Creating a More Inclusive and Accessible Digital Transport System: Developing the INDIMO Inclusive Service Evaluation Tool

Current transport planning and policy are mostly focused on the transport system itself, without actually focusing on those using the system, resulting in the idea that a decent or good working system is enough to provide transport for everyone, indirectly indicating that all people have the possibility to participate in activities (Brown et al. 2009). It cannot be denied that this approach has provided an ever increasing accessibility to a significant part of the population, however, now it is obvious that this does not mean this method has proved to be sufficient for everyone (Lucas 2012). The ever-increasing digitalization of the transport system, and consequentially the increased complexity of the same demands a different approach.

The need for a new approach emerged in the 1990s, with the introduction of the capabilities approach, a theory developed by Amartya Sen and Martha Nussbaum (Nussbaum 2000, 2011; Nussbaum et al. 1993; Sen 1985, 2001, 2009). Various definitions and descriptions have been adopted, for this paper, we adopt the following definition as proposed by Sen (1995, p.1)” A person’s capability to achieve functionings that he or she has reason to value provides a general approach to the evaluation of social arrangements, and this yields a particular way of viewing the assessment of equality and inequality”. The capabilities approach was also promoted as a methodology for the appraisal of transport systems, with its foundation in the contribution it provides to a persons’ opportunities and wellbeing, which is a basis for consistent evaluative approaches to influence transport planning and policies (Alkire 2003; Vecchio and Martens 2021).

Besides its ability for evaluating transport systems, the capability approach inherently promotes accessible transport systems and thus has a positive impact on the groups in society that are more poorly served by the current transport system than others (Lucas 2012; Martens 2017). Furthermore, it is especially useful when it is used for evaluating a diverse set of people, each with their own capabilities and constraints, keeping in mind the distribution of mobility resources and how these are differently available and used by different people (Vecchio and Martens 2021).

The capabilities approach, especially in relation to mobility has been approached in multiple ways, from a very broad interpretation: ‘the ability to be mobile’ (Beyazit 2011), from a physical, social and financial point of view, to ‘being able to use public transport’ (Ryan et al. 2015). Another possible approach was introduced by interpreting accessibility as a capability, rather than just being mobile. This interpretation focused on the participation in society, for which a person needs the be mobile to a degree (Martens 2017).

Lastly, another important difference in interpretation and use of the capabilities approach is combining it with a top-down or bottom-up approach. For the creation of a service or product for people vulnerable to exclusion, the bottom-up approach is preferred as this examines how each person attributes different values to an activity and how this results in participation in activities due to the accessibility provided through the transport system (Vecchio and Martens 2021). In other words, the bottom-up approach includes those people in the development process for whom the product or service is meant. Contrary to the top-down approach, where users are not involved in the development process. A main disadvantage of the top-down approach in this case is the lack of knowledge about barriers and needs that are experienced by people vulnerable to exclusion, as well as other stakeholders of the digital transport network. Therefore, in this paper, we consider the bottom-up approach to be the most suited approach when researching and working with citizens vulnerable to exclusion. In combination with the capabilities approach of Randal et al. (2020), this results in a policy and service evaluation tool that was developed with input from users, developers, operators etc. so that their capabilities, requirements and needs are integrated in the tool.

Universal Design Principles in Digital Service Design

Universal design (UD), a concept first mentioned and used by Ronald Mace, an architect who worked on social inclusion of people with disabilities, is described as “the design of products and environments to be usable by all people, to the greatest extent possible, without the need for adaptation or specialized design” (Aarhaug 2019, p. 2). This resulted in a first, comprehensive approach to develop a more inclusive world. Earlier attempts resulted in a segregated approach, one infrastructure for those without ‘disabilities’ and one for those who needed an adapted approach. These attempts usually were considered to be ugly and required additional investment (Mace et al. 1998). Throughout the years, the way UD was approached has evolved and a vision emerged that UD promotes a new approach to design that celebrates diversity and provides equal opportunity of access to mobility and services (Audirac 2008). In the EU, as part of the General accessibility act (2019), the concept of UD is defined and integrated within the legislation for the development of products and services (European Commission 2019). A more recent practical evolution is the implementation of UD in digital services, with Begnum and Bue (2018) concluding that there is still a significant lack of awareness of UD among designers and other stakeholders.

The widespread emergence of services, especially digital services, has resulted in the fast growth of service design, focusing on the holistic experience of users, with the goal to make services user friendly, easy to use and more intuitive (Polaine et al. 2013; Scott et al. 2016). According to Begnum and Bue (2021) a widely accepted definition of an inclusive digital service is still not available, so, in this paper, we adopt the working definition as described in their work: “A service is universally designed when its customer journey is usable to all people (to the greatest extent possible and without the need for adaptation or specialized design), by selecting suitable touchpoints” (Begnum and Bue 2021, p. 22). The design has a significant impact on the value creation of the digital services (Law et al. 2008) and consequently on society as well (Kuk and Janssen 2013), with digitalized services dominating society at an ever increasing pace (Newman 2020). The impact service design has, is significant, and even though there is an idea on the relevant concepts linked to universal service design, awareness about this method is overall lacking (Begnum and Bue 2021; Delaere et al. 2020).

The introduction of universal design, or inclusive design, which are often used as synonyms (Goodman-Deane et al. 2010; Clarkson and Coleman 2015) in digital services has resulted in the development and design of more inclusive digital services. Adopting universal design principles when developing services is not necessary for most users, but it does provide the opportunity for vulnerable to exclusion people to make use of the services as well, resulting in a more accessible and inclusive service for all users. In this regard, one set of regulations that has proven impactful are the Web Content Accessibility Guidelines, which are guidelines defining how to make Web content more accessible to people with disabilities. Accessibility involves a wide range of disabilities, including visual, auditory, physical, speech, cognitive, language, learning, and neurological disabilities (W3C 2008). Within the European digital landscape, the Web Content Accessibility Guidelines 2.0 (WCAG 2.0) are mandatory for National agencies, and their contractors (European Union Agency for Fundamental Rights 2014). Additional guidelines are available in the WCAG 3.0, which have significant overlap with WCAG 2.0, but it introduces an alternative to previous versions (W3C 2021). Although WCAG has resulted in an increase of the accessible character of web based applications and platforms, since social and spatial inclusion is not really integrated into the definition of the WCAG, it does not provide an answer or guideline to all barriers of digital services (Begnum et al. 2018). Tools have been developed to assess the inclusive character of projects, programs, organisations and companies, but only a few have been created for the evaluation of digital services (Department for Digital, Culture, Media and Sport 2017), with none focusing on digital mobility services.

The implementation of UD principles in legislation shows an effort for a more generally accessible design and use of services and products. Nevertheless, there is still a lack of knowledge among developers, designers regarding UD in service design. Moreover, inclusivity of services and products is often only a sidenote in the company goals or not mentioned at all (Delaere et al. 2020). This is where the SPET can fill an existing gap, it can provide assistance to evaluate the digital inclusivity of digital mobility services and can provide guidelines to design, develop and implement accessible, inclusive digital mobility services.

Pappers et al. (2020) stated that the use of co-creation, as a form of public participation has become more present in multiple industries, especially within health and education. In transport development and policy, co-creation has only recently become more of a standard, advocated by the European commission with projects like those within HORIZON2020.

Developing digital mobility services that are accessible and inclusive require user- centred development, and applies a bottom-up approach, meaning that the users are involved in the development process and have a significant impact on the final output. One way of involving users in the design of a service is using ‘co-creation’, that can be interpreted as a more intense, further reaching form of customization, involving “collaboration with customers for the purpose of innovation” (Kristensson et al. 2008, p. 47). In a co-creation process the focus shifts from the firm or company to the users, who are significantly involved in the development process addressing user- specific needs (Chathoth et al. 2013).

The first stakeholder group, mainly the users, are mostly the sole focus of inclusivity measures, as these are the people that might potentially be excluded. But, for the development of a digital service, it is not sufficient only to keep in mind the users, as this can result in demands that cannot be fulfilled by developers, operators or policy makers. Therefore, it is important that both sides have sufficient input in the development of the tool. This will eventually result in a service that is inclusive and accessible for all users, but will also make sure that all inclusion related changes and measures are feasible for the developers, operators and policy makers to implement as well.

Conceptual Framework

The Service and Policy Evaluation Tool, that we developed, addresses the gap between the provided capabilities of potential users, and the required capabilities of a digital mobility or delivery service based on the capabilities model of Randal et al.( 2020). It is intended to align the set of requirements posed by a digital service to match the capabilities of potential users to facilitate their participation in society. The tool aims to help the key stakeholders that have an influence on the development of digital applications and services, i.e. developers, operators and policy makers, to make digital mobility services universally accessible and inclusive.

The goal of the SPET is to intervene in matching both sides of the capabilities model (Fig. 1). Provided capabilities (e.g. digital skills), have to match the required capabilities (constraints) in order for a person to participate in a specific activity (e.g. visiting a family member), which are influenced by the actual activity (e.g. using a ride-sharing service) and the environment, i.e. digital and physical context in which the activity will take place (e.g. the smartphone application that is needed to book a ridesharing service as well as being able to find a physical meeting point to get to the vehicle)(Vecchio and Martens 2021). Applying universal design to the service and the environment could help to design a service that accommodates the capabilities of as many people as possible rather than designing specialised services addressing capability limitations. In order to facilitate universal design, developers need to take it into account when designing the service and its interface; operators need to consider it when they operate a digital mobility service; and policy makers need to create guidelines and regulations that incentivise developers and operators to comply with accessibility and inclusivity requirements. If the Universal Design principles are considered, the activity and the environment can be designed in a way that accommodates the requirements of the users in a broad sense and the capability gap disappears or it is at least decreased. In this way, a user would be able to book a ridesharing service through a smartphone app, communicate with the driver and find the meeting point and board the vehicle in order to reach her/his destination irrespective of her/his level of digital skills.

The SPET has both an assessment and a steering role in this process. On the one hand, it would allow policy makers, as well as developers and operators to assess to what extent a service complies with minimum and recommended accessibility and inclusiveness standards; but on the other hand, it would also steer service design by giving recommendations on applying universal design to improve specific features of services and applications.

Before evaluating the digital inclusivity and accessibility of a mobility service, the evaluator will have to select what kind of service will be evaluated and which type of stakeholder (developer, designer, operator or policy maker) she or he is. Based on these choices, a specific set of topics and questions is presented to the evaluator.

The services currently available for evaluation are those linked to the INDIMO pilot projects (food delivery, ride- and car-sharing, electronic parcel lockers and multimodal route planners), with ‘other’ as an additional option, to provide the evaluator with the option to evaluate another type of service.

Once this information has been entered in the tool, the actual evaluation starts, in which the evaluator will have to provide answers to the questions linked to each of the evaluation-topics (Table 1). The structure of the questionnaire is threefold, linked to the three themes present within the tool: Universal Design Principles, service features and assistance provided.

As explained in previous sections, input for the tool is based on multiple interviews with each of the stakeholder groups, after which the information was used to develop the UDM, UIL and SCG.

To create a clear structure three themes, based on the collected information, were created, the first theme is the universal design principles, which is based on an adaptation of the standard universal design principles. From the original seven principles, five were kept, the other principles (simple and intuitive, size and space for approach and use) were left out. In the digital world, ‘simple and intuitive’ was considered to be very different for each stakeholder group and service, also, for a large part, this aspect is covered by the other principles. The ‘size and space for approach and use’ was not necessary as the physical aspect is not evaluated by the SPET, rather this principle was replaced by digital (and spatial) wayfinding. The last topic that was included in this theme, is the data protection and privacy of the users. Furthermore, one principle was adapted slightly, a cognitive part was added to the low physical effort principle, to better fit with the digital approach.

The second theme ‘service features’ is an umbrella term for all the topics related to the inclusivity of the services considering pricing, payment methods, information provision, communication and spatial accessibility.

Finally, the third theme ‘assistance offered’ refers to the topics that provide help for those who have issues using the application or the service, as well as focusing on the iconology used in the application. This topic was identified as a separate topic as one of the key findings of the interviews with users and non-users was that people would need human assistance when using digital mobility services.

This resulted in 18 topics, divided among three main themes as presented in Table 1.

Each of these themes, with their appropriate topics form the core structure of the tool. For the evaluation of individual topics, a number of questions were developed for each theme.

The entire questionnaire is made up of three types of questions, categorized by the way in which they are answered. The first type of questions requires a ‘yes or no’ answer, the second type are answered in a Likert scale with 3 or 5 potential answers, the last type of questions are the ones that can be answered by selecting suitable option from multiple choices. Once the evaluator has answered all questions the topic and theme-wise scores are calculated.

However, the final score does not only depend on the answers from the evaluator, but it is also influenced by the weights allocated to each of the topics.

Each of the topics was described and presented to, and discussed by, 29 experts in the field of digital mobility, policy and inclusivity in order to determine the relative importance of the topics. Experts concluded that all three themes are equally important for an accessible and inclusive application or service, but topics under a certain theme may differ in their relative importance. For allocating weights to each of the topics, a survey was distributed among the experts (operators, developers, researchers and mobility professionals) participating in the INDIMO project. To quantify the relative importance of the topics, the experts were asked the question: ‘Relative to the importance of other topics in this theme, how would you score this topic out of 20’. It was decided to use a scale of 1–20 with the consideration and objective of providing experts a scale wide enough to comfortably express the differences in the perceived importance of each topic. Based on the scores provided by experts, the topics within a theme were compared and weights, relative to each other, were allocated to the topics Table 2. The standard deviation of the weights allocated to each topic is also included in Table 2, showing the variance in answers provided by the experts.

Calculating the weights for each topic happens as follows: for the simplicity of calculation and convenience of understanding and presentation, total weight for all themes is considered to be 1. After this the total weight was equally distributed among three themes which are equally important from the perspective of accessibility and inclusivity of an application and service. This way, the weight allocated to each theme is 0.33 (approximately). Then 0.33 was divided among topics according to the ratio of the average score (out of 20), allocated by the experts to each topic to find the topic weights. The actual weight is then calculated by dividing the topic weight score by the sum of all topic weight scores within a theme. The result of this calculation is presented in Table 2 in the column on the right ‘Topic weight’.

The final score for each theme is calculated based on the answers for each of the questions that topic contains. Table 2 shows the number of questions for each topic, the number of questions does not have an effect on the importance of a theme. Due to the fact that yes/no questions, multiple choice and different Likert scales are used, a transformation is necessary (Table 3). Each of the scores is re-distributed on a 20 point scale. After the re-distribution, the average of all the unweighted score for each question results in the unweighted topic score.

Once the evaluation of the service is finished, i.e. the evaluator answered all the questions, the average of all weighted topics scores results in the theme performance score. Finally, the average of the three themes results in the overall performance score for the evaluated digital transport service. An example is given in Table 4.

After this step all calculations are finished and the results are presented. For each theme a spider diagram is shown (Fig. 3), containing each of the topics included and their performance score in percentage giving the evaluator an easy-to-interpret result. Besides the spider diagram, the evaluator receives the performance scores for each of the three themes, accompanied by a general recommendation, as well as an overall performance score. Three general recommendations are possible, depending on the score of the theme (low, medium or high).

If the evaluator wants a more detailed representation of the results, this is possible as well. For each of the evaluated topics the evaluator receives a score and recommendations that would help to improve the service so that the performance score for a specific topic increases, this of course with the ultimate goal to improve the digital inclusivity and accessibility of the mobility service.

The recommendations are the final result produced by the SPET, providing the evaluator with relevant information on how to make the service more digitally inclusive and accessible. Each of the recommendations (Fig. 4) are specifically linked to the questions in the tool, providing detailed and focused interventions to the evaluator. Depending on the score for each topic different recommendations are presented in three categories: low, med, high, depending on their importance for a more inclusive service.

Recommendations included in the tool are all linked to the aforementioned INDIMO toolbox providing a more elaborate explanation about every recommendation. The recommendations are all organised in the INDIMO toolbox and explained in these documents: (Capaccioli et al. 2020; Ciommo et al. 2020a, b; Hueting et al. 2020).