Keywords

1 Introduction

As one of flexible and convenient means of public transportation in cities, taxicab has been one of the important vehicles for people’s daily personal mobility [1]. The quality of taxi service and the satisfaction degree of user experience of passengers reflect from a certain side the city’s integral civilization image and overall service quality [2].

Therefore, having impact on a city’s civilization image and the quality of service supply, it is an important subject that the key influencing factors critical to deter-mining the satisfaction degree of user experience and its interrelationship are explored and discovered so as to take tailored measures to overcome the potential problems in taxi industry, to elevate taxi service quality, and as a result, to finally raise the levels of user experience and the satisfaction degree of taxi passengers as consumers.

2 Literature Review

In the field of taxi service and taxi industry, numerous studies have been conducted by researchers. Many theories and methods in management science and economics are employed to explore pricing, management mode, business model, competition and service in taxi industry. Curry et al. [3] studied the advantages and disadvantages of taxis and modes of public transport by establishing a queuing model for passengers to take a taxi at an airport. Cairns and Liston-Heyes [4] studied the competition and regulation issues in the taxi industry, considering that price and access controls are necessary. Xiong [5] put forth the proposal of getting rid of the single mode of taxi market operation by breaking up the administrative monopoly. Yang et al. [6] studied the non-linear pricing of taxi service. Gu and Zheng [7] explored the rationality of taxi pricing system. Anderson [8] and Rayle et al. [9] compared two for-profit models with traditional taxis and the new share rides, and its impact. Zhang et al. [10] probed into the relationship between taxi service strategy and income. Hai et al. [11] and Shi and Lian [12] studied the demand-supply equilibrium between passengers and taxis. Leng et al. [13] studied the impact of the appearance of taxi app and the apps’ commercial competition on the taxi market pattern and service mode.

Meanwhile, some principles in computer science and mathematics are used in solving the taxi-related technical problems. Osswald et al. [14] carried out optimal design of HMI for electric taxis. Moreira-Matias et al. [15, 16] studied the time and spatial distribution regularity of taxi passenger. Yan et al. [17] proposed a method for the most efficient taxi scheduling. Hu [2] ranked passengers’ needs according to their importance in optimizing taxi service quality by quantifying these needs with QFD method.

In addition, some studies focus on the behavioral and psychological aspects of taxi driver and/or passenger groups. Zhang and Wang [18], Brandenburg et al. [19] and Tang et al. [20] studied why taxi drivers refuse to take passengers, what factors arouse a taxi driver’s anger when driving, and how taxi drivers perceive the routes. Shen [21] evaluated single taxi and its affiliated company with multiple indicators by introducing Customer Satisfaction Index and developed recommendations about service improvement. Li [22] studied the effective mechanism for dealing with customer complaints in taxi service to integrate passengers’ feedback into the taxicab management by Customer Relationship Management. Zhou [23] targeted to solve problems in user experience by using Customer Experience Management methodology in taxi service research.

In the field of research on passenger’s user experience involving in the mental feelings and physical sensations [24] in the whole process of travel, researchers’ interests mainly focus on exploring the space design, emotional experience, service design, experience design and efficiency improvement in public places such as trains, buses, airplanes or airports and platforms, and on planning the public transport system from the perspective of management to optimize passenger’s user experience. Hensher et al. [25] investigated the role of service quality in enhancing user experience of bus passengers in deregulated markets with Generalized Ordered Choice preference model. Stradling et al. [26] put forth an approach to improve people’s experience in mobility by city bus after surveying and analyzing the factors that result in people’s dislike for taking a bus. Acevesgonzález et al. [27] compared the behavioral differences between young passengers and elderly passengers in keeping a stable stance while traveling by bus, and suggested improving the experience of older passengers by strengthening the regulation of bus drivers. Ahmadpour et al. [28] studied interior design of airplane cabins from the perspective of passengers’ physical and psychological comfort, and developed design recommendations and an evaluation model. Bogicevic et al. [29] and Ahmadpour et al. [30] explored the factors affecting passenger’s experience and comfort in airplane trips. Harrison et al. [31] discussed the framework for adaptive terminal design from the perspective of passenger-orientation based on a concept model of user experience in the airline industry. Chou and Kim [32] employed Structural Equation Model method in analyzing the influential factors to the degree of satisfaction and royalty of passengers of high-speed trains, and in making recommendations from the aspects of marketing strategy and corporate image. Foth and Schroeter [33] probed into how real-time passenger information systems can make a difference in improving the passenger experience. Li and Li [24] put forward user experience-oriented solution by planning optimization so as to enhance the passenger’s experience of bus service. Fu et al. [34] discussed interior space design of trains in terms of space experience, spatial scale, the behavior of passengers. Bai [35] analyzed the passengers’ perception of interior design of underground carriages, and developed recommendations about function design optimization to improve passengers’ visual and emotional experiences [36]. Zheng [37] and Zhou [38] explored, with methods of service design and passenger behavior research, how to improve terminal transfer efficiency for passengers to enhance passengers’ degree of satisfaction.

Generally speaking, there are less studies on user experience and its degree of satisfaction in taxi passenger service. The purposes of the existing researches mainly lies in (1) instead of more targeted and detailed studies on user experience of taxi passenger, taxi passenger’s user experience is usually used to analyze other problems, for example, to improve taxi management mode and operation mode; (2) There are few researches focusing on influential factors critical to determine the degree of satisfaction of taxi passenger while currently focusing on the overall feeling of taxi passenger in a general way, for example, feeling of taxi’s inner environment experience or service process in temporal dimension. In this study, by employing Decision-Making Trial and Evaluation Laboratory (DEMATEL) method, the interrelationship between main influencing factors to user experience in taxi passenger service is explored to extract the key factors critical to determining the satisfaction degree of user experience and to discover the structural elements in the satisfaction of user experience in taxi passenger service.

3 Method

3.1 DEMATEL Procedure

Decision-Making Trial and Evaluation Laboratory method was originally developed by the Science and Human Affairs Program of the Battelle Memorial Institute of Geneva for exploring and solving complicated and intertwined problems by understanding the structure and the cluster of intertwined problem, and identification of workable solutions by a hierarchical structure [39,40,41,42,43,44].

As an effective methodology for analyzing systematic elements of complicated and difficult problems using graph theory and matrix tools [45], DEMATEL method has been successfully applied in many fields involving in social security reliability [46], market strategy [47,48,49], research and development in manufacturing and management [50,51,52], suppliers and supply chain [53,54,55,56], logistics industry [57], business partner selection [58], performance evaluation [59], investment strategy, construction projects [60, 61], project risks [62], whole vehicle and auto component design and manufacture [50, 63,64,65], disaster response education [66, 67], 68 gender equality [69], data transmission technology [70, 71], e-tailer credit evaluation [72], sustainable development and environment protection [73,74,75,76,77], and agricultural field [78].

DEMATEL method is used in this study as a tool to extract the critical influencing factors by defining main influencing factors and exploring the interrelationship between them, aiming at unveiling the specific structure of taxi passenger’s satisfaction of user experience for taxi service quality improvement.

The DEMATEL procedure in this study comprises the steps of (1) defining the main influencing factors that impact upon the satisfaction degree of taxi passenger’s user experience in the taxi service; (2) inviting subjects to estimate the direction of influence and the degree of priority when each of the main factors is compared with each of other main factors in DEMATEL questionnaires with one of four level values; (3) generating the direct relation matrix Z; (4) calculating with developed DEMATEL tool the λ value, the normalized direct relation matrix and the direct/indirect relation matrix T, and the corresponding Prominence value, i.e., (D+R), and Relation value, i.e., (D−R), of each main factor; (5) drawing a causal diagram in which the directions and degrees of the impact factors can be directly observed to help build the problem structure; and (6) extracting the key influencing factors from the main factors to identify the relationship between the main factors.

3.2 Twelve Main Influencing Factors

In order to mine the information on both user experience of passengers in the taxicab and what factors have impact upon their perception of the satisfaction with taxi service, a large number of consumers who often take taxi and a number of taxi drivers are investigated by combined user research methods such as brain-storming, observation, self-report and in-depth interview. Meanwhile, the problems taxi passengers face in taxi rides are collected as supplementary information to be analyzed by reviewing related literatures and online forums.

After a large quantity of descriptive statements and sentences gathered by channels motioned above are processed and analyzed, the following twenty-one factors are summarized and defined as the preliminary influencing factors concerning taxi passenger’s user experience. According to temporal dimension, they are organized into the following factor categories: (1) four factors occurred before getting in the taxi-cab, including ‘time to wait for pick-up’, ‘rising in price during peak times’, ‘unmatched vehicle information’, and ‘canceling order’; (2) twelve factors during taxi ride, including ‘driver’s talking on the cell phone when driving’, ‘the driver’s familiarity with the local roads’, ‘driver’s chatting to passenger(s) when driving’, ‘long detours and intentional slowly driving’, ‘information on the destination’, ‘weather conditions’, ‘designated location for pick-up’, ‘picking others up during taxi ride’, ‘pettish driving mood’, ‘buckling up seat belts when driving’, ‘driver’s smoking when driving’, and ‘time-phased pricing’; and (3) five factors occurred after taxi ride, including ‘driver’s using counterfeit money’, ‘driver’s using fake invoices’, ‘driver’s returning articles left on taxicab by passengers’, ‘an additional charge’, and ‘evaluation on driver’s service’.

At last twelve main influencing factors are sorted out and given serial number Factor 1 to Factor 12 as the following: (1) Factor 1 - ‘driver’s talking on the cell phone when driving’; (2) Factor 2 - ‘the driver’s familiarity with the local roads’; (3) Factor 3 - ‘driver’s chatting to passenger(s) when driving’; (4) Factor 4 - ‘long detours and intentional slowly driving’; (5) Factor 5 - ‘information on the destination’; (6) Factor 6 - ‘weather conditions’; (7) Factor 7 - ‘designated location for pick-up’; (8) Factor 8 - ‘picking others up during taxi ride’; (9) Factor 9 - ‘pettish driving mood’; (10) Factor 10 - ‘buckling up seat belts when driving’; (11) Factor 11 - ‘driver’s smoking when driving’; and (12) Factor 12 - ‘time-phased pricing’.

3.3 DEMATEL Questionnaire

The DEMATEL questionnaire (Table 1) is used for users as consumers and passengers to estimate the direction of interaction and the degree of relative priority of each factor listed in first column to each factor listed in first row in Table 1. The priority degrees of factors in first column over factors in first row are defined in four levels, i.e., value ‘0’ means ‘do not affect’, value ‘1’ means ‘slightly affect’, value ‘2’ means ‘fairly affect’, and value ‘3’ means ‘strongly affect’. Totally, thirty-two pieces of effective data are collected from DEMATEL questionnaires.

Table 1. The DEMATEL questionnaire
Full size table

3.4 DEMATEL Operations

Direct Relation Matrix.

By processing and averaging the thirty-two pieces of effective data, the elements in direct relation matrix Z of the interactions between twelve main factors are obtained as listed in Table 2.

Table 2. The elements in direct relation matrix Z
Full size table

The degrees of interactions between main factors are then calculated with a DEMATEL operation tool developed, as shown in the figure in Appendix. And the λ value of 0.0653168, the elements in the normalized direct relation matrix, the elements in the direct/indirect relation matrix T, and the values of Prominence and Relation of every factor among twelve main factors are derived.

Direct/Indirect Relation Matrix.

The elements in direct/indirect relation matrix T are as listed in Table 3. By listing all elements in matrix T in a sequence of number, the quartile deviation, i.e., 0.084, of this sequence is calculated as the threshold to measure the strength of interactions between factors. If all values in the row and the column that correspond to an element in the matrix T are below this threshold value at the same time, the corresponding row and column will be removed.

Table 3. The elements in direct/indirect relation matrix T
Full size table

Prominence and Relation.

The Prominence value, i.e., (D+R), and Relation value, i.e., (D−R), of each of twelve main factors are calculated and listed in Table 4. The Prominence indicates the relative weighting (percentage) of one factor’s influencing strength in total influencing strength of all factors, while the Relation indicates to what extent one factor has influence on or is influenced by other factors and the absolute value of the Relation value (D−R) represents the influencing strength. The positive Relation value of one factor means it has an impact upon other factors while a negative Relation value means it is influenced by other factors. It can be observed in Table 4 that Factor 2, i.e., ‘the driver’s familiarity with the local roads’, has the maximum value of Prominence, and Factor 10, i.e., ‘buckling up seat belts when driving’, has the minimum value of Prominence. This indicates that the degree of driver’s familiarity with the local roads is the most important influencing factor related to passenger’s user experience in taxicab while the drivers’ buckling up seat belts when driving is the least significant influencing factor.

Table 4. Prominence and Relation values
Full size table

On the one hand, all factors’ values of Prominence in Table 4 can be averaged as mean value, 2.400. Eight factors with Prominence value above 2.400 can be lined in order from larger Prominence value to smaller Prominence value as follows: Factor 2 (‘the driver’s familiarity with the local roads’), Factor 4 (‘long detours and intentional slowly driving’), Factor 3 (‘driver’s chatting to passenger(s) when driving’), Factor 1 (‘driver’s talking on the cell phone when driving’), Factor 8 (‘picking others up during taxi ride’), Factor 9 (‘pettish driving mood’), Factor 5 (‘information on the destination’), and Factor 7 (‘designated location for pick-up’).

On the other hand, as for Relation value, it can be observed in Table 4 that five factors have positive Relation values, indicating that they have impact upon other factors. More specifically, Factor 6 (‘weather conditions’) has the strongest impact upon other factors, and both Factor 2 (‘the driver’s familiarity with the local roads’) and Factor 1 (‘driver’s talking on the cell phone when driving’) have great impacts upon other factors, respectively, while Factor 11 (‘driver’s smoking when driving’) and Factor 5 (‘information on the destination’) have minimal positive influence on other factors.

On the contrary, seven factors left have negative Relation values indicating that they are influenced by the five factors with positive Relation values mentioned above. To be specific, Factor 7 (‘designated location for pick-up’) is influenced strongly by other factors while Factor 9 (‘pettish driving mood’), Factor 4 (‘long detours and intentional slowly driving’), Factor 3 (‘driver’s chatting to passenger(s) when driving’) and Factor 8 (‘picking others up during taxi ride’) are also influenced clearly by other factors.

3.5 DEMATEL Causal Diagram

The causal diagram for twelve main factors influencing taxi passenger’s user experience is built and drawn as shown in Fig. 1, where the horizontal axis stands for the Prominence value (D+R) while the vertical axis for the Relation value (D−R).

Fig. 1.
figure 1

The causal diagram

Full size image

In the causal diagram, a vector line indicates the direction of influence of one factor to another factor pointed to, and solid lines indicate stronger influencing degree while dotted lines weaker one. As illustrated in Fig. 1, Factor 2 (‘the driver’s familiarity with the local roads’) has stronger influence simultaneously on five factors, i.e., Factor 3 (‘driver’s chatting to passenger(s) when driving’), Factor 4 (‘long detours and intentional slowly driving’), Factor 5 (‘information on the destination’), Factor 7 (‘designated location for pick-up’) and Factor 8 (‘picking others up during taxi ride’). This indicates that whether or not a taxi driver is familiar with local road networks and, as a result, with the corresponding route to the designated destination, plays a pivotal role in determining the satisfaction degree of taxi passenger’s user experience because of its direct influence on driver’s related behaviors, including talking with and inquiring to passengers while driving, detouring and driving slowly, knowing basic information on the destination, and picking passengers up timely at the designated location.

4 Conclusion and Discussion

4.1 Key Factors Influencing the Satisfaction Degree of Taxi Passenger’s User Experience

The top three factors with the largest Prominence values including Factor 2, Factor 4 and Factor 3, and the top three factors with the largest Relation values including Factor 6, Factor 2 and Factor 1, are fetched from Table 4 and reorganized in order as listed in Table 5.

Table 5. The top three factors in Prominence and the top three factors in Relation
Full size table

By reviewing Table 5 and analyzing mentioned-above messages in Fig. 1, it can be concluded that the top three factors with largest Prominence values, i.e., Factor 2 (‘the driver’s familiarity with the local roads’), Factor 4 (‘long detours and intentional slowly driving’) and Factor 3 (‘driver’s chatting to passenger(s) when driving’), are therefore the key factors critical to determining the satisfaction degree of taxi passenger’s user experience when taking a taxi ride. By the way, it shows by reviewing Table 4 that Factor 11 (‘driver’s smoking when driving’), Factor 12 (‘time-phased Pricing’), Factor 6 (‘weather conditions’) and Factor 10 (‘buckling up seat belts when driving’) have the relatively slight influence on the satisfaction degree of taxi passenger’s user experience in taxi ride.

4.2 Relationships of Influencing and Being Influenced Between Main Factors

It is observed in Table 5 that Factor 6 (‘weather conditions’), Factor 2 (‘the driver’s familiarity with the local roads’) and Factor 1 (‘driver’s talking on the cell phone when driving’) are the top three factors with the largest positive Relation values, implying that these three factors have the strong impact upon other factors. Mean-while, the three factors with negative Relation values in Table 4 are Factor 7 (‘designated location for pick-up’), Factor 9 (‘pettish driving mood’) and Factor 4 (‘long detours and intentional slowly driving’), indicating that these factors are ones being influenced most greatly.

4.3 Discussion

In this study, twelve main influencing factors are obtained and defined by analyzing twenty-one preliminary factors collected by means of brainstorming, questionnaire and in-depth interview and so on. And by using DEMATEL method and operation tool, three key factors critical to determining the satisfaction degree of user experience in taxi passenger service are extracted out of these main factors, including (1) the driver’s unfamiliarity with the local roads; (2) long detours and intentional slowly driving; and (3) driver’s chatting to passenger(s) when driving. And furthermore, three factors influencing greatly on other factors are discovered, i.e., (1) weather conditions; (2) the degree of familiarity with the local roads; and (3) whether or not a taxi driver talks on the cell phone when driving.

Among five different factors listed above, it is found that four factors except the non-human factor ‘weather conditions’, are related to the taxi driver side. The findings in this study show that user experience and its degree of satisfaction of taxi passenger during taking taxi ride are related to and influenced by comprehensive factors.

On the one hand, two key factors, ‘the driver’s familiarity with the local roads’ and ‘long detours and intentional slowly driving’, are directly related to the amount of passengers’ payment. This implies that the passenger’s feeling that a taxi driver is unfamiliar with the road networks, has heavily negative effect to the degree of satisfaction of the passenger’s user experience.

On the other hand, the driver’s behaviors stated by one key factor, ‘driver’s chatting to passenger(s) when driving’, and one influencing factor, ‘driver’s talking on the cell phone when driving’, have direct impact upon and are related to driving safety and the passenger’s safety during taxi ride. This indicates that the satisfaction degree of the passenger’s user experience will be also greatly reduced if a taxi driver frequently acts in behaviors such as chatting to passenger(s) or talking on the cell phone while driving.

In addition, it can be elicited from interpreting the causal diagram that whether or not a taxi driver is familiar with the corresponding routes to the passenger’s destination, has also direct impact upon the passenger’s user experience and their evaluation on the degree of satisfaction with taxi passenger service.