The Autolib’ barometer is back after a summer break. A sustained decrease in the number of subscribers as well as in the number of rentals supports the hypothesis of a long-term trend.
As usual, all the figures used are presented in an Excel document: scroll to the end of the article to download it.
Since our last barometer (July 2017), the number of active yearly subscriptions has continued to decrease. The decrease was particularly sharp in September (-2.2%), a pattern that may be caused by a failure to renew their subscriptions from those users who subscribed in September last year. Since November 2016, the number of active yearly subscriptions (estimated at the end of each month) has been diminishing each month without exception. Between November 2016 and November 2017, i.e. a 10-month period, a total decrease of 5.4% in yearly subscriptions has been observed.
The number of rentals undergoes strong seasonal variations: for instance, it sharply decreases in July and August (i.e. French summer school holidays). In order to assess the long-term evolution of rentals, it is therefore necessary to compare yearly figures on a month-to-month basis. The two graphs below respectively show:
– the average number of Autolib’ rentals per day, by month, since June 2014 (blue curve);
– the evolution of the number of rentals relative to the figure observed on the same month in the previous year, for each month since June 2015 (yellow dots).
Since October 2016, each month has recorded fewer rentals than it did on the previous year, with decreases spanning from 4 % to 16 %. The decrease in the number of rentals looks like a long-lasting trend, as does the decrease in the number of yearly subscriptions.
In our article about Autolib’s business model, released last January, we noticed an inverse relationship between the number of yearly subscribers and the frequency of Autolib’ rentals. This inverse relationship led us to hypothesize that, as the number of subscribers grows, it may become harder and harder to find available cars, which may in turn partially deter subscribers from using Autolib’.
The graph below confronts the evolution of the number of yearly subscriber per car – which is an indicator of the cars’ availability – to the evolution of the frequency at which subscribers use Autolib. The decrease in the number of yearly subscribers hasn’t resulted in an increased frequency of use. In other words, the inverse relationship that we identified when the number of yearly subscribers was rising no longer exists when the number of subscribers is decreasing. One possible interpretation would be that users who turned away from Autolib’ due to the cars’ lack of availability are lastingly deterred and do not return to their former level of use when the cars’ availability rises again.