TechCrunch recorded short pitches from all the competitors. At 3pm Benjamin, my partner and CEO met an investor. And at 3:45pm we were at last invited into the backstage : a lovely room with a private bar. There conference speakers were working on their talk, other competitors came back from the stage and some others were waiting like us. Then, around 4pm we reached the semifinal stage. Even if it’s a small stage that’s still impressive and even more when we saw the jury is made of two people from Accel Partners a big fund that invests mainly in companies that could see Cozy Cloud as a competitor (if it  becomes big).

Then we started our pitch, I didn’t talk, I was just there to manage the demo. Benjamin was  a little bit stressed but that was ok, hi s talk was clear and he made the audience smiled several times. The demo time came, it works pretty well until the mobile part. The application crashed (in fact I was signed out but because of the browser cache, I didn’t understand what was happening). As I learned with playing music, we had to go on like if everything was ok. So after a short and weird break we go on: the rest of the demo worked correctly and Benjamin finished his pitch. Then came the Q&A session. The jury make like they didn’t understand our value proposition (centralizing data around user to give him new capabilities, make him more productive and take smarter decisions). So Benjamin explained again what he said but without bringing new elements. That was not bad, but that was not really what was expected. The pitch ended, we were not satisfied by our performance. After that we were exhausted, one hour later we left the event and decided to go back to our beds.

The second day we were more relaxed, we met a lot of people again. Around 11am they announced secretly the finalists. We understood quickly that we were not selected. That was a little bit frustrating but we were not surprised due to our previous performance. Whatever the day went well and we met a lot of people. Among them, Philippe Raynaud CEO of Audio Fanzine, a boostrapped company (1M+ euros of annual income). Then we went together to see the finals. The stage was really impressive. Our friends from PricingAssistant won, we were quite happy for them. After that we left to enjoy some fish and chips!


After the competition and final thoughs

Now let’s talk about immediate results. Even if we didn’t win, being part of LeWeb brought to us several very nice “goodies”:

  • Our website was listed on websites with good page rank;
  • We get a little mention on TechCrunch;
  • We were featured on a well known French blog with 3000 feed subscriptions;
  • We had meeting with 5 VCs (early and late stage funding);
  • We got several tweets from accounts with 30 000+ followers;
  • It was like an intensive training of two days to improve our pitch and clarifiy our value proposition.

To conclude, I would say that LeWeb, by pushing us forward, was a great opportunity to reach new people. The motivation and the trust given by this event helped us to go one step higher. In other words, LeWeb was the kick in the pants we needed to reach the next level.


last photo credits: Jessica Bernard

In Bruges

In January, I spent two days in Bruges, a lovely city in Belgium. Due to its numerous canals, some call this town The Venice of The North. I have never been in Venice, whatever I had a pleasant time there, in a snowy version of Bruges.


The Market Place, one of the main place of the city.There are two frie sellers who make daily battles where the winner is the one who sells the more fries.


Windmills, there are several of them around the city. They make a nice landscape.


A frozen canal, by night.

Building a startup: beta stage (december – april 2013)

Last months were very busy at Cozy Cloud: we started our new beta program, we hired new members, we communicated more widely and we worked hard on the product. Here are the details of what happened.

beta stage

We previously started a beta stage that would be better called an alpha stage: only a few close friends were able to test our product (named Cozy). In december we decided to make things wider. This time we took hundreds mails from the list of people who subscribed on our website and we mount up a Cozy for each of them and send an email about what was happening. We were very excited by that and it was time to battle-hardened all our stack (infrastructure and product).

hard times


Unfortunately, four bad news happened to us after our beta launch:

  • Things didn’t work as expected. We had very few feedback. Almost noone wrote to us. Some people didn’t even use their Cozy, despite the mails we sent to them. We learnt one thing : the capabilities of the product were too poor. Even if people knew the potential (it’s an extensible product) of our platform they didn’t get interested in it… except a very few ones who gave us some hints on what was wrong, what was not working.
  • We met a lot of problems with our hosting infrastructure. Providing hosting services is not as easy as you could imagine. Even with 100 instances we had to set up automatic backup, precise monitoring and look for optimizations. And we are still working on the provisioning and logging aspects. It’s like we have to be able to scale early.
  • We wanted to be part of Le Camping, a kind of YCombinator for French startups (we are located in France). We spent a lot of energy in our application material. We rewrite most part of the website, we changed our linkedin accounts, we spent days to record the required video. But that was not what they expected: we were busted in the first round.
  • We also send applications to have a stand and talks at FOSDEM. Our application was refused too.

This clearly brought us some doubts about the value proposition of Cozy and the feasibility of providing clean hosting services. Fortunately, people that follows us since the beginning kept encouraging us and help us with bug reports, feedback or  developments. That maintain our faith in what we do. So we decided to go on and find more workforce to help us.

NB: The reasons of why we get rejected surprised us: The Camping said that we need an experienced marketer or a first round of fundraising. FOSDEM told us that we should have a fast growing community. For both, we candidate to them to find help to do what they ask for…

new team

We made several recruitments : 3 developers (interns) and one marketer. That was a great move : they all performed pretty well and fix broken things. Today, they are still improving the product,  and they are bringing a lot of new ideas.  Moreover, what we learnt from our previous intern session, made it easy to set up our new organization (see my recently posted slides). All the materials we wrote before (documentations, tutorials, public repository…) helped them to get operational very quickly.



Then we started to think marketing seriously. First we improved the overall aspect of the product : we redesigned the user interface. Second,we improved the tools provided to Cozy external contributors to help them to set up their development environment.

Then, we applied the principle of content marketing : you produce great content (blog and newsletters) to give a good reason to people to come. We  also contacted blogger in our field (self-hosting). That’s important to reach the community to let them know that we are doing something that could interest them. We learnt from that we should separate more the project aspect (Cozy is open source) from the corporate aspect. Then we get more feedback on our app and our installer. So we could figure on what to focus on. We also received a lot of encouragements which were very important for our moral.


So we learned the hard way two things: that’s not because you are able to convince people of your field (self-hosting) that you will be able to convince people from another one even if it’s a close one (innovative people). The technical part should never be underestimated too. We had a lot of experience in IT and we get still astonished by the required amount of work. We also tried to think simple all the time, but adding a lot of simple things make a complex thing. We also learnt that bringing new people to the project and receiving positive feedback give a lot of energy. Building a startup is a long journey and moral must be kept on top all the time.

Next months for Cozy will be dedicated to our contract with our partner (FING), to growth hacking and to fund raising. We hope you will hear about us very soon and not only on this blog!

A startup with no office, hipster tools and open source products

Talk performed at LyonJS, April 2013