An Integrated Methodology for Analysing the Acquisition and Use of Mobility Products and Services

An Integrated Methodology for Analysing the Acquisition and Use of Mobility Products and Services


S Le Vine, M E H Lee-Gosselin, J Polak, A Sivakumar, R Krishnan, Imperial College London,UK


This paper presents a new framework for the analyse the choice of competing ?bundles? of travel products/services. The research is motivated by the need to analyse shared-car services, a small but rapidly-growing part of the mobility marketplace.


Shared-car services offer people on-demand access to a fleet of vehicles distributed throughout a service area. In contrast with private vehicle ownership, there are typically lower fixed costs and higher marginal costs. Participation in such services has grown rapidly in recent years to approximately 350,000 members worldwide, with 200,000 in Europe. Concurrent with this period of growth, researchers have been increasingly interested in both the direct travel behaviour impacts of shared-car service participation and secondary impacts such as on parking demand, transport network operations, and transport-related emissions.

Existing empirical research indicates that participants tend to use shared-car services for less than 10% of their journeys, with public transport and non-motorised travel modes accounting for much of the rest. In general, people choosing to participate in shared-car services do so with a view to satisfying limited and occasional, but foreseeable, needs for car travel. Provided that alternative means of travel can adequately serve one?s residual travel needs, accessing a car via a subscription service may be more attractive than either owning a private automobile or relying on less-flexible methods of car access (e.g. traditional car rental). Shared-car services, therefore, can plausibly serve as substitutes or complements to other travel methods depending on specific circumstances.

Modelling the adoption and use of shared-car services as a ?gap-filler? method of travel therefore presents particular methodological challenges including the need to accommodate complex patterns of substitution and complementarity and the need to account for the selective nature of the catchment of travel served by shared car scheme.

This paper puts forward a new random utility based framework for modelling the adoption and use of shared-car services. In this framework, an individual?s choice of whether or not to participate in a shared-car service is subsumed within a larger strategic choice of which ?mobility bundle? to acquire. A mobility bundle is conceived of as a set of elemental travel modes and services, each of which may make available particular methods of travel. For example, one mobility bundle might consist of simply owning a private car, whereas a second might consist of purchasing a public transit season ticket and a shared-car service membership, a third might consist of shared-car service membership and non-motorised modes, and a fourth might bundle the partial use of one car owned by the individual?s household with the supplementary use of a shared-car service.. The concept of mobility bundles captures the patterns of substitution and complementarity that can exist between the elementary travel modes and services. The choice between alternative mobility bundles is assumed to be made with respect to their ability to satisfy the mobility requirements of a set of key activities, termed the ?reference activity space?. The utility of different mobility bundles is the combination of the utility of each bundle with respect to each activity in the reference activity space. The framework permits flexibility in the composition rule used to combine activity utilities and in our empirical results we explore a number of alternative specifications of the composition rule, inspired from the literature on decision-making under uncertainty and multi-criteria-multi-individual decision-making. The proposed treatment accommodates activities in which a person is observed to participate, as well as those which may be prospective, retrospective, or otherwise unobserved in the estimation dataset. This framework represents a significant extension of the current state-of-the-art which ignores the bundling elemental travel modes and services and treats separately the dimension of product acquisition and use.

The paper is divided into a number of sections. The first section provides a brief overview of existing research into the adoption and use of shared car services. The second section provides a synthesis of findings from in-depth interviews of transport users that provided input to the design of the new modelling framework, and discusses the merits of this framework compared to existing approaches. The third section briefly describes the data collection strategy underlying the operationalisation of the model. This is based on a stated response survey which is linked to the analysis of the UK?s National Travel Survey dataset, (a detailed 7-day travel diary). Data from the NTS is used to estimate participants? reference activity space and in the stated response exercise they are presented with choices between alternative mobility bundles. Novel methods are used to estimate the level-of-service attributes for unchosen travel methods from the NTS sample. The fourth section presents preliminary model estimation results and discusses their interpretation. The final section presents some overall conclusions from the work to date.


Association for European Transport