Matti Schneider Équipe permanente @matti_sg_fr Incubateur Numériques de Services Incubator Digital 

Matti Schneider @matti_sg_fr Digital Services Incubator Numériques Incubateur Core team member 

Media Partners Eco-System Partners Support Sponsors Organizer Thank you to AT Beirut We work in the heart of the State, at the Prime Minister task force for Modernisation of Public Action. Our aim is to spread the culture of digital innovation throughout the administration. You might have heard that France has a bit of a bureaucracy… and digital innovation is the very opposite of bureaucracy. It is all about delivery, disregarding established rules and practices. So how do we show what is production in such an environment? 

Startups State By incubating what we call “State Startups”. 

Startups State (not to scale) Usually, when I say “State Startups”, people hear “oxymoron”. I’m told the two entities produce two very different things. It’s true, and it’s legitimate: they both have to handle very different constraints. But they do have a common trait: they exist to create and manage services for a set of users. They simply do it with very different communication and interaction ways. And it happens that one of these ways is more common for citizens in 2015. How come it is so easy to buy a used computer from a foreign country, follow its trajectory across the globe and get a text half an hour before delivery… and so hard to know when my immigration card will be done? Which of those has the greatest impact on my life?? 

Startups State (not to scale) We hire people who have shown abilities to make digital products to decrease the gap between citizens’ demand and governmental offer. We insert them in the heart of the State with a clear mission: remove a specific friction between the administration and the citizens. A small, autonomous agile team is created, containing an “intrapreneur” within the administration. This team will have to deliver a first public, experimental version of its service within 6 months. A State Startup is a group with the culture of digital startups, but working entirely within public jurisdiction. It is not a separate private entity. This is critical in making our goal of transformation of administrative thinking achievable. 

Startups State (not to scale) By applying methods and values of the digital startups, the State becomes able to build digital products that are at least as good as their private counterparts. So, here I must pause for a second. I’ve been told there are very different profiles attending this session. If you’re used to agile production, nothing here will really blow your mind. As I said, it’s only about applying known practices in an environment that sorely needs them yet rejects them. If you already know the practices, the more interesting parts will be in the openness that our public position allows. If you discover agile, you will be able to see what it looks like. 

M E S - A I D E S . G O U V. F R O N L I N E C O M P U TAT I O N O F F R E N C H S O C I A L B E N E F I T S Here is a first example: a web application that helps citizens know which benefits they are eligible to. I am CTO of this State Startup. The base friction is that of “non-recours” to social benefits. That is, not claiming benefits people have a right to! In 2008, 80% of people eligible to financial help for health insurance (“ACS”) did not ask for it. 68% of low-income workers who were eligible to a complement of revenue did not ask for it. After several months of research, the answer was to consolidate all forms in a single claim file. However, many administrations are involved in giving the benefits. And each of them has different rules for giving them! So the final paper form ended up with a dozen sheets and thus a terrible user experience. That’s when we started incubating a State Startup. 

This application is available at mes-aides.gouv.fr since more than one year. 

We simply ask for a few basic information, such as your age and nationality. 

Same questions for your children and spouse, if you have any. 

You then give in some information about your housing conditions, and your income. 

And then, under 6 minutes for a first-time user, you get a list of all the benefits the family you described is eligible to, with computed amounts. How did we create this under six months, and improved over the last 12 months? 

2 weeks Iterations 2 weeks By making it public very early to collect feedback, and by continuously improving the product based on that feedback. Every two weeks, we deliver improvements in a fully transparent way. Transparent both in delivery (no user experience disruption, small incremental changes) and in roadmap: every task we’re working on is publicly visible, as are all the discussions that lead to the choices we make. 

2 weeks Iterations 2 weeks waffle.io/sgmap/mes-aides-api By making it public very early to collect feedback, and by continuously improving the product based on that feedback. Every two weeks, we deliver improvements in a fully transparent way. Transparent both in delivery (no user experience disruption, small incremental changes) and in roadmap: every task we’re working on is publicly visible, as are all the discussions that lead to the choices we make. 

2 weeks 2 weeks Iterations ? Indeed, the real question is: how do we decide what will be done in the coming fortnight? We, as a product team, make choices based on our workload, opportunities and partnerships. But also from information we collect from different stakeholders. 

Collaboration ? First of all, the end users. We collect their needs and uses through user testing. These user tests can be either made to challenge or validate hypotheses; or on a more exploratory setting, to discover possible improvements and new use-cases. That’s how we improve user experience. Obviously, we actively engage the administrations to participate in the evaluation of our computations. To help in this endeavor, we created an acceptance testing tool that allows one to create tests right from the end-user interface. Simply click a button at the end of a simulation to suggest a correction in the presented results. That’s how we improve correctness. Finally, we federate mediators. People who help citizens understand their rights are critical to our mission, as not everyone is able to use digital tools. However, our tool can empower humans to help other humans. We invite all interested mediators to a monthly meeting that we hold in a different town each time, to both collect feedback on their uses and disseminate best practices. 

Collaboration ? First of all, the end users. We collect their needs and uses through user testing. These user tests can be either made to challenge or validate hypotheses; or on a more exploratory setting, to discover possible improvements and new use-cases. That’s how we improve user experience. Obviously, we actively engage the administrations to participate in the evaluation of our computations. To help in this endeavor, we created an acceptance testing tool that allows one to create tests right from the end-user interface. Simply click a button at the end of a simulation to suggest a correction in the presented results. That’s how we improve correctness. Finally, we federate mediators. People who help citizens understand their rights are critical to our mission, as not everyone is able to use digital tools. However, our tool can empower humans to help other humans. We invite all interested mediators to a monthly meeting that we hold in a different town each time, to both collect feedback on their uses and disseminate best practices. 

Collaboration ? First of all, the end users. We collect their needs and uses through user testing. These user tests can be either made to challenge or validate hypotheses; or on a more exploratory setting, to discover possible improvements and new use-cases. That’s how we improve user experience. Obviously, we actively engage the administrations to participate in the evaluation of our computations. To help in this endeavor, we created an acceptance testing tool that allows one to create tests right from the end-user interface. Simply click a button at the end of a simulation to suggest a correction in the presented results. That’s how we improve correctness. Finally, we federate mediators. People who help citizens understand their rights are critical to our mission, as not everyone is able to use digital tools. However, our tool can empower humans to help other humans. We invite all interested mediators to a monthly meeting that we hold in a different town each time, to both collect feedback on their uses and disseminate best practices. 

Collaboration ? First of all, the end users. We collect their needs and uses through user testing. These user tests can be either made to challenge or validate hypotheses; or on a more exploratory setting, to discover possible improvements and new use-cases. That’s how we improve user experience. Obviously, we actively engage the administrations to participate in the evaluation of our computations. To help in this endeavor, we created an acceptance testing tool that allows one to create tests right from the end-user interface. Simply click a button at the end of a simulation to suggest a correction in the presented results. That’s how we improve correctness. Finally, we federate mediators. People who help citizens understand their rights are critical to our mission, as not everyone is able to use digital tools. However, our tool can empower humans to help other humans. We invite all interested mediators to a monthly meeting that we hold in a different town each time, to both collect feedback on their uses and disseminate best practices. 

Collaboration ? First of all, the end users. We collect their needs and uses through user testing. These user tests can be either made to challenge or validate hypotheses; or on a more exploratory setting, to discover possible improvements and new use-cases. That’s how we improve user experience. Obviously, we actively engage the administrations to participate in the evaluation of our computations. To help in this endeavor, we created an acceptance testing tool that allows one to create tests right from the end-user interface. Simply click a button at the end of a simulation to suggest a correction in the presented results. That’s how we improve correctness. Finally, we federate mediators. People who help citizens understand their rights are critical to our mission, as not everyone is able to use digital tools. However, our tool can empower humans to help other humans. We invite all interested mediators to a monthly meeting that we hold in a different town each time, to both collect feedback on their uses and disseminate best practices. 

Collaboration ? First of all, the end users. We collect their needs and uses through user testing. These user tests can be either made to challenge or validate hypotheses; or on a more exploratory setting, to discover possible improvements and new use-cases. That’s how we improve user experience. Obviously, we actively engage the administrations to participate in the evaluation of our computations. To help in this endeavor, we created an acceptance testing tool that allows one to create tests right from the end-user interface. Simply click a button at the end of a simulation to suggest a correction in the presented results. That’s how we improve correctness. Finally, we federate mediators. People who help citizens understand their rights are critical to our mission, as not everyone is able to use digital tools. However, our tool can empower humans to help other humans. We invite all interested mediators to a monthly meeting that we hold in a different town each time, to both collect feedback on their uses and disseminate best practices. 

Collaboration ? First of all, the end users. We collect their needs and uses through user testing. These user tests can be either made to challenge or validate hypotheses; or on a more exploratory setting, to discover possible improvements and new use-cases. That’s how we improve user experience. Obviously, we actively engage the administrations to participate in the evaluation of our computations. To help in this endeavor, we created an acceptance testing tool that allows one to create tests right from the end-user interface. Simply click a button at the end of a simulation to suggest a correction in the presented results. That’s how we improve correctness. Finally, we federate mediators. People who help citizens understand their rights are critical to our mission, as not everyone is able to use digital tools. However, our tool can empower humans to help other humans. We invite all interested mediators to a monthly meeting that we hold in a different town each time, to both collect feedback on their uses and disseminate best practices. 

embauche.sgmap.fr Another example, with the same ways on a different topic. Here, we created an embeddable widget that computes the cost of hiring an employee, with all taxes and refunds included. I was responsible for the discovery and alpha phases of this Startup. The initial friction was to simplify the edition of payrolls. However, the possible solution were too risky to deploy under six months. We thus pivoted on a first step that would more easily get interest: cost of hiring. 

This widget is very simple so it can be embedded in any layout and made available to employers in their ecosystem. We want to solve a friction, not to create new official portals with partial information. 

Here is the widget embedded in the website of a private startup that helps SMEs make their business plan. 

We produced the first public version of this tool under 6 months and one full-time person. 

Free & Open-Source software github.com/sgmap/cout-embauche What’s important to understand is that this tool can be used for free by any entity (administration, association, commercial…). We create a digital common that makes law computable. However, distributing for free doesn’t mean “without any counterpart”. We replace the monetary counterpart by a contribution counterpart: every improvement that is made by a user must be offered back to the community. We then merge those we find useful for all. This is how we intend to scale the development of the more exotic rules. 

A State Startup… 1. Starts with a friction. 2. Learns, makes, repeats. 3. Is built together. 4. Takes advantage of uncertainty. Nothing magical, only a different approach to risk management. The administration is usually focused on reducing uncertainty. We accept uncertainty and use it to, somehow paradoxically, reduce risk. There’s an increased risk of execution, but you will know it much earlier, and thus decrease the risk of delivering something no one really cares about, or delivering something that made sense two years ago but is already obsolete. So, we accept the risk of failure. At small scales, iterations may fail. But startups themselves may “fail”. However, we mitigate failure through openness. And here is the second half of the strategy. 

apientreprise.fr Each of our products is mainly an “appeal product” for APIs we build that make law computable and opens doors between the administration and the outside world. With these products, we get time and budget to develop APIs, have leverage to negotiate openness with administrations, and raise public awareness of the APIs we create. This is how we both empower the ecosystem and spread innovation: even if reluctant at first, the administrations soon realize the untapped potential of exposing their data and services. 

github.com/sgmap github.com/etalab government.github.com/community Even beyond APIs, all of our code is free and open-source. I already mentioned how we ensure contributions happen, through licensing and tooling. But there’s something more to it. By valuing public contributions, we hope to visibly align the politician’s and the people’s goals, as they should always be: each contributor is a voter. 

Startups State (not to scale) So, this is how we make the State able to build digital products that improve public services. I showed how focusing on the vision, iterating, collaborating and embracing uncertainty allow for great products to emerge with little cost. And I showed how we use transparency and openness to make sure we constantly deliver value while still taking risks. 

Startups (not to scale) State By incubating State Startups for different agencies, and perhaps even creating other incubators, we will then spread around the culture we’re demonstrating. 

Matti Schneider @matti_sg_fr Digital Services Incubator Core team member Thanks! Questions? Thanks for your attention.