Romain Payet — Now came the time to estimate our potential revenue, in other words the amount of money we would make on each of the lines that we envisaged operating. First we made a list of all the lines compatible with a night-train journey, based on two key criteria: they were less than 1,500 kilometres long (that is, medium-haul), and all used electrified tracks and not LGV lines that couldn’t be used in the middle of the night.
Once we’d settled on our potential lines, we got hold of the air-traffic figures for these routes in normal years – 2018 and 2019, so we weren’t looking at data skewed by Covid-19 – and we came up with a ranking. For each of them, we also looked at how seasonality might affect things. Since the data is in the public domain, it was all easy to find out and, fortunately, much of it was very detailed too. We could see how many passengers were passing through each airport, broken down month by month, with a distinction made between departures and arrivals.
As well as enabling us to rank the various lines by footfall, this data allowed us to start to build out our idea for a ‘hub’, a central point from which several lines would run. We were convinced from the start that this would allow us to make serious savings in operational costs and would allow us to build awareness for our brand and sales network. These would all prove key factors in the future success of Midnight Trains.
Although all this new information was crucial for us, it didn’t give us an exact number of passengers who would be willing to take Midnight on each of these lines. So we now had to estimate a plausible market share for the night train, and Midnight in particular, that could be gained from these air-passenger numbers.
To do that, without spending a fortune on forecasts from a consulting team, we lent on: data from trains that were already run by an operator. Because although it may have been in decline for some years, the sleeper train remained an existing market, with clear results; historic data from lines that had been operated in the past; and forecasts from external public bodies.
One of our first sources of public data from an operator were those from the Spanish Trenhotel and, above all, its international lines. Launched in 1991 by Renfe, these trains had the aim, as their name suggests, of being real ‘hotels on rails’, which you won’t be surprised to hear very much interested us. All the more so because many of them cross borders, much like ours intend to: the Francisco Goya between Paris and Madrid, the Joan Miró Paris and Barcelona, the Salvador Dalí Barcelona and Milan, and the Pau Casals Zurich and Barcelona. We also looked at data from Thello, whose night trains between Paris, Milan and Venice still run, even though they had been paused during the Covid-19 pandemic, as well as all the international lines run by the Austrian OBB Nightjet firm.
From each of these sources, we managed to get hold of a reliable source of passenger numbers (without elements of seasonality) for a given year, which we then compared with air traffic over the same period, so we could estimate a market share for the sleeper train. Our various analyses led us to the conclusion that the market share for sleeper trains would be between 5 percent and 12 percent of air traffic on each of the lines we had analysed.
Since we now had visibility over the market, we started making passenger forecasts. These were largely lifted from the French government’s TET report, in particular its section on international sleeper trains, as well as research from the French group Oui au Train de Nuit.
Finally, to obtain our passenger-number predictions for each of the selected lines, we applied the market shares (minimum and maximum) obtained from air traffic in 2019, based on the hypothesis that pre-Covid levels would have returned around 2024.
These passenger forecasts, though preliminary and rather rough around the edges, would prove very useful in two ways.First, it would allow us to refine our requirements for rolling stock. In effect, the advantage of having locomotives pulling several carriages (rather than self-propelling carriages, combined in an indivisible unit of 200 or 400 metres) lies in the flexibility that it allows over the layout of the train. It is thus possible (and wise) to create a carriage that matches the potential size of the market and predicted occupancy levels.
For each line, we estimated the number of carriages we would need, and on top of that, all the other requirements involved in our business development plan. This allowed us, as we touched on last week, to figure out the costs relating to rolling stock, as well as, to a lesser extent, infrastructure costs. Second, those forecasts would allow us to estimate potential revenue for each line and a rough overall revenue figure by applying an average price over the entirety of the routes.
Thanks to these sums, we were in a position to work out whether our business could be economically viable or not. And since we didn’t just abandon the project right there and then, you can guess our conclusions. However, putting the figures to one side, we couldn’t make Midnight Trains a reality alone. We needed experts, consultants to guide us along the way. But the search for the right people would prove rather surprising and difficult.