National Academies Press: OpenBook

Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook (2008)

Chapter: Chapter 2 - Introduction to Market Research and ITS Data

« Previous: Chapter 1 - Introduction
Page 7
Suggested Citation:"Chapter 2 - Introduction to Market Research and ITS Data." National Academies of Sciences, Engineering, and Medicine. 2008. Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook. Washington, DC: The National Academies Press. doi: 10.17226/13917.
×
Page 7
Page 8
Suggested Citation:"Chapter 2 - Introduction to Market Research and ITS Data." National Academies of Sciences, Engineering, and Medicine. 2008. Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook. Washington, DC: The National Academies Press. doi: 10.17226/13917.
×
Page 8
Page 9
Suggested Citation:"Chapter 2 - Introduction to Market Research and ITS Data." National Academies of Sciences, Engineering, and Medicine. 2008. Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook. Washington, DC: The National Academies Press. doi: 10.17226/13917.
×
Page 9
Page 10
Suggested Citation:"Chapter 2 - Introduction to Market Research and ITS Data." National Academies of Sciences, Engineering, and Medicine. 2008. Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook. Washington, DC: The National Academies Press. doi: 10.17226/13917.
×
Page 10
Page 11
Suggested Citation:"Chapter 2 - Introduction to Market Research and ITS Data." National Academies of Sciences, Engineering, and Medicine. 2008. Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook. Washington, DC: The National Academies Press. doi: 10.17226/13917.
×
Page 11
Page 12
Suggested Citation:"Chapter 2 - Introduction to Market Research and ITS Data." National Academies of Sciences, Engineering, and Medicine. 2008. Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook. Washington, DC: The National Academies Press. doi: 10.17226/13917.
×
Page 12
Page 13
Suggested Citation:"Chapter 2 - Introduction to Market Research and ITS Data." National Academies of Sciences, Engineering, and Medicine. 2008. Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook. Washington, DC: The National Academies Press. doi: 10.17226/13917.
×
Page 13
Page 14
Suggested Citation:"Chapter 2 - Introduction to Market Research and ITS Data." National Academies of Sciences, Engineering, and Medicine. 2008. Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook. Washington, DC: The National Academies Press. doi: 10.17226/13917.
×
Page 14
Page 15
Suggested Citation:"Chapter 2 - Introduction to Market Research and ITS Data." National Academies of Sciences, Engineering, and Medicine. 2008. Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook. Washington, DC: The National Academies Press. doi: 10.17226/13917.
×
Page 15
Page 16
Suggested Citation:"Chapter 2 - Introduction to Market Research and ITS Data." National Academies of Sciences, Engineering, and Medicine. 2008. Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook. Washington, DC: The National Academies Press. doi: 10.17226/13917.
×
Page 16
Page 17
Suggested Citation:"Chapter 2 - Introduction to Market Research and ITS Data." National Academies of Sciences, Engineering, and Medicine. 2008. Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook. Washington, DC: The National Academies Press. doi: 10.17226/13917.
×
Page 17
Page 18
Suggested Citation:"Chapter 2 - Introduction to Market Research and ITS Data." National Academies of Sciences, Engineering, and Medicine. 2008. Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook. Washington, DC: The National Academies Press. doi: 10.17226/13917.
×
Page 18
Page 19
Suggested Citation:"Chapter 2 - Introduction to Market Research and ITS Data." National Academies of Sciences, Engineering, and Medicine. 2008. Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook. Washington, DC: The National Academies Press. doi: 10.17226/13917.
×
Page 19

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

7Understanding Customers: Traditional Market Research Techniques Market researchers in the transit industry collect informa- tion about their customers and the traveling public in a variety of ways. The most commonly used techniques for recovering information include intercept surveys of persons riding the sys- tem or at given locations, telephone and mail surveys of area populations, field observations (recording the delivery and consumption of service), and focus groups. One way of clarifying the distinctions among the various market research techniques is to delineate their key features with respect to the population(s) targeted, the information that is recovered, and the questions analyzed from the in- formation recovered. This is illustrated in the typology pre- sented in Figure 2-1. Each technique is defined by a distinct target population, the specific kinds of information sought, and the research questions that are analyzed. In the practice of market research, it is probably more logical to rearrange the columns of the typology to correspond with the research process. This process usually begins with a set of questions (e.g., “Who is using our system?”). These questions then determine the identity of the subject population and the information that will be sought from them or, rather, their sampled representatives. Questions driving the research process can range from simple and direct to complex and multifaceted. At one end, for example, is the singular, direct question “How many peo- ple are riding our system?” The “analysis” of this question requires recovering information about users of the system. The type of information needed is simple: counts of board- ings and alightings from a representative sample of times/ locations in the system. Among the range of approaches listed in Figure 2-1, the technique that is best suited to recover customer count information at lowest cost or least effort would be a field observation method. This would involve assigning ride checker staff to record boardings and alightings on sampled trips or trip segments. The analysis of the sample data would be limited to expanding the sample counts to infer the system’s total ridership. Much more complex sets of questions reside at the other end of the market research spectrum. For example, the ques- tion “How do riders on our system differ from non-riders?” indicates that the target group is the general population of travelers residing in the service area. The only way to recover the necessary information from this group is through a mail or telephone survey. The information that must be recovered is determined by the desired scope and depth of understand- ing of the distinctions between the two groups. Usually, a threshold information set would cover frequency of use, age, sex, race, ethnicity, income, residence and work locations (if employed), and vehicle ownership. This information set would provide the necessary data for market segmentation analysis along traditional lines. More advanced approaches to market segmentation analysis, however, seek to further dis- tinguish riders and non-riders along attitude, opinion, and preference dimensions. Doing so requires recovery of scaled response information on attitudes, preferences, perceptions, and opinions on a range of topics known to influence per- sons’ travel choices. Incorporating this information into the analysis not only provides an opportunity to gain a greater depth of understanding of the distinctions between riders and non-riders, it also allows for further identification of sub- groups within the two populations. Identifying such subgroups holds great strategic importance in transit marketing programs. Among transit rider subgroups, for instance, one can usually identify a segment whose members would rather not be using the system. Knowing why this is the case is the first step toward taking action to retaining riders who may be on the verge of leaving the system. Some members of this group may prefer another means of travel, but cannot presently afford to act on that preference. Should their incomes increase, there may be little that a marketing campaign can do C H A P T E R 2 Introduction to Market Research and ITS Data

to sway their decision to leave. Other members, however, may be uncomfortable or dissatisfied with their experiences on the system, and getting to the bottom of their perceptions presents an opportunity to take corrective action that will mitigate prob- lems they encounter or perceive. In contrast with rider segments, analysis of attitudes and preferences among non-riders often identifies a subgroup whose members express very favorable opinions, perceptions, attitudes, and preferences toward transit as a travel option, but have not followed through in their actual travel mode choices. For some, the underlying reason may be simple: they lack reasonable access to service. Analysis can then shift to the question of whether this group is sufficiently large and geo- graphically concentrated enough to be cost-effectively served. For others, access may be adequate, but they lack a compelling reason to change their travel choices. Analysis of their opin- ions, attitudes, and preferences may reveal the values that may motivate a change in their choices. This group would be receptive to marketing initiatives promoting transit that res- onate with these values. These two subgroups represent what may be the most im- portant latent markets in the transit industry. Their affiliation with transit lies very close to the choice point with other modes. To build a high ridership system it is necessary to identify these groups and understand what they want. Their allegiance as riders is, at the same time, most readily gained in a successful market development program and most easily lost when market researchers ignore them. In some instances, rather than initiating the market re- search process with a set of questions, analysts turn to a target population for their views on how questions should be defined and interpreted. This can be motivated by an interest in gain- ing a deeper understanding of issues that are believed to be important to customers, or to explore customer reactions to new ideas and issues. For example, regarding the former, sat- isfaction surveys often reveal that service reliability and safety and security are important issues in customers’ minds. Yet, it is also known that riders rarely consult schedules, and crime statistics suggest that the system is commonly safe and secure. Thus it is apparent that riders are expressing perceptions or opinions that may not be reflected in measures such as sched- ule adherence or crime incidence. Probing issues or questions through focus groups may provide better definition of customer perceptions. For ex- ample, riders’ concern with “reliability” may, upon deeper consideration, be revealed to be grounded in the uncertainty they experience wondering when the next vehicle will be arriving. Alternatively, it may turn out to be a reaction to see- ing platoons of vehicles on high frequency routes that are failing to maintain scheduled headways. Probing concerns about “safety and security” may reveal that riders are un- comfortable with unfamiliar places, or when riding with others who are different from them. In both examples, get- ting to the bottom of perceptions can help to identify actions that would improve riders’ satisfaction with service. Posting schedules at stops or, in an ITS world, posting the next vehicle’s expected arrival in real time may be the best action that can be taken to improve riders’ dissatisfaction with reli- ability. Providing information at stops and adding various treatments on vehicles or along riders’ common access and 8 Approach Target Group(s) Information Obtained Questions Analyzed Intercept Surveys: • On-board • On-street Riders Non-Riders Personal Demographics Travel Characteristics Fare Payment Used Attitudes, Preferences, Perceptions & Opinions Who is using our system? How often do they ride? Where are riders coming from and where are they going? Why are they traveling? How are they paying? What are their travel options & preferences? How satisfied are they with their experience on the system? What do they think about a possible change? Telephone & Mail Surveys Riders & Non-riders Household Demographics Travel Characteristics Attitudes, Preferences, Perceptions & Opinions How do riders differ from non-riders? How are travel market segments defined? What should we do to retain & attract riders? How would riders and/or non-riders respond to a service change? Field Observation Riders Agency Employees System Infrastructure System Usage Service Delivery Customer Service Infrastructure Condition How many customers are being served? Is service reliable and on time? How easy is it for riders to use the system? How are riders treated? What are conditions like on the system? Focus Groups Target Group Varies by Topic Perceptions, Opinions, Attitudes & Preferences How would riders and/or non-riders respond to new ways of doing things? What is really important to riders and/or non-riders? What should we do to make the system better? Figure 2-1. Typology of transit market research approaches.

egress paths may be what is needed to address concerns with safety and security. Focus groups can provide valuable information in the con- sideration of new practices or systems. In the area of fare pay- ment, for example, transit providers can turn to focus groups to gain insight on customer perceptions and opinions of a new fare option or a new system for fare payment. Consider- able time and resource investments are at stake in each case, and focus group analysis can serve to reduce investment risk by gauging market acceptance. Some of the most elementary yet critical market research questions are analyzed from information gathered through field observation techniques. Some of these questions focus on vehicle operators who, in delivering service, represent the principal contact between a transit agency and its customers. Do they operate their vehicles with customers’ safety and comfort in mind? Do they treat riders with respect and re- spond to their questions? Do they announce stops? Do they provide assistance when requested to riders with disabilities? The information required to evaluate these questions is com- monly recovered by “mystery shoppers,” who pose as riders and record observations on riders’ experiences. The knowl- edge gained from evaluating this information can help to identify areas needing greater emphasis in ongoing training programs. Operators themselves and other field personnel have tra- ditionally served as sources of information on conditions in the system affecting riders’ travel experiences. Operators are relied on to report problems with the state of repair and cleanliness of vehicles, stops and stations, potential hazards that riders and others may encounter, and situations that threaten rider safety or the safe operation of vehicles. More informally, vehicle operators have also commonly reported back their assessment of the adequacy of schedules, the lay- out of routes, and the location of stops. Field observation approaches also include recovery of basic service delivery data by staff “ride checkers.” These data include boarding, alighting, and load counts; vehicle running times; and schedule, headway, and timed transfer adherence. The scope and frequency of ride checker data recovery are defined, at a minimum, by the annual reporting requirements of the Federal Transit Administration’s (FTA) National Transit Database (NTD) program. Beyond NTD reporting, transit agencies commonly seek more extensive recovery of service delivery data to support service planning, scheduling, and operations management needs. Intercept surveys, fielded either on-board vehicles or at specific locations, are the workhorses of most transit market research programs. Much of what a transit agency under- stands about its customers is learned through analysis of intercept survey data. On-board surveys can provide demo- graphic profiles of riders on the system; information on the purposes of their trips; where their trips begin and end; the path of their travel through the system; how they pay for ser- vice; their dependence on transit to meet their travel needs; their preference for transit in relation to other modes; and their satisfaction with the services provided. On-street sur- veys recover information about how people travel to or from given locations; and their perceptions, preferences, and opin- ions about location, program, or product-related attributes of service delivery. Information from O-D surveys provides a foundation for the design and delivery of transit service to a community. Un- less more detailed or extensive travel information is being sought, O-D surveys are designed to be fielded on transit vehicles. An example of an on-board O-D survey instrument is shown in Figure 2-2. This instrument was fielded on TriMet’s bus, light rail, and streetcar system in 2005. The figure provides a useful illustration of several features of an on-board survey. First, it is designed to recover as much in- formation as can be reasonably expected during the course of a typical rider’s time on a vehicle. Second, despite the re- sponse time limitation, the scope of information sought in the instrument nevertheless extends beyond that needed to identify trip characteristics to include limited examples of most other types of information recovered in on-board sur- veys. This additional information allows researchers not only to document travel patterns occurring on the system, but to relate these patterns to personal characteristics. The typology of transit market research approaches pre- sented in this section does not include information that flows into the agency from customer-initiated contacts by tele- phone or the Web. Information from these contacts includes commendations, suggestions, and complaints, as well as travel queries recorded by trip planning software and real time vehicle arrival queries. Logs of the information obtained from customer-initiated contacts are usually maintained for tracking and analysis. In particular, the incidence of com- mendations and complaints is usually included among the system performance indicators that are closely tracked by senior management. What distinguishes customer-initiated information from the information recovered by the approaches identified in the market research typology is the ability of the latter to sup- port inferences from a limited number of observations to a definable population. For surveys and field observation ap- proaches, the ability to make such inferences is ensured by following statistical sampling plans. For focus groups, it is ensured through careful screening of participants to repre- sent an intended audience. Alternatively, inferences are diffi- cult, if not impossible to make from customer-initiated data. For example, it is hard to know what population is being rep- resented by persons querying trip planning software. Thus, it is very unlikely that a trip table “inferred” from trip planning 9

10 Figure 2-2. TriMet origin-destination survey instrument.

software logs would correspond to one inferred from an O-D survey. Likewise, “inferences” from customer complaints, which often refer to a person’s unpleasant encounter with an operator, are unlikely to correspond to riders’ assessment of their treatment by operators from a customer satisfaction survey. Subsequent sections of this report include the use of ITS data from customer-initiated contacts. It should be noted that ap- plications drawing on such data mainly support customer serv- ice and human resource rather than market research functions. Inventory of ITS Data for Market Research Deployment The transit industry’s use of ITS data for market research is conditioned by the extent of deployment of the respective technologies. The deployment status of advanced public transportation technologies has been systematically tracked over time by the U.S. DOT’s Volpe National Transportation Systems Center. The Volpe Center’s most recent report docu- ments ITS deployment in 2004 (Volpe Center 2005). Data on ITS deployment status in 2004 were recovered by a survey of 516 transit properties that report information to the NTD. There were 327 responses to the survey. While the deployment status of advanced technologies among non-respondents is unknown, it is reasonable to assume that properties with operational or planned technologies were more likely to respond to the Volpe Center survey than those without the technologies. Deployment information from the 2005 Volpe Center re- port is presented in Table 2-1 for the technologies of interest to this Guidebook. Information in the table is presented for 2004 and 1995, with the latter being the year covered by the Volpe Center’s first deployment survey. For AVL, the num- ber of properties with operational systems grew from 22 in 1995 to 157 in 2004, a 614% increase. Adding planned de- ployments, the number of AVL properties grew from 86 in 1995 to 257 in 2004, an increase of 199%. Under the limiting assumption that none of the non-responding properties had or were planning to deploy an AVL system, the correspon- ding industry penetration rates for this technology in 2004 were 30.4% (operational) and 49.8% (operational plus planned). The deployment of APC systems in the transit industry has been more limited than AVL deployment. Only 11 properties reported operational APC systems in 1995, and the growth to 75 properties in 2004 resulted in an industry penetration rate just under half the corresponding AVL rate. Adding planned APC deployments yields totals of 32 properties in 1995 and 129 properties in 2004, with industry penetration of 25%. Departing from previous surveys, the Volpe Center did not report fleet coverage information for vehicle-related tech- nologies in its 2005 report. Earlier reports indicate that AVL and fare payment technologies are commonly deployed fleetwide, while APC systems are not. A general “rule-of- thumb” in the industry is that 15% fleet coverage is the minimum necessary for APCs to satisfy NTD reporting requirements. Anecdotal evidence suggests that properties with APCs are increasingly deploying the systems well beyond the rule-of-thumb threshold in order to support internal reporting and analysis needs. Coverage of electronic fare payment technologies in the Volpe Center report is limited to magnetic stripe and smart card systems. In 2004, there were 159 properties with opera- tional card systems, up from 22 properties in 1995. Adding planned systems increases the respective totals to 65 and 293. 11 AVL APC Electronic Fare P’mt Automated Transit Info. 1995 Operational 22 11 22 48 Operational + Planned 86 32 65 93 2004 Operational 157 75 159 488 Operational + Planned 257 129 293 512 Percentage Change, 1995-2004 Operational 614% 582% 623% 917% Operational + Planned 199% 303% 351% 451% Industry Penetration, 2004 Operational 30.4% 14.5% 30.8% 94.6% Operational + Planned 49.8% 25.0% 56.8% 99.2% Properties Surveyed, 2004 516 516 516 516 Table 2-1. ITS deployment in the transit industry, 1995–2004.

The industry penetration rates for electronic fare payment systems are comparable to the rates for AVL systems. Automated transit information technologies are defined as “. . . systems that provide information to the public, without human intervention . . .” (Volpe Center 2005: 122). The media through which information is provided include the Web, automated telephone systems, automated voice an- nunciation systems (AVA), signboards, cell phones, pagers, and personal digital assistants. The types of information pro- vided include schedules and fares, trip plans, and real time vehicle arrival times. The deployment data on this collection of technologies indicate that survey staff was able to recover information about virtually all non-responding properties, most likely by visiting the properties’ websites and calling their phone systems. In 2004, 488 operational and 512 oper- ational plus planned systems were identified, compared to 48 and 93, respectively, in 1995. The 2004 figures indicated that the penetration level in the industry is near-complete. In addition to the technologies listed in Table 2-1, the Volpe Center began reporting of the deployment of mobile data terminals in its 2002 survey. When installed on transit vehicles and integrated with AVL systems, mobile data ter- minals can record pre-programmed events when operators press a key. There are a number of conceivable events whose documentation would provide useful information to market researchers. The survey reported 115 transit properties with operational mobile data terminals and 137 additional prop- erties with planned deployments in 2004. This corresponds with industry penetration estimates of 22.3% (operational) and 48.8% (operational plus planned). There are a number of general conclusions that can be drawn from the information in Table 2-1 and from closer in- spection of property-level data provided in the series of Volpe Center deployment reports: • The pace of technology deployment over the 10-year pe- riod has been rapid, suggesting parallel transformations in data archiving infrastructure, data reporting and analysis practices, staffing needs, and skill requirements. • ITS deployment was initially concentrated among larger properties and later spread to smaller properties. Current deployment levels are still higher among larger properties. Two implications follow from these observations: (1) there has been an opportunity for smaller properties to learn from the ITS deployment experiences of larger prop- erties, through peer exchange and other means of commu- nication and (2) if industry penetration measures were based on the number of customers affected rather than on properties, the extent and impact of deployment would be much greater than indicated in Table 2-1. • Among electronic fare technology deployments and planned deployments over the decade, smart cards have become increasingly favored over magnetic stripe cards. Especially when registered, smart cards are capable of re- covering more information about riders and their use of the system. • The deployment of APCs has not been as extensive as the other technologies covered in Table 2-1, yet the passenger counts provided by these systems hold great potential for leveraging traditional market research applications. ITS Data Inventory The ITS technologies recovering data for market research can be grouped into two categories. The first includes systems that are installed on vehicles, including AVL, AVM, MDTs, and control heads, APC, magnetic and smart card readers, and electronic registering fareboxes. The second category includes systems supporting customer service, including ticket vending machines, automated phone systems, and the Web. Traffic counting systems are also included in the second category. Data elements recovered from on-board systems are pre- sented in Figure 2-3. The role of AVL systems in recording time and location information is central to the viability of other on-board systems, in addition to its independent use- fulness. When integrated with AVL, the data recorded by the other on-board systems become identified with respect to where and when a specific event occurred. With systems in- tegration, the records generated through AVL also serve as the basis for aggregating other on-board data, from unique events that are recorded at specific locations and times, to summaries at higher levels, such as route and system (spa- tially) and hour and month (temporally). The integration of AVL with other on-board technologies has been problematic, especially for properties that acquired on-board systems over time in a piecemeal approach (Casey 2000). Casey pointed out that the complex proprietary soft- ware provided with the early systems hampered integration efforts, as did each property’s desire to customize a given sys- tem to meet its specific needs. Integration problems have been serious enough in some cases to require installing multiple AVL units on a vehicle, with each dedicated to pro- viding time-location referencing to one or several on-board systems. AVL integration with other on-board systems has im- proved, following the development of the “smart bus” con- cept (Furth et al. 2006). The concept is organized around use of a vehicle logic unit for storing data recovered from on- board systems. On-board systems are now manufactured to comply with the J1708 integration standards used in most AVL systems (Society of Automotive Engineers 1993, 1996). Focusing on AVL independent of the other systems, the most useful data for market researchers are the time-location 12

vehicle status data. At the lowest level of aggregation these data are collected at stops and route origins and destinations. Aggregating the time data over all route locations yields a vehicle’s actual running time for a given trip. Analysis of the pattern of running times for all trips for specified time periods reveals typical running times and the variability of running times for a route, both of which contribute important infor- mation to the schedule writing process (Levinson 1991). Time-location data recovered by AVL systems can also be used to assess deviations in the actual delivery of service relative to the schedule. Analyses of such deviations across time points at the route and system level are the basis for reporting on-time performance. The general practice in the transit industry has been to define service as being “on time” when departures from time points are no more than 5 min. late or 1 min. early (Bates 1986). The pattern of depart times recorded by AVL for successive vehicles can also be compared against the scheduled head- ways to assess the degree of regularity in the service that is de- livered to customers. Especially in frequent service situations, a deterioration in regularity results in “bus bunching,” which leads to longer waits for riders, a higher incidence of crowd- ing on vehicles, and a loss of effective service capacity for the transit agency. The ability to know where and when such problems occur facilitates the operations control process (Strathman et al. 2003). For some properties, the additional waiting time resulting from bus bunching has been standard- ized into performance measures that are reported to senior management (see the CTA and TriMet case studies, Appen- dices A and C, respectively). AVL systems also record the location status of vehicles at specified time intervals (ranging from 30 to 90 s). These are re- ferred to as poll data. Poll data are transmitted over the vehi- cle’s radio system to the dispatching center. In addition to supporting dispatch functions, AVL poll data are used in real time to support software that broadcasts vehicle arrival time estimates through an agency’s website or automated phone system. While real time use of AVL poll data has benefited dis- patchers and customers, archived poll data are of limited use- fulness to market researchers because they cannot be easily 13 Automatic Vehicle Location 1. Date & time 2. Vehicle ID 3. Route ID 4. Trip ID 5. Operator ID 6. Location (latitude & longitude, Stop ID) 7. Direction 8. Arrive Time 9. Depart Time 10. Speed (average or maximum, from previous location) Automatic Vehicle Monitoring 1. Door open & close 2. Lift deployment 3. Signal priority request Mobile Data Terminal/Vehicle Control Head 1. Event Code (fare evasion, pass up/overload, security issue, etc.) Automatic Passenger Counter 1. Boardings 2. Alightings 3. Load Electronic Farebox 1. Transaction ID (if transaction-based) 2. Transaction Type (ticket/token/cash) 3. Transaction Amount 4. Ridership (operator initiated) 5. Event (operator initiated, e.g., fare evasion, lift use, pass use, transfer use, etc.) Magnetic Stripe Card (Reader) 1. Card ID 2. Card affiliation (when registered) 3. Transaction ID 4. Transaction Amount Smart Card (Reader) 1. Card ID 2. Person ID (when registered) 3. Transaction ID 4. Transaction Amount Figure 2-3. Inventory of ITS data for transit market research: on-vehicle systems.

joined with other ITS data (given the absence of common lo- cational references). By counting boardings and alightings, and calculating pas- senger loads, APC systems provide important data to market researchers and service planners. These passenger data, com- piled at the stop, route, and system levels, establish the sampling frames for rider surveys and provide the expansion factors needed to infer survey results to the riding population. The in- tegration of APC with AVL systems has greatly improved the quality of passenger count data. Prior to AVL, passenger count data from APCs were related to stops by inferring location from time and odometer stamps on the data records, a clumsy process that screened out much of the data recovered (Strathman and Hopper 1991). AVM technology recovers data on a vehicle’s mechanical and electrical systems. While AVM data primarily support maintenance functions, they are sometimes useful for market research. AVM data on lift deployments, for example, document where riders with more serious mobility impair- ments access the system, supporting surveys of this special population. AVM systems also record the transmission of sig- nal priority requests when this function exists. These data could be coordinated with corresponding signal system data to assess time savings for riders. MDTs and control heads allow operators to record prede- fined events by pressing an icon or button. Conceivably, the range of “events” that affect customers’ riding experience is extensive. These systems are capable of recording when and where such events occur. Event data commonly recovered by these devices include overloads and pass-ups, safety and security incidents, fare evasions, medical emergencies, and breakdowns (see the TriMet case study in Appendix C). Electronic registering fareboxes record fare transactions data, including fare type and transaction amount. Fareboxes that are equipped with keypads also allow operators to record the fare media used and other predefined events (as described for MDTs and control heads). Among the advanced technology systems covered in Fig- ure 2-3, electronic registering fareboxes have often been the first to be installed on a transit property’s vehicles. When AVL systems were subsequently acquired, their integration with existing fareboxes was sometimes not feasible or was not pursued. Without AVL integration, farebox data are defined by vehicle and time stamps, resulting in similar location referencing problems experienced by APC systems in the pre- AVL days. Where electronic farebox and AVL systems have not been integrated, some effort has been made to identify transaction locations by “matching” farebox and AVL time stamps (Cui 2006). Magnetic stripe and smart card systems recover fare trans- actions data of the cardholders. For each transaction, magnetic card systems record the identity of the card and the transaction amount. Magnetic stripe card IDs can be registered, making the transactions of defined groups identifiable. For employer and university/school pass programs, this feature allows reli- able documentation of usage and supports program develop- ment efforts (see the City of Madison Metro Transit case study in Appendix B). Smart cards have stored and recharge value ca- pability and can be registered to individual users in connection with personal accounts. Thus the transaction data from smart card systems can identify individuals when they are registered. Through integration with AVL systems, the transactions data for smart and magnetic card systems typically identify riders’ boarding locations. Analysis of the time sequence of transactions, however, appears to be fairly successful in inferring riders’ alighting locations (Rahbee and Czerwinski 2002). Similarly, when electronic fareboxes are integrated with AVL systems, summary totals for exit locations have been successfully inferred from entry location totals (Navick and Furth 2002). Data elements for the remaining technologies covered in this Guidebook are presented in Figure 2-4. Data from ticket vending machines include each machine’s location, the date and time of each transaction, the number and type of fare items purchased, the price per item, the total value of each transaction, the method of payment; and the status of the transaction. Web tracking software documents characteristics of Internet contacts, including the date, time, and duration of each visit; the entry page, the page path through the site and exit page; and the identification of the referring website. For websites that broadcast vehicle arrivals in real time, tracking software record the route and stop location selected and the time of the request. For websites with trip planning software, tracking data cover the requests related to time of travel, origin and destination locations, preferred travel options (e.g., shortest path or time, fewest transfers), and the final travel itinerary. Most websites include features that allow customers to communicate commendations, sugges- tions, and complaints to the agency. In responding to these communications, customer service staff usually documents characteristics of the communication (e.g., type, date, time, route, location, and vehicle/operator identification, where possible). Tracking software for automated telephone systems doc- ument the characteristics of each call received, including the date, time, and duration of the call; the routing of the call through the system; and the origin of the call. Similar to the Web, for phone systems with trip planning and real time vehicle arrival services, tracking software documents char- acteristics of the information requested. Documentation of customer communication of commendations, requests, and complaints via the phone system also occurs in a simi- lar fashion. 14

Traffic loop detectors are the final technology given in Figure 2-4. The data recorded by loop detectors include location, direction of traffic, date, and time; and vehicle classification, count, speed, and occupancy for a defined time interval. Benefits and Limitations of Combining Traditional and ITS Data While it may be tempting to contrast the “anonymous customers” represented in ITS data with the “identified cus- tomers” in traditional market research data, such a dichotomy masks the differentiation among ITS technologies in their abil- ity to define customers. It is true that ITS data cannot capture customers’ stated attitudes, opinions, perceptions, and prefer- ences, the core information sought in traditional market research. At the same time, if individually stated preferences occupy the core of the market research paradigm, then ITS data form rings of highly useful information around the core. Much of this information would otherwise be gathered only by using more costly and irregular manual collection techniques, if at all. Some ITS data offer entirely new sources of information or new data combinations. Furthermore, the layers of data around the core of traditional market research can improve understanding by linking customer perceptions and attitudes to extended data from ITS technologies. Finally, ITS data can serve an enabling role by assisting the collection and analysis of data gathered using traditional techniques. Figure 2-5 displays the range of ITS technologies and their respective functions, applications, and resolutions related to the primary objectives of market research. As one moves from the periphery toward the center of the figure, the “cus- tomer resolution” improves; that is, the characteristics and travel activities of specific customers and customer groups become increasingly identifiable. At the center of the diagram is the highest level of customer resolution, representing the traditional market research goal of uncovering customers’ preferences and perceptions. 15 Ticket Vending Machines (TVM) 1. TVM ID 2. TVM Location (latitude/longitude) 3. Transaction Date and Time 4. Article Purchased (e.g., Adult All Zone pass, Youth/Student ticket, etc.) 5. Number of Article Units Purchased 6. Price per Article Unit 7. Type of Purchase (cash, credit, etc.) 8. Total Amount of Sale 9. Transaction Status (complete, cancelled, etc.) 10. Transaction Characteristics (credit card number, expiry date, service host identifier, etc.) Website Tracking Software 1. Date and Time of Visit 2. Duration of Visit 3. Pages Viewed 4. Entry Page 5. Exit Page 6. Path Through Site 7. Referrer (outside site referring visitor to host website) 8. Route Selected (for schedule or real time information queries) 9. Stop (for real time information requests) 10. Origin/Destination/Time/Path (for trip planning queries) 11. Communication Received (e.g., commendation/suggestion/complaint) 12. Files Downloaded Automated Telephone Systems 1. Call ID 2. Date and Time of Call 3. Duration of Call 4. Number Called 5. Caller ID 6. Call Abandoned 7. Call Routing (e.g. trip planning, real time info., complaints, lost and found) Traffic Counters (Loop Detectors) 1. Date and Time 2. Location (latitude/longitude, street/highway name; jurisdiction) 3. Direction 4. Vehicle Count 5. Vehicle Classification 6. Speed 7. Occupancy Figure 2-4. Inventory of ITS data for transit market research: other systems.

In the outermost ring of Figure 2-5, ITS technologies, such as AVL, provide information related to service delivery. At this level, customer identities and characteristics are com- pletely unresolved within the data. One sees only the general service environment that a hypothetical customer would encounter. This level of resolution is suited to assessing re- search questions about the characteristics of service delivered to customers, such as on-time performance. Moving in to the second ring, technologies such as APC, magnetic stripe and smart cards, and Web and phone logs enumerate customers. At this level, anonymous customers can be located and counted, and these counts can address questions pertaining to when, where, and how many customers are using the system. Or, in the case of Web and phone logs, questions can be answered about how many customers are requesting information about particular services or are communicating information back to the agency. Moving in to the third ring, technologies such as smart cards, magnetic stripe cards, and Mobile Data Recorders can relate data to specific customer groups. For example, a smart card or magnetic stripe card can be linked to a specific pass program participant, or records of lift deployments could link customers with disabilities to specific stops or trips. The data in this ring are useful for answering research questions about which customer groups are using a specific service or the system as a whole. Finally, in the ring closest to the traditional market re- search core, technologies such as smart cards and magnetic stripe cards can potentially document the travel activity of an individual customer using the unique ID associated with each card. This ring represents the highest level of customer resolution obtainable with ITS technologies. Individual cus- tomers can be tracked through time and space within the sys- tem. For example, the transfer patterns of a specific customer could be tracked through the system as linked trips and com- pared over time. Deeper analysis of data at this level can even begin to reveal customer preferences. For example, given sub- stitutable nearby transit services, the choice of one service over another may reveal a customer’s valuation of selected service characteristics. Zhao (2004) uses analysis of this type to determine how CTA rail customers trade off travel time for comfort in the form of available seats and ease of transfers. Thus, ITS technologies collect transit data over a wide range of customer resolution levels, from general service delivery characteristics to individual customer actions and choices. It is important to recognize the linked nature of ITS data. Because most transit ITS data are time and location stamped, data from different “rings” can often be joined. For this rea- son, it is better to think of the data layers moving toward the center as being complementary and cumulative. For instance, using the unique ID of an electronic fare card, a customer can be tracked through the system. These data exist within the customer behavior ring. Working toward the periphery, additional ITS data such as fellow rider attributes (customer attributes), passenger loads (customer enumeration), and schedule adherence (service delivery) could all be linked to 16 AVL EFB: electronic registering farebox APC EFB WEB SC MAG Event SC MAG Surveyed attitudes and preferences Customer Resolution Figure 2-5. ITS resolution in a market research context.

the individual customer’s travel, providing useful context data for analysis of travel choices. Relevant ITS data from any level can also be linked to the traditional market research core. For example, APC load data and AVL reliability data could be linked to surveyed customer perceptions of crowd- ing and reliability, respectively, providing an opportunity for comparing perceived and actual conditions. ITS data cannot be easily generalized. Different technolo- gies record data at different levels of resolution. The resulting data have widely varying applications in market research. While ITS data cannot replace the core of traditional market research—individually stated preferences—in many cases, ITS technologies provide lower cost and expanded data that support and extend the core. The ability to use and link data from different resolution levels allows a more complete understanding of customer preferences and behavior. When combined, ITS and traditional data are highly complemen- tary in addressing market research questions. ITS technologies gather data on many aspects of service de- livery and customer activity continuously, systemwide, and at low cost when compared to manual data collection. Tradi- tionally, market researchers have relied on manual collection for such service delivery and consumption data. In an inte- grated market research system, different ITS data elements may replace, extend, or complement traditional market research data. However, certain information and data, especially those related to attitudes and non-rider characteristics, remain within the exclusive purview of traditional market research methods. Figure 2-6 presents a graphical representation of the po- tential contribution of ITS data in six traditional market research applications: attitude studies, market segmentation analysis, customer satisfaction surveys, O-D studies, fare studies, and area analysis. Each research application’s diagram in the figure provides a rough portrayal of how ITS data relate to traditional market research data sources, and also notes the primary data available relevant to the specific application. First, the area uniquely within each rectangle in the figure represents the data requirements for each traditional mar- ket research application in the absence of ITS data. Next, the area uniquely within the ovals represents the range of data that ITS technologies make available. Finally, the area of overlap represents the opportunity for ITS data to combine with traditional market research data in each application. In this role, ITS data can facilitate or enrich traditional data collection and analysis. For instance, a customer satisfaction study on crowding could use APC-generated passenger load data to determine sampling times and locations. In addi- tion, surveyed satisfaction data on crowding could be com- pared with actual passenger loads to relate perceptions of overcrowding to its actual incidence. Ideally, the APC data would be drawn from the actual trip where the survey was administered. 17 Attitude Studies Area Analysis Market Segmentation Fare Studies O-D Studies Customer Satisfaction Rider & non-rider attitudes covering all aspects of the transit service & traveling experience Actual behavior related to attitudes; inferred attitudes Rider & non-rider attitudes and behavior Continuous data on usage; service delivery by user-defined analysis areas Surveyed fare usage and preferences Continuous, comprehensive fare usage data and immediate ridership data pre & post fare changes Continuous O-D estimation within system; online (unidentified) trip requests O-D surveys or person trip diaries Surveyed satisfaction Surveyed satisfaction linked to continuous data on service quality indicators Riders & non-riders surveyed usage, demographics, and attitudes By time, day, season, area; limited demographic data Figure 2-6. Traditional and ITS data in market research applications.

Traditional market segmentation studies survey riders and non-riders to delineate distinct groups based on demograph- ics, reported travel characteristics, and attitudes. ITS data collected from APCs and electronic fare cards can extend traditional data by providing continuous records on actual system use by time and location. For instance, using APC data, market researchers can segment different temporal user groups (e.g. time of day, day of week, or season) and analyze each group by location, ridership share, and trends over time. For example, using smart card data, Utsunomiya et al. (2006) segment CTA customers by residential location, transfer pat- tern, route/stop consistency of use, general frequency of use, and system access distance. Finally, ITS data can be combined with traditional data to enable or enrich market segmentation studies. A study seeking to segment off-peak riders, for in- stance, could use off-peak counts from APCs to target surveys or to provide sampling weights for expanding the survey re- sults. APC data could also enrich traditional data by linking known market segments to actual system use. For example, areas with a preponderance of certain attitudinal groups (e.g., “transit lifestyle” or “necessity riders”) could be linked with the areas’ actual ridership data and further analyzed along temporal dimensions (e.g., weekday/weekend, peak/off-peak, seasonal). Customer satisfaction is usually gauged by surveys. ITS technologies allow stated satisfaction and customer feedback to be linked with actual service delivery data. For example, TriMet (Appendix C) has compared surveyed perceptions of “overcrowding” with actual passenger loads on specific trips recorded by APCs. The comparison provides a more nuanced understanding of how actual conditions relate to customer perceptions and satisfaction. In addition, important factors influencing customer satisfaction that emerge from satisfac- tion surveys may be monitored continuously to better under- stand trends between surveys. This monitoring function may permit earlier market research findings to be used preemp- tively to forestall decreases in customer satisfaction. For ex- ample, periodic satisfaction surveys at TriMet (Appendix C) revealed that both reliability and overcrowding were increas- ingly affecting overall rider satisfaction. Analysis of AVL and APC data suggested that deteriorating headway maintenance (i.e., bus bunching) was related to the downward satisfaction trend. Thus, the combination of traditional and ITS data within market research can provide managers with both satis- faction metrics and related service delivery measures to address conditions related to customers’ concerns. In this example, operations managers at TriMet became aware of the importance of maintaining headways in order to improve customer satisfaction rather than simply promoting headway adherence as “good operations practice.” O-D studies are conducted by surveying riders or a sample of households and extrapolating the data to construct trip tables. ITS data can both support traditional survey tech- niques and provide independent O-D estimates, with or without survey-based validation. Survey-based O-D trip tables can be checked for correspondence with APC or smart card data. APC boarding/alighting totals by stop can be com- pared with survey-based estimates, or individual household travel surveys can be compared with actual transit trips recorded by household smart cards over the survey period. CTA plans to use both of these approaches in future O-D survey efforts (see Appendix A). Furthermore, ITS technolo- gies can provide either direct O-D data or inputs suitable for O-D estimation. Rahbee and Czerwinski (2002) report accu- rate O-D estimates for rail passengers using fare card data on boardings only. As electronic fare card use becomes more widespread, ITS data also promise low-cost updating of trip tables between the times when traditional surveys are undertaken. Fare studies are traditionally undertaken using a combina- tion of surveyed fare preferences and manually collected fare usage data. ITS data from electronic fare cards and electronic fareboxes can largely replace manually collected usage data and provide continuous, accurate fare usage data. For exam- ple, City of Madison Metro Transit (Appendix B) has used magnetic stripe card data on pass program users in negotiat- ing pass program contracts with employers and institutions in their district. ITS data can also support ongoing monitor- ing of customer responses to fare changes and validation of survey-based fare models. The CTA (Appendix A) analyzed actual fare card usage data in updating its survey-based fare change model. APC data might be used to monitor ridership changes following a fare change. ITS data on fare use can also combine with traditional surveys to better understand changes among different market segments. CTA compared demographics and perceptions of smart card users with those of other riders and non-riders to better understand why cer- tain customer groups prefer specific fare options. Thus, ITS data on fare choice and response to fare changes improves the currency of fare data and extends analysis possibilities. Area analysis is often performed by surveying riders and non-riders in a defined geographic area, and recovering their attitudes and preferences regarding existing or planned tran- sit service. Because most ITS data are location stamped, or- ganizing them around specific analysis areas is relatively easy. Furthermore, the ability to define the analysis area after data are collected provides a considerable advantage over tradi- tional surveys. ITS data provide continuously updated infor- mation on both service delivery and usage within any defined analysis area. For example, Utsunomiya et al. (2006) examine consistency of station and route choices in different areas using smart card data, finding that some areas’ usage patterns vary considerably more than others. Such analysis provides the market researcher with a better understanding of the 18

customer types served in specific areas. There is also potential for leveraging ITS data for specific areas with survey-based area analysis. Perceptions about service quality could be com- pared with actual area service performance to determine whether, for instance, an area with below average perceptions of reliability actually experiences lower than average reliabil- ity, as measured by AVL. Attitude studies constitute much of the core activities of traditional market research practice and are typically admin- istered by intercept, phone, or mail methods. While ITS tech- nologies cannot recover customer attitudes directly, there are opportunities for ITS data to support and extend traditional attitude studies. In the simplest case, ITS data on stop-level ridership can be useful for planning and expanding user atti- tude surveys. Surveyed attitudes and attitude trends can also be compared with corresponding service delivery trends to discern whether customers are becoming more or less sensi- tive to specific service quality conditions or whether the service conditions themselves are changing. In addition, in certain cases ITS data provide a sufficient basis to directly infer customer preferences. For example, Zhao (2004) estimates the implied monetary value of service amenities like seat availability and ease of transfer using smart card data on pas- sengers’ choices among substitute rail services. Thus, while traditional techniques remain the primary tool for gathering rider and non-rider attitudes, ITS data can in many cases facilitate data collection or enrich the interpretation of results. It should be mentioned that Figure 2-6 simplifies poten- tial research applications somewhat. In fact, data typically associated with one application may be useful for answering questions in other applications as well. This is especially true with combined traditional and ITS data. For instance, the CTA has considered combining fare card data on college stu- dent pass riders (fare study) with surveyed perceptions of safety (customer satisfaction) to assess whether the presence of college students on certain routes improves riders’ per- ceptions of safety. Thus, the diagram probably understates the total contribution of ITS data to each market research application. ITS data provide useful inputs for various market research applications. In many cases, ITS technologies provide data that are useful for addressing market research questions directly. Some of the data are entirely new and some replace traditional data gathering techniques at lower cost and with improved cur- rency. In other cases, ITS data enable or enrich traditional data gathering and analysis. Finally, some data remain within the exclusive purview of traditional market research methods. The size and scope of the contribution of ITS data to market research vary by application and by the data analysis capabili- ties of a transit agency’s market research staff. 19

Next: Chapter 3 - ITS Data Applications in Market Research »
Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB's Transit Cooperative Research Program (TCRP) Report 126: Leveraging ITS Data for Transit Market Research: A Practitioner's Guidebook examines intelligent transportation systems (ITS) and Transit ITS technologies currently in use, explores their potential to provide market research data, and presents methods for collecting and analyzing these data. The guidebook also highlights three case studies that illustrate how ITS data have been used to improve market research practices.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!