Virtuo’s approach to data science
Virtuo was launched thanks to two simple observations :
- Owning a car is getting more and more absurd in large cities, from a practical point of view, economically and sustainability-wise. Let’s not forget, individual cars are used less than 5% of time.
- Renting a car is also a painful and slow process at traditional car rental companies (many of whom were founded more than 100 years ago)
⇒ We needed a hassle-free, simple, quick, clear and user-centric alternative.
Virtuo’s “North Stars”
And we did it thanks to three principles:
- Capitalising on tech advancements
- digital key: queues at the counter, pickup and deposit limited by the working hours of the car rental agencies were the two most striking pain points in the rental experience
- collecting and centralizing user data, so that he/she doesn’t to submit again and again the same papers and contact details
- but with a human customer service
2. Offering a state-of-the-art service that is distinct from what the competition can offer. We believe that car rental can really be a daily alternative to car ownership, within the mobility landscape.
3. Keeping our feet on the ground: our core business requires to excel in logistics and ensure we do the best for our customers on a daily basis.
Thanks to those “North Stars”, Virtuo has been providing for more than 5 years the best car rental experience first in French large cities, then in Western Europe. We’ve progressively introduced better services, electric cars, better suited to city driving, delivery, which makes car rental as simple and invisible as possible: “your” car is now available directly at home. We’ve just introduced carbon offset.
That’s how Virtuo addressed the issue of car ownership in large cities, and that’s how we conduct our data science projects.
Identify what we need to solve
First, we identified what we needed to solve, and offered tailor-made solutions instead of creating the need for pre-existing techniques. In data science, it means we try to predict or model only for phenomena on which we know we have levers on. A bad move would be to generate many KPIs and predictions we wouldn’t know what to do with.
Transparency, data democracy, iterations
We offer transparent solutions. For the end customer, a transparent and readable pricing. We provide clear and unequivoqual KPIs to Virtuo teams, it means, and we compute simple, readable and actionable predictions.
We provide autonomy to our users. For the end customer, it means picking up his/her car in an autonomous fashion, letting him/her sending directly his/her papers, two time-consuming steps of the traditional car rental experience where the customer feels infantilized. In data science, autonomy means data democracy : all Virtuo teammates have access to an unified BI tool, and can create reports and alerts according to a “subsiadirity” principle. Everyone uses the same single source of truth, uses the same figures. They’re all trained to build themselves alerts and dashboards, and escalate to the data team only when their needs exceeds their competences.
We try to do our work humbly, through iterative processes, by questioning ourselves in order to improve our tools step by step, so does Virtuo (first the digital key, then delivery, etc.). In data science, all our BI tools and algorithms run a continuous integration and undergo post mortem analyses.
Smarter decisions
Our role as data scientists is to simplify life Virtuo’s teammates and customers, as Virtuo simplifies the life of its customers. To be a catalyst of logistic excellence, to help make informed and data-driven decisions, to make smarter decisions thanks to algorithms that deal with data that would be too complex and intricate for human to handle.
Thus, we have developed a supply-demand optimization algorithm, that recommends car buyings or acquisition spend. In such problem, there are many variables to optimize (supply-demand match, cost of acquisition, etc.) and many, hundreds, thousands micro-decisions to take (do I really need to buy this batch of cars, for how long, where should I allocate it, etc.)
Arbitraging all of these principles
These are principles, North Stars, not hard rules; we try our best to follow them. We sometimes turn away from these, but always at our expense. And there’s a plural at principles, they’re not synonyms, one for another. For that reason, they may compete: a tailor-made solution is not necessarily transparent or simple. Our role is to find the best arbitrage between them.
Interested in knowing more about that or joining the team, do not hesitate to contact us and to follow our next posts on the matter.
Alexandre Journo, Head of Data @Virtuo — The New Generation of Car Rental
and Romain Thoyer, Data Engineer Virtuo - The New Generation of Car Rental
