The year is 2017 AD. The whole web is occupied by centralized services. Well… Not entirely. One small village of indomitable free software lovers still holds out against the invaders…

This is how the French association Framasoft presents itself. They have achieved something remarkable in the last two and a half years: Framasoft offers anyone free, libre and open-source alternatives to services like Google Docs, Dropbox, Skype, Slack, Facebook Groups, Doodle. For free. Framasoft is hosting roundabout 30 services ranging from Etherpad, Nextcloud, Jitsi, Mattermost, Loomio and Studs used by hundreds of thousands of users every month (here is an quick overview of all services Framasoft offers right now). They did spend 400.000 Euros in the last two and half years on building this – most of the money came from donations.

There is much to learn from Framasoft on how to make free, libre and open-source software popular and accessible. There are many great alternatives – but you need to help people to use, understand and love them. To make it as easy as possible to use libre alternatives (here is a great talk on this at Fosdem).

The Framasoft team shares their insights in the interview below – answered collaboratively in an etherpad by six persons (the illustration above was done by Simon « Gee » Giraudot, I recut it, it is freely reusable under Creative Commons BySA 4.0).

Why do you offer free services for anyone based on free software?

From its early age, Framasoft’s approach was– and still is – very pragmatic: we do not intend to convert people to the Libre as to a theoretical cause. We provide tools and services that you can use in your everyday environment – even on a proprietary OS! And from there on, we hope we can help people think by themselves, about their privacy, about their data, about the control they are entitled to claim on their digital lives.

Framasoft started 15 years ago, with teachers who created a directory website for free libre open source software (FLOSS), so that they could share the software with their friends and colleagues. And we kept on going along this path and proposing practical projects to bring more free libre software (DVDs, USB-Keys and so on) and culture (Blog, translations, publishing house…) to Mr-Mrs Everyone. Because we are (or were) Mr-Mrs Everyone: we still have a minority of tech-savvy hardcore developers among our midst, and we try to act as a “missing link” between this world and the widespread audience.

A turn point was 2011, when we started hosting our first Etherpad instance (the former branch). It was really powerful: as soon as we showed people how to collaborate on writing a text in real-time, online and without opening an account, they were amazed. We went on with hosting an Ethercalc instance (collaborative spreadsheets) and a now-homebrewed Doodle alternative named Framadate.

Nowadays, our association gathers people from all over France who seldom meet in the same physical space. So we have been the first to use our collaborative online services. As a matter of fact, we often started hosting them for our own private needs before making them publicly available.

But it’s only in October, 2014 that we launched our “Let’s De-Google-ify the Internet” program and made the bet that over the next three years we would host more than 30 Free-Libre and ethical alternatives to big-data services.

What do you offer that is unique compared to other web services or hosting providers?

It can all be summarized in one word: trust. We try to offer the best conditions for users to trust us with their data, thus giving them the questions everyone should ask themselves when using a hosted solution. To us the conditions of this trust stand on a few non-negotiable points:

  1. Exclusive use of Free Software (as in free speech), for both the services we provide and our system administration. Their source code being open, everyone can audit them and therefore have trust in them. It also induces the use of open formats, which is key for interoperability and importing/exporting one’s data.
  2. A transparent and data-friendly economic model. We chose to finance our services on user’s donations, allowing us to claim that we don’t have any financial incentive in collecting data and digital lives. Note that other economics models can also be data-friendly, but our model fit’s our goals. We aim to be as transparent as we can: our Terms of Services are user friendly.  We edited a “tl;dr”-version which only takes five minutes to acknowledge and is legally binding. Our charter describes the values we defend, we publish our administrative, technical and financial information and are always present to discuss and explain further all these information.
  3. Net and social neutrality: We respect and try to protect net neutrality. Moreover, our economic model allows us to offer free services (as in free beer), so there are no differences in the services you get based on your income. Last, as we don’t need to know who you are. We don’t want nor need to get a “user profile” or a “social graph” from you. We don’t (and can’t, and won’t) discriminate the service you get based on your (non-)gender, skin color, origin, orientations, political preferences and the like.
  4. Solidarity and education: We try to facilitate as much as we can the use of our services by providing startpages, quick-start guides, documentation, self-hosting tutorials, support (both individual and through our Frequently Asked Questions page), users forums. We both provide and ask for contributions on the software we use and their documentation, so our users community can also become a contributors community.
  5. Decentralization: Our services are offered as a kind of proof of concept. They demonstrate that FLOSS can be an alternative to Big Data’s services, and that it is possible for the users to keep the control over their digital lives. Users can try different software and use it as much as they need, and (if and when they are ready) leave our services for even more digital independence, because they are able to host the software they need themselves. We often provide guidance for those who wants to cooperatively or self-host these services and migrate from our services to their own servers.

What are the benefits of doing this as an association instead of a cooperative for example?

Not dealing with clients :p!  OK, behind that poor-tasted-joke lies some kind of truth: we want to empower people in their digital lives, and we feel we won’t be able to do so if we place them in a passive customer role.

On a general point of view, our activities are non-commercial and we intend to keep them on a small scale and to maintain a democratic balance between the employees and the volunteers in the association. By doing so, our relationship won’t become anonymous and everyone is involved in the global project. Like in a cooperative, 1 person = 1 voice

Since we don’t sell our services (nor won’t we one day propose “premium fees” and such), we are dealing with users. It changes everything: people are more understanding, less demanding. They know we do our best (we would settle for nothing less) and they can accept when our best isn’t enough, when there is some downtime, for example.

We don’t aim to host and concentrate as many people (and people’s data) as possible. Our goal is to demonstrate that the Free-Libre world has already worked on alternatives to GAFAM’s services (Google, Apple, Facebook, Amazon, Microsoft). We want people to come and try services with us. That’s why we contribute through user & self-hosting documentation, translations into French, presentation and a (tiny bit of) design, some code. But our final goal is that users leave our services because they liked to use them and were so convinced that they started self-hosting them. Or that they have at least found a local ethical Free-Libre service provider – decentralization of data is important and forgotten too often.

Those purposes make us not compatible with a model based on profit. Being a non-profit looking only for self-sustenance allows us to explore, to take time to educate people and to experiment – and with experiments come failures, which are great lessons.

Last, being an association, under the French status of “Association de loi 1901” allows our organization to be officially recognized of general interest. This benefits directly to our more than 2,000 donators who can get a tax relief of a third of the amount they gave us. Thus it’s an incentive for them to keep on supporting us. And donations are 95% of our revenue stream and the basis to pay for our 6 employees, our servers, and such.

How much of the collective work of framasoft goes into the free services?

Nowadays: most of it!

Before the “Let’s De-Google-ify the Internet” project, we had some balance between free software projects (the directory, USB keys, DVDs, etc.), free culture projects (blog, translations, publishing free-libre novels, comics and handbooks, etc.), free-libre services (pad, calc and doodle-like), and the life of the association.

Now, our translation group and our publishing house are still very active, our free software directory has been completely re-modeled, but half to three-quarters of our energy are focused on the 30+ services we are hosting, and the few alternatives left we intend to complete.

What we didn’t realise before starting this project, is how much support and communication (public relations) we would have to provide. And it’s important, as we both want to help people adopt free-libre services and to care about the stakes of data silos such as Google, Apple, Facebook, Amazon and Microsoft (the infamous “GAFAM”).

Even though, we still take and find time to share our experience so it can be reproduced and adapted, mainly through a network of local ethical free-libre services hoster we initiated: CHATONS (French for “Kittens”).

How do you measure the performance of framasoft – what are your goal, what are your key performance indicators?

When we started the “Let’s De-Goole-ify the Internet” project, in October 2014, it almost seemed like a fools’ errand, an impossible bet we took on ourselves. So, at first we didn’t set any other goals than trying to do our best to achieve what we had announced: To inform the audience of the stakes of data-concentration, to show that FLOSS can provide ethical alternatives, and to help those who could to achieve digital independence.

Two and a half year later, many people trusted us and gave us the means to achieve what at the time could have looked as a moonshot project. Now, our main goal is to share this experience so other hosters can reproduce, adapt and use it. We have initiated a network of French-speaking ethical sevice hosters (named CHATONS), and now we are trying to see if such a network can be expanded or reproduced in other languages and countries.

As for the measurements, we have of course some quantitative indicators: for instance, we get more than 2 million visits per month on our network of websites and services, which means that hundreds of thousands of people are at least trying to change their digital habits every month.

Since we don’t want to collect detailed users’ data, we don’t have qualitative performance indicators. So, right now, we can only share some quite informal feedback.
Until now, we mainly get feedback by interacting with people, online via social networks, comments, e-mails and so on, and in real life because we take part to roundabout 100 events every year. And it is a great pleasure to get thanked by people we helped get rid of the services of Google and the other big players of the Internet. It allows us to think that what we do is useful and that we are doing it in the right way.
Of course, a key indicator is the volume of donations, our main resource. It has been significantly growing with our current “De-google-ify” campaign, which allowed us to hire and to grow from 2 to 6 employees in the last three years. If we weren’t responding to a need, people would stop giving, wouldn’t they? And if they weren’t satisfied, they would give less.

This De-google-ify campaign has brought more light to our little association: Now French journalists come to us from time to time and ask for our opinion. Some media have mentioned our services, not only the specialized and tech press, but also national newspapers. We sometimes are invited to talk on the radio or on television. Meanwhile, known figures from the free software movement are supporting our methods, which is an important recognition because this is where our services are rooted. We didn’t want to respond to a few invitations by politicians (for instance by a political group at the Assemblée nationale) and explained that our association’s goals were not political and that we were addressing people from the civil society, regardless of their political position.

When we met for our general annual meeting last January, we were thinking that at this point, we needed to know better our users and their expectations. So we decided that we should launch a survey to find out: “Who are you, Framasoft services users?”. It could help us improve our services to better fit the people who actually use them and we expect to have a better and more reliable insight next year (barring unforeseen circumstances,)

How did you key performance indicators develop over the last years?

Great as our project thrived and users adopted more and more our services.
In 2014, we participated in around 30 events, and our support team got about 10 messages a day. In 2016, we attended more than 100 events (conferences, workshops, debates and so on) and our support team processed more than 100 tickets a day.
We can’t compare visitors numbers, since we changed this tool during this time, settling on a self-hosted Piwik instance to analyze visits on our website.
Everyone can check to get some metrics (but unfortunately it’s only in French for now)

What are your aspirations, what are your promises to users in regard to availability, security, service for framasoft free services? Or asked differently: Would you recommend an association or local volunteer groups to use your services for organizing themselves and getting work done?

In our charter, we pledge to give our best efforts, but not to get the best results (at least, not at any cost).
Practically, it translates into doing everything we can for the best uptime. We hired a full-time system administrator, help by other tech-savvy employees and volunteers. 100% uptime isn’t a sensible goal for a non-profit like us, the cost would be overwhelming. When our services fail (need to reboot), it’s usually for less than 15 minutes. If there is a bigger issue, we inform our users (through social networks and a dedicated website), taking this chance to educate them about what it means to host a service and administrate a system. We only had one major incident in January 2015, when our ethercalc instance became unstable after an update. It was closed for a fortnight, but when we re-opened it, all the user data were there and safe.
We take every security step we can for our users, with multiple backups of data at different geographical places.
Service is very important to us: our support team respond to each and everyone, as it is the occasion to know our users and to try and help them on their way to digital independence. Of course we provide a Frequently Asked Question page, but you will never get a reply like “read the fucking manual” (RTFM) from us ;).

Any association and local group is welcome to use our services, we know how they can be a very important vector of digital education, awareness and empowerment. That being said, we would advise them to consider our services as a “first step”: use it, try anything you think would help your self-organization, and when you know the tools you need, try to take the second step and host them for your group. We’ll be here to provide help and share our knowledge!

Do you have plans and activities to educate interested users in the configuration and usage of tools like Mattermost and the like? The learning curve is steep for people who never worked with other tools besides E-Mail and Office. How can this change?

Our approach is a progressive step by step enticement to change. We consider it to be one successful step if people adopt LibreOffice instead of MS office and Firefox instead of Chrome. We don’t invite people to jump immediately into the full-fledged Libre world. We believe in suggesting alternatives that fit the needs of users. Rather simple tools like Framapad and Framadate are good examples: They have huge success because they need almost no initiation. The success proves that this one possible way to change usage. We provide two kinds of documentation for our alternative services: One for the mainstream user and one for more tech-savvy people that can install services themselves. It is crucial to swarm/spread.

Pouhiou, the PR person at Framasoft, had never used Slack (or something similar) before Luc (our sys-admin) proposed a Mattermost instance. He obviously proposed it as a side-feature of our Gitlab instance, aimed mainly for developers. But when I tried it, I claimed in surprise “Wow! This can be THE alternative to Facebook Groups!” And we all realized the potential of this tool, aimed to developers, but easy to apprehend by most.
Lots of people already know FB Groups, that can be public, confidential or private, where you have to join with your account, where you can talk both in real time or with delay, where you can share pictures and links… Well, Mattermost is quite the same (just better with its search and text-formatting features).
That is how we decided to aim and present it to the widespread audience: a Facebook Groups alternative. We translated user-documentation into French, created a detailed example for our blog presentation to show the power of this tool.
We really think that being a mixed crowd with software developers and enthusiasts help us to reach the everyday person. Because, for most of us, we are this person (or used to be not so long ago).

Whom are you building the free services on framasoft for?

We are working for two categories of persons:

  1.  A clear majority of users who we call the Dupuis-Morizeau, the average French family with zero to some knowledge about computers and the Internet. We consider it is our main mission to bring them step by step a little more freedom in their digital usage. They are the people we wish to take advantage of our degooglisons campaign in term of awareness, new habits, discovering free software, the importance of privacy, etc.
  2. A more limited group of tech-savvy people pertaining to the FLOSS community that can bring knowledge and technical tools to the population. They can install applications and services on their own servers. That is why each service we offer to the public is released with an installation tutorial to spread alternative tools.
    CHATONS is the name of our other recent initiative for decentralization.
    The name itself is a play on words (it means Kittens in French), It is an acronym for Collective of Hosters which are alternative, transparent, open, neutral and solidaric. The project aims to bring together players who offer free, ethical, decentralized and solidarity-based online services in order to enable users to find – quickly – alternatives to Google products (among others) but respectful of their data and privacy. Since Oct ’16 the first wave of “CHATONS” is 20 hosters strong, and more are coming. The validation of wannabe CHATONS is made via the collective itself on the basis of a manifesto and charter that were elaborated collectively.

What are framasofts sources of revenue and what are spending on?

Note that 2016’s books are being revised by an independent account commissary (a legal obligation for us that we love, as it enforces transparency), so we’ll talk based on 2015’s… but it will roughly be the same shares.
90% of our income in 2015 (~178,000 €) came from donations from our 2,000+ supporters. The rest comes from state employment aid (5%) and selling things (such as books and goodies – 5%).
We mostly spend it on paying our six employees (70%), because we cannot depend on volunteer’s’ energy if we want to provide reliable services and user support. Then we spent it on conventions, conferences and meeting fees (10%) as “evangelization” is a big part of our work. Then comes technical fees (only 6%!), operating costs (6%, too) and the rest is divided between printed communication, supplies and bank fees.
What never cease to amaze us, is that with ~400,000€ over the last two and a half years, we’ve been able to host and maintain roundabout 30 services (along with our ~20 other projects), and change the digital customs of hundreds of thousands of users each month! It’s about the price of 50 meters of a highway!

If people like your work: How can they ensure that Framasoft flourishes?

We accept donations which are almost our unique source of money – donations are welcome here. But we are also glad when code contributors join us because our tech team is only 3 people strong. The same with language and translation contributors, as we wish to spread our experience and services to other languages.

Are you planning to offer localized versions of Framasoft’s free services in Germany?

Well, they are already offered over the Internet to everyone, but only a few services are already internationalized enough to be translated into various languages. We need some effort and contributors and an English version first since English is the lingua franca nowadays. All of our services are on a Git repo. When there are “locale” files to be translated, the German-speaking community is welcome to do it!
Note that the majority of our services are based on well-established free software which already is available in English and German. It’s mainly our design modifications and the quick-start guides that are not translated. In case German people are interested in having an English version of a page/tutorial/service/interface which is currently French-only, please tell us and we will do our best to facilitate the transition to a German-speaking resource.