La startup qui a aidé 120 000 personnes à retrouver un job

Quand 75% du marché de l’emploi est dit caché avec des recrutements qui ne sont pas publics, comment favoriser le retour à l’emploi ?

C’est la problématique qu’a adressé Eric Barthelemy, conseiller à l’emploi chez Pôle Emploi, avec La Bonne Boîte.

Très vite, Pôle Emploi accepte de lui laisser sa chance et il démarre accompagné par l’incubateur Beta.Gouv qui l’assiste d’un coach et d’un développeur.

“On se donne le droit de rater avec les startups d’état, c’est le propre des startups d’état !”

Il a d’abord l’idée d’un réseau social pour mettre en relation les entreprises et les demandeurs d’emploi. Finalement, il va construire un algorithme puissant qui à partir de 40 Millions de données de l’URSAFF récoltées chaque année permet de prédire les entreprises susceptibles de recruter et sur quel poste.

Cet algorithme facilite la recherche d’emploi et l’envoie de candidature spontanée, il permet à plus de 120 000 personnes de trouver.

Eric nous raconte comment :

  • Il a créé son algorithme en allant à la rencontre des demandeurs d’emploi en 6 semaines
  • Il a testé l’outil et amélioré son prototype jusqu’à atteindre 80% de prédiction
  • L’algorithme s’est intégré aux services de Pôle Emploi jusqu’à accueillir plus de 3M de visiteurs par an.

Durant son aventure, il a dû faire sa place malgré la méfiance de certains pour avoir un réel impact sur le retour à l’emploi. Il a su bénéficier de l’accompagnement de BetaGouv et du soutien de son Directeur Générale pour avancer.

Retrouvez l’histoire complète racontée par son principale protagoniste.

The one metric you need to track – Lean Analytics

I recently finished another online course: Lean Analytics by Alistair Croll & Ben Yoskovitz, authors of the book of the same name. Those guys are so smart. They explain everything very simply and clearly – masters of their domains.

This course is very rich and complete with vivid examples, but the one thing that I don’t want to forget is: Focus on ONE metric when doing improvements. Which metric ? Well depends on your business model and which stage of development you are in. The stages being: empathy, stickiness, virality, revenue, scale.

OMTM: the One Metric That Matter.


So ask yourself: what’s the ONE metric I need to track at this stage ?

3 ways to market your service on social media

Beyond the +1, Like and Tweet button, here are 3 ideas to sell or market your online service on social media platform.

1- Broadcast – The most basic way is to uses social media to broadcast your message and market your services. In this approach you establish a presence on a social media platform (a Facebook page, a twitter account or a company page on LinkedIn) and communicate about your service and content. Add a follow-us and a like button and you are set. This is mostly about retention and reaching your users on other platforms.

2- Engage – Another way to use social media is pro-actively engage users and start the conversation. It’s about finding who talks about what you business does and share thinks that you think may be of interest to them. Respond to their expressed reactions. For instance, you are selling cookies; send a message to someone when they are hungry. This approach will allow to increase your user base, but depending on how you to id, it is very time consuming.

3- Force – The most aggressive way I have seen uses of social media is on the services that will only work on top of them. They leverage the virality to the maximum, pushing the users to share events and actively seek participation of other users. Good examples are the so-called social games a-la Farmville and services such as This aggressive approach is very effective to increase dramatically your users base. It also requires quite a bit of thinking getting it right.

Those are just 3 ways I have seen frequently used. Each have their own level of complexity behind them of course.

Twitter search: now relevant

Previously, Twitter search results were only offered according to time, not any form of relevancy.

“We want to make real-time search even more valuable by surfacing the best tweets about a particular topic, by considering recency, but also the interactions on a tweet. This means analyzing the author’s profile, as well as the number times the tweet has been retweeted, favorited, replied, and more. It’s an evolving algorithm that we’ll be iterating on & tuning until practically the end of time.”

The two point I found interesting in that story:

  • how they compute relevance, by considering many factors (not just popularity)
  • how they are going to tweak the algorithm… forever.

Read the full story on the ReadWriteWeb.

You, your privacy and your digital footprint

Image by alancleaver_2000 via Flickr

Yesterday, I was invited to the You+technology event organised by Olswang, a tech/media law firm. The event regrouped an interesting mix of investors, entrepreneurs, visionaries and bankers to exchange on the topics of location and digital footprint. Two guest speakers, Alan Moore (SMLXL) and Tony Fish (author of My Digital Footprint) sparked the conversation with ideas and examples touching privacy and location services in the digital space. Coming with an open brain, I expected to connect my neurons to some good thinking. Here are some thoughts the exchanges triggered.

Alan dived straight into the subject quoting the co-founder of Sun Microsystems, Scott McNealy – “You have zero privacy, get over it.” … Which is true and not just in the digital world, think CCTV, phone book…

Combining location with other services to provide hyper-local services, Alan mentioned Otetsudai Networks, in Japan, which allows local employers in need of immediate help for a task to bring-up a map of the neighbourhood that pin points registered helpers available for the job now. But revealing your location (or not revealing it) can be dangerous. A good example is Please Rob

We can agree that in general, knowing about a person should help a company provide a better service to that person. People are willing to *give* some information away when they get something back for it. Alan’s word of advice was “Trust”. Trust, in the sense that services utilizing personal information, will need to have the Trust of the audience.

Tony talked about his book (My Digital Footprint) – which I am currently reading. He argued that your mobile phone knows a lot about you and your habits, such as which exact TV channels and programs you watch and what type of shampoo you use (by making a lot of assumptions too).

Obviously, your digital footprint is a gold mine for marketers. They aim at understanding people’s profile to serve them more relevant advertisement. No harm when done correctly, both people and the marketers win. But the mix needs to be right.

Tony gave his insights to successfully deal with such a large amount of user footprints. The key is in extracting behaviour DNA of users. Not trying to understand everything the user is doing, most of it useless – but instead focusing on understanding the behaviours. The true value for a business lies in getting actual meaningful and usable data, gathered over time and not easily to replicable.