Season 8 - Implementing the IT

Episode 3 - The best of both worlds

Romain Payet —  To understand how we began creating the Midnight Trains digital product, we need to rewind a little, to just before we hired Claire, the newest member of our team. After the first two missions we assigned to Maxime and Yann, we began working with Capgemini Invent and Frog.

This collaboration is multi-pronged. The first part consisted of working with UX designers who explained how a classic railway digital product works, how it must be organised, what are the absolute must-haves and what is optional. It helped us see what we need on a purely technical level. The second part took place with Frog, a company recently acquired by Capgemini, with whom we tried to define what the digital product for the night train of the future should look like. We want to produce something very different from what’s available today - something contemporary, intuitive, and effective.

Doing so required a process of elimination. By separating them into categories, they presented us with the different elements that we could integrate. Some are essential, others are nice to have - the ones that aren’t mandatory, but would be a bonus. Or features which are absolutely essential, and without them we’d relegate ourselves to clunky, old-school railway interfaces. The differences between these categories are subtle, but important. If we choose these elements right, we will get a head start. So we gradually sorted these options until we got a solid foundation.

Adrien Aumont — At the end of the project, which we made rapid progress with thanks to Capgemini’s experience and our skill, we arrived at a number of conclusions. First, it seems that even if we had a precise vision of what we wanted, complexity remains a sticking point. Unlike the hotel industry, the railway sector faces unforeseen events: train delays, maintenance work, passengers that need to be urgently evacuated, and bad weather. All these must be managed by the IT part of Midnight Trains. In the event of a problem, it has to be able to warn the stations, and the passengers, who may need to be reimbursed, receive help to make their connections, and make accommodations if they miss them. The list goes on. In short, you have to code thousands of possible scenarios, and ensure that the digital product is capable of handling them.

This turns out to be all the more complex because with a classic digital product, we can carry out stress tests to check that everything is working correctly. But it's impossible to run a faux train with faux passengers to check what happens if one of them is faux sick, or if faux snowfall obstructs the tracks.

Romain Payet — Our second conclusion is that the ideal digital product for Midnight Trains is a nice balance between that of the hotel industry, and that of the railway sector. It must provide a level of service similar to that of large hotel apps, while relying on the technical tools enabling trains to run correctly. And for good reason - SNCF Connects or Trainline offer few or no services like this. Luxury hotel apps however are full of them, but aren’t backed by rail traffic management tools. If we go in that direction, our trains won’t go anywhere. So we need the best of both worlds.

Now we know what we want to take from each universe, we have to use the In or Out method; basically separating what we’re going to adopt as existing programs, and what we’re going to code ourselves. To determine this, we’re kicking off major tours and roadshows, through which we hope to discover the different tools available on the market. The selection includes three or four hotel tools, a few others designed for rail and even hybrids, like those used by long-distance luxury trains. We’re going to give each of them our specifications, to evaluate the gap between what we want and what they do, so we can establish a relationship with and commission additional code.

Adrien Aumont — All these companies have already developed products that can be connected with the main interface. They can create additional code if necessary, as long as they believe the code will be of interest to other companies. Therefore we have to persuade them that night trains have a future ahead of them, and that other companies will follow after us and require the same things.

Our role is therefore to centralise and verify what works and what doesn’t work with the tools we discover, and then we can delve into new levels of detail. Maxime and Yann will be identifying the holes, the pieces of code necessary for the software — payment systems, entertainment, and catering — to communicate with each other. As for Claire, she’s giving us a roadshow on revenue management tools. The lucky one will be used to create our very own digital product. Essentially, it’s a work in progress.

Midnight Trains Logo