Background

beta.gouv.fr is the Digital Services Incubator, a mission of the French government’s Secretariat-general for Government Modernisation. Our aim is to spread the culture of digital innovation throughout the administration. We do this through State Startups.

A State Startup is not just an oxymoron, but a match between a team and a mission. There is no capital investment, no separate juridical entity, only the usage of the tools and practices of the digital startups to solve a friction in an interaction an administration has with citizens. We help it identify an “intrapreneur”, one of its civil servants ready to hustle her way into delivering a digital public service, and create a team of experts around her. We hire developers, datascientists, designers, product managers and all needed talents, either under short-term contracts or as independent contractors, and coach that team so it stays laser-focused on solving real-world issues for real-world users. This team of 2 to 4 people will have 6 months to prove a digital product can improve the situation.

Objectives and main results achieved

The new version of the French Open Data Portal (data.gouv.fr) was the first State Startup to come into existence, back in June of 2013[1]. Made publicly available in December that same year, it brought social features to a tool that used to be a catalog of spreadsheets —as in every other country at the time. The new interface created feedback loops from civil society, that could now publish reuses, towards the institutions producing data, thus allowing them to increase the quality of their datasets by leveraging the crowd. Being open-source, this tool is now used in other countries such as Luxembourg, transferring capabilities EU-wide.

Another example is Mes Aides, a web application that helps citizens know which benefits they are entitled to. The base friction is citizens not claiming benefits they are eligible to due to the complexity of law. As a matter of fact, in 2012, up to 80 % of French citizens eligible to financial help for health insurance did not request it. A State Startup was incubated in February 2014, and made a service publicly available at mes-aides.gouv.fr in September. Two years in, over 1400 individuals assess each day their eligibility to 22 social benefits, for the majority in under 7 minutes. A large-scale evaluation of this service, led by independent researchers with controlled cohorts, is currently underway.

We were able to scale this product thanks to strong collaboration between different government agencies and the greater public. Indeed, we started with 7 target benefits and only 70 tests provided by partner agencies. We now compute 22 benefits and have over 600 tests, including edge cases, provided by new partner agencies, foreign territories, but also NGOs using the tool on a daily basis, and even individuals in specific situations. And once again, by opening the code powering this service, strong opportunities were enabled. Two new governmental services, independently from the Incubator, will use the same computation engine to target specific populations. During the Open Government Partnership worldwide summit hackathon, teams of 4 could make tax law computable for countries such as Tunisia and Senegal under 48 hours.

Another service incubated at beta.gouv.fr uses that same engine to provide an embeddable widget that computes the cost of hiring an employee, with all taxes and refunds included, with a single line of code. The first public version of that tool was once again delivered under 6 months and one full-time person. It is now used in over 8 different partner websites, from administrations, startups and SMEs all alike.

Our continued growth, our now 27 products, as well as our increasing number of partner ministries, all prove that building digital services for and with the public, ensuring collaboration across government agencies and beyond, ignoring usual committee-based, siloed decision-making, yields greatly improved services at a fraction of the cost and fosters interest and support from the civil society.

Lessons learnt

We have learnt that the best strategy in digital transformation is delivery. Rather than trying to convince other stakeholders —be they fellow administrations, the private sector or individuals— to work towards a specific goal, it is more efficient to start working on it, while opening at the same time as many ways for contribution as possible. As feedback is truly listened to and taken action upon, a community starts shaping itself and the power of the crowd is leveraged. That support can then be shifted into resources, but also into political power to support larger transformation efforts, convincing more reluctant actors to consider digital transformation as an opportunity rather than a challenge.

Notes

[1] Several academic papers present this history and analyse the public value impact of the incubator. Alexandra Foucaud’s When the French State works in startup mode can give more detailed insight.