Office of Operations
21st Century Operations Using 21st Century Technologies

Expanding Traveler Choices through the Use of Incentives: A Compendium of Examples

3. Applying Incentives to Shift Time of Travel

Incentivizing travelers to change their time of travel can distribute the use of transportation facilities over a longer period and mitigate the crush of peak congestion, especially in highly urbanized areas where thousands of people often attempt to access these facilities at the same time. These shifts are also important to manage demand during special events. For example, businesses in a popular Colorado ski town would offer incentives to skiers to motivate some to stay longer and avoid the congested I-70 Mountain Pass during the prime departure time each weekend.

Encouraging a Shift to Off-Peak Travel Times

  • Online resources.
  • Point-based rewards.
  • Carpool matching.

Metropia is a for-profit enterprise that offers Metropia Total Mobility, a smartphone-based platform to support congestion management strategies and policies. Metropia's Total Mobility platform, which uses artificial intelligence (AI)-based algorithms, data analytics, and behavioral economics, provides a demand management framework that supports shifts in travel by time of day, mode, and route.

To support adjustments to off-peak travel, Metropia's platform allows users to compare routes and time-of-departure choices, depicting travel time savings as well as informing the user about the external impacts of their choices, such as reductions or increases in carbon dioxide (CO2) emissions. The goal is to incentivize commuters to shift their departure to off-peak travel times using a system of reward points. By making this shift to what are called the "shoulders" of the peak period, users are able to earn more reward points, based on a variable point profile, than they would by departing during the peak periods.

In earlier versions of the platform, the level of rewards provided by the system depended on the travelers' degree of behavioral change and their contribution to traffic congestion alleviation. The next generation version of the platform (expected for release in late 2018), utilizes a behavior engine (INDUCE), where rewards are based on the traveler's "persona" (i.e., observed travel behaviors, information captured from micro-surveys, and inferred activity type from destinations and time of day). By using AI-based algorithms, parameters are then matched to user persona characteristics to compute a likelihood of change coefficient. Users with high coefficients are more likely to try a suggested alternative. The value of INDUCE is that it helps pair attractive mobility options with more receptive users, with the ultimate goal of convincing the user to switch to a more sustainable mode (e.g., transit, carpool, etc.). Users can redeem points within the app for rewards such as gift cards from national retailers, gift cards from local stores and restaurants (supplied by partners), or they can redeem their points to plant a tree.

Metropia incorporates a user communication module to alert users to planned construction zones and lane closures as well as any unplanned events such as a flooding, incident, or signal outage at a major intersection, which may impact their trip. Additional in-application features include:

  • A message inbox, where transportation agencies can directly inform users about planned and unplanned roadwork, lane closures, new corridor openings, traffic incidents, etc.
  • Turn-by-turn navigation, guiding users along the fastest route possible. This is designed to identify underutilized roadways when traffic levels rise and reward drivers who opt to use those routes as a means of easing congestion on more overburdened roadways.
Screenshot of the Metropia app.
Figure 3. Screenshot. Rewards from Metropia program.
Source: Metropia.

The social carpooling "Driving Up Occupancy" (DUO) module of the platform allows users to arrange carpools by creating or joining groups. The app verifies passenger participation in the carpool, and each person receives reward points. The reward points are based on a spinning wheel incorporated in the smartphone app as part of the gamification element of the platform. The driver also receives a portion of each passenger's points as an additional reward. Another feature is Metropia's integration with transportation network companies (TNC), which allows the user to request to share a TNC ride with others, rewarding them with additional points for opting to share their ride.

Over the past 3 years, Metropia has conducted several experiments on changing traveler behavior. To test drivers' willingness to shift travel times, Metropia revised its point distribution system and conducted a month-long experiment among its El Paso, Texas, users. To incentivize more users to shift their departure times from 7:30 a.m. to 8:30 a.m. and from 4:15 p.m. to 5:15 p.m. the service decreased the number of points earned for travel during rush hour both in the morning and evening peak hours to just 10 points. The points available during the shoulders of rush hour spiked to 100 in an effort to lure drivers to travel at that time. A study (Metropia, 2018) of user behaviors during this experimental period indicated:

  • A 13 percent overall decrease in trips taken during the morning rush hour.
  • A 7 percent rise in trips taken during the subsequent, less-congested hour of the morning commute period, reflecting a shift in departure times.
  • A similar reduction in trips during the evening peak period, with trips increasing before and after peak congestion but dropping below previous levels during rush hour itself.

Similarly, the "Austin Don't Rush" mobility challenge, issued by the mayor of Austin, Texas, on May 11, 2016, urged commuters to consider the simplicity of avoiding peak traffic times and carpooling when their schedule permitted. As with the El Paso experiment, Metropia lowered peak period rewards points and doubled points for travel on the shoulders of the morning and afternoon peak periods. Points for using Metropia's DUO social carpooling feature were also doubled throughout the day. Observed improvements (Metropia, 2018) include the following:

  • The Austin areas achieved a systemwide 4 percent drop in a common travel time reliability performance measure, the travel time index (TTI) for morning peak period and a 3 percent drop in the TTI for the afternoon peak period. A drop in TTI indicates an increase in travel time reliability.
  • Those who shifted their morning commute experienced a 10 percent reduction in their travel times, and those who shifted their afternoon commutes realized a 6 percent reduction in travel time.

The Metropia system, which is funded through agreements with local governments and regional planning authorities, integrates an active demand management and data analytics platform to support the application. It has a back-end server system that calculates real-time traffic information from multiple sources and predicts traffic conditions, calculates routes, and manages all of the subsystem processes to support the front-end services. This backend analytics capability provides travel and traffic information to both commuters and cities during normal operations, pre-planned special events, or unexpected, extraordinary circumstances.

Metropia has active deployments in Austin, Texas; Tucson, Arizona; a binational deployment covering the El Paso–Juárez-Las Cruces region; and an upcoming deployment in Taiwan. Metropia's platform also supports the Regional Transportation Authority of Pima County's Mobility-on-Demand Sandbox grant from the from Federal Transit Authority and the Bay Area Rapid Transit's Perks 2 incentive program. Future deployments are currently planned for a number of cities in the United States.

For more information, see:

Commuter Connections Flextime Rewards Program

  • Online resources.
  • Point-based rewards.
  • Carpool matching.

In the Washington, D.C. region, one program has been developed that targets a specific set of corridors particularly prone to bottlenecks during the morning and afternoon peak congestion periods. The Commuter Connections Flextime Rewards Program (Flextime Rewards), launched in December 2017 and currently running as a pilot, was developed by Commuter Connections, a regional network of transportation organizations coordinated by the Metropolitan Washington Council of Governments, in conjunction with the University of Maryland. The program is sponsored through an award from the U.S. Department of Energy's Advanced Research Projects Agency-Energy (ARPA-E), and is jointly funded by the District of Columbia, Maryland, and Virginia Departments of Transportation, as well as through grants from the U.S. Department of Transportation.

Flextime Rewards combines Commuter Connections' software with the analytics behind the University of Maryland's National Transportation Center incenTrip program to provide personalized, real-time traveler information that can calculate the estimated time of arrival. This data can be calculated up to 24 hours in advance and recalculates as traffic conditions change or an incident is detected. Commuter Connections uses this data to alert registered Flextime Rewards users that significant congestion has been identified along their route and will give alternative departure times that could improve the commuter's trip.

The Flextime Rewards program is open to commuters who are able and willing to commute during off-peak hours to avoid congestion along the four selected corridors targeted by the program. At companies that allow their employees to utilize a flexible schedule, the Flextime Rewards program offers the option to adjust the times they arrive at and depart from work, reducing the overall number of people attempting to travel during the morning and afternoon peak periods. Currently, the pilot program is available to employees working in the Washington, D.C. air quality non-attainment region.

The image is of a man at his desk looking at his cellphone.
Source Ridofranc, iStock/Getty Images Plus.

To be eligible to participate and receive rewards, users must commute at least 2 days per week on one of the four eligible corridors. They must download the Commuter Connections app, which is the mechanism by which users are notified of higher-than-average congestion along the eligible corridor which they commute. In response, the user must adjust his/her normal commute trip and record the action using Commuter Connections' software to be eligible for a reward. A valid response by a user to a Flextime notification results in an entry into a lottery for a prize drawing to be held at the end of the month.

During the month, each time the user responds to a notification by altering their commute, they are entered into the drawing. A $100 raffle drawing is held each month. The winner receives either direct payment, gift certificates/ cards, or debit cards. If no notifications are sent throughout a month, no prize drawing occurs.

The Commuter Connections program managers chose to conduct a pilot or "soft launch" of the Flextime Rewards program so that they could analyze the initial results, get feedback from participants, make the program more robust, and then use marketing funds to increase participation. During the 2019 fiscal year, an upgrade to the system will include geolocation services to automatically detect whether or not a user is delaying their trip. During the pilot period, the program managers are seeking to gather feedback on ease of use and understandability. One of the early lessons learned is that the program needs to make it easy for participants to log their trips and to continue to obtain feedback from channels such as the system's call center so that program managers can increase the effectiveness of the program.

For more information, see Commuter Connections Flextime Rewards Program webpage at

Instant (Infosys-Stanford Traffic) Program: A "Nudge System" in Bangalore, India

  • Shuttle service.
  • Weekly drawings.
  • Point-based rewards.

The INSTANT experiment, conducted from 2008-2009 in Bangalore, India, served as an early exploration of the incentives concept. INSTANT rewarded commuters who chose to travel at less congested periods as a means of alleviating peak period congestion. The philosophy behind this experiment contrasts with strategies related to "self-routing," in which travelers are encouraged to simply use different routes at the same time that they normally begin their trip rather than varying their trip time.

Gridlock had become a concern among executives at Infosys, a large digital services and consulting firm that employs around 20,000 people in Bangalore, India. Like many booming Indian information technology (IT) companies, the Bangalore campus of the firm is located about 9 miles south of the city center. The IT boom in Bangalore resulted in a population explosion from 4.13 million in 1991 to nearly double that in 2007, which in turn caused the city to sprawl over nearly 750 square miles. (Merugu et al., 2009) Every morning, nearly three quarters of the Infosys employees commute through the congested city to reach the company's offices. Of the nearly 15,000 Infosys commuters, around 9,000 commuted by buses chartered by Infosys. The company was able to maintain extensive and detailed data on these commutes, such as commuting times and bus occupancy levels. A study of this data revealed that commuters who left for work after 7:30 a.m. had commute times that were about 1.5–2 times longer than those who left before 7:30 a.m.

One solution for potentially beating this congestion nightmare was a "nudge engine" that would encourage Infosys employees to come in early and beat the morning rush. A nudge engine is a program that uses mobile, cloud, and social networking technologies to sense individual behaviors; for example, the number of times employees swipe their identification badges at work. The tool then "nudges" these individuals to change those behaviors through the use of friendly competitions and incentives.

Congested bus stand in Bangalore city.
Source: shylendrahoonde, iStock, Getty Images Plus

The primary incentive for the INSTANT experiment, developed by Professor Balaji Prabhakar of Stanford University's Electrical Engineering and Computer Science Department, was an algorithm-driven lottery. The algorithm had three components: credit allocation, weekly reward drawings, and credit deduction.

For the experiment, commuters using Infosys buses earned credits based on their time of arrival, with those arriving earliest receiving the most credits, and those arriving later earning fewer credits until a certain time after which they accrue no credits. The more credits a commuter earned, the higher the amount of prize money they could win and the greater the chance that they could win a prize at all. At the end of each week, the algorithm divided the commuters into different levels based on the quantity of credits they had earned; those who had fewer credits were in lower levels, and the drawings were for lower amounts. Those who had higher numbers of credits were in higher levels with the potential for larger winnings. Rewards varied in value from the equivalent of about $10 for those who had accrued fewer credits to about $240 for those who had accrued a greater number of credits. This potential to win a greater reward was designed to incentivize commuters to accrue more credits each week.

After each drawing, the algorithm deducted credits on a sliding scale for both the winners and non-winners to ensure that commuters would continue to arrive early to rebuild their credit balances. This was also to ensure that winners would have to build up their credit balance over a few weeks before being able to win again, thus giving others a chance to win larger prizes.

The INSTANT experiment was considered to be very successful. During the course of the experiment, the pickup times for about 60 buses had to be advanced between 15 and 30 minutes as a result of commuter demand for earlier arrival of their buses, and several buses were shifted from arrival after 9:30 a.m. to before 8:30 a.m. At the conclusion of the 6-month experimental period, the number of Infosys commuters arriving before 8:30 a.m. had doubled, and the average morning commute time per person had dropped to 54 minutes from around 71 minutes before the scheme was launched. This resulted in a net savings of about 2600 person-hours per day. (Merugu et al., 2009)

For more information, see "An Incentive Mechanism for Decongesting the Roads: A Pilot Program in Bangalore" by D. Merugu, B. S. Prabhakar, N. S. Rama at:

Stanford University's Congestion and Parking Relief Incentives (CAPRI) Study

  • Point-based rewards.
  • Raffle rewards.
  • Online resources.

One of the largest employers in the San Francisco Bay Area, Stanford University, signed a General Use Permit with the County of Santa Clara that requires the university to manage its transportation impacts under a "no net new commute trips" standard: The amount of traffic during peak hours must not increase by more than 1 percent during the morning and afternoon peak hours (based on traffic count data from 2000). By 2012, while existing measures had been effective in reducing the total number of commuters who drive alone, they did not directly address peak-hour commuters, whose numbers were increasing.

In an effort to address this imbalance, the university, with a $3 million grant from the Federal Highway Administration Value Pricing Pilot Program, launched the Congestion and Parking Relief Incentives (CAPRI) program, which ran from April 2012 through September 2014. Its goal was focused on shifting driver commutes away from peak hours, but was expanded in 2013 to incentivize walking and bicycling commutes. The approach behind the project is based on the understanding that "congestion is a 10 percent phenomenon." (Abadi et al.) In other words, a small reduction in demand can lead to a significant drop in congestion. By targeting peak period commutes, a corresponding decrease in peak period congestion around the university could be achieved.

Chutes and ladders type game interface with a checkered game board, points balance, and colored buttons.
Figure 4. Screenshot. Chutes-and- ladders game for redeeming cash reward in CAPRI.
Source: Zhu et al 2014.

In addition, rather than penalizing undesirable behavior, such as increasing the cost of transit during peak periods, the CAPRI project approach was designed to incentivize decongestion by using "carrots" to influence driver behavior. This methodology leverages game theory, in which games with low stakes see players become more risk-seeking, resulting in greater user responsiveness achieved by paying out random "chunky" rewards rather than small, deterministic payments. CAPRI built on the incentives first proposed in the previously introduced INSTANT (Merugu et al., 2009) program, as well as Steptacular (Gomes et al., 2012) and Insinc (Pluntke and Prabhakar, 2013) programs in terms of both behavioral interventions as well as technological elements.

In April 2012, Stanford University parking permit holders who parked inside the "congestion cordon" were invited to participate in the program. Those who enrolled were given passive radio-frequency identification tags to place on their windshield. Entries and exits were tracked by sensing devices at 10 main access points on the Stanford campus during the 7–10 a.m. and 4–7 p.m. periods each weekday, with peak-hours defined as 8–9 a.m. and 5–6 p.m. For each automobile detected by the sensors during the off-peak shoulder hours (i.e., 7–8 a.m. and 9–10 a.m., and 4–5 p.m. and 6–7 p.m.), the participant was awarded 10 points.

Additionally, CAPRI assigned each participant a "boost day," or a day on which their off-peak trip earned them 30 points. Beginning in May 2013, the project was expanded to incentivize walkers and bicyclists by awarding them between 10 and 25 points, depending on the length of their commutes. Walking and biking activity was monitored using the "My Beats" smartphone app developed for this project.

Participants were incentivized by receiving points for commutes during off-peak periods and non-motorized commutes. These points could be redeemed in one of two ways:

  1. Deterministically, by trading 100 points for $1 (or a full week's worth of off-peak trip points).
  2. Randomly, by playing a "chutes-and-ladders" type game using their points on the CAPRI website.

The game gave cash rewards ranging from $1–$50. The follow-on study of the project found that 87.3 percent of the participants used the random rewards option, validating the theory behind the project. (Zhu et al., 2014) Notably, since participants were allowed to change the manner of redeeming rewards, 13.2 percent of the participants ended up switching from the deterministic option to the random option at some point during the program. (Zhu et al., 2014)

The interface for the "My Beats" smartphone app for walkers and bikers. The image notes that commutes are rewarded by duration, and that on boost days, participants can earn three times the credits for eligible trips.
Figure 5. Screenshot. Interface for "My Beats" smartphone app for this CAPRI.
Source: Abadi et al., n.d.

CAPRI also tapped into basic human traits such as the desire to improve one's social status, the desire to connect with friends, and the desire to feel understood to increase the popularity, engagement, and behavior shift among the participants. This was accomplished through a rewards system that encouraged the desired behaviors and made them fun:

  • Status system. Participants began at the bronze level and were able to advance through silver, gold, and platinum levels based on the number of off-peak shoulder hour trips they made on a weekly basis. At the silver, gold, and platinum levels, failure to make the number of off-peak shoulder hour trips required for that status level resulted in a degrading of the status by one level. Recognizing that status is only worth something if it is associated with a privilege, CAPRI gave participants with higher status higher odds of winning rewards in the game, and higher-valued rewards were only available at the higher status levels. For example, a $50 reward was only available at the Platinum level.
  • Friends. Leveraging the popularity of social media, CAPRI participants were allowed to invite friends who were eligible to participate in CAPRI to join the program as well as to connect with their friends on the CAPRI portal. Participants could see their friends' recent updates, including status upgrades, any cash awards won, etc. This feature provided a basis for social influence to spread.
  • Magic Box. Based on a participant's tracked preference for commuting off-peak, this incentive offered weekly personalized opportunities to gain additional points through a tab in a commuter's portal called "Magic Box." For example, filling out an optional survey might garner the user an additional 200 bonus points.
  • Trendjacking. Because Stanford University participates in numerous high-profile sporting events, CAPRI offered tickets to some of these events and used them to incentivize behavior shift or increase enrollment.


Over the 30-month study period, 4,057 Stanford affiliates completed the registration process; this includes 3,082 car commuters and 975 biking/walking commuters. These car commuters comprised about 30 percent of the 10,290 car commuters in Stanford who were ever eligible to participate. (Zhu et al., 2014)


A post-project study identified the following results of the CAPRI program:

  • CAPRI users avoided peak hours. For CAPRI participants, the peak-hour trip ratio is only 30.1 percent in the morning and 32.4 percent in the evening, which is a 21.2 percent and a 13.1 percent reduction, respectively, from the Stanford-wide traffic. (Zhu et al., 2014)
  • CAPRI users responded to incentives. The commute density for CAPRI participants peaks adjacent to (but just outside) the peak hours. Furthermore, CAPRI users preferred commuting during the hour before the peak hour as compared to the hour after the peak hour. (Zhu et al., 2014)
  • CAPRI rewards had a direct effect on participants' commute time. Results of the study show that participants will shift their commute time away from peak hours when receiving rewards in the recent past. (Zhu et al., 2014)
    • Early commuters who have friends winning rewards in the past week travel around 1.5 minutes earlier. Early commuters also advance their commutes by an additional minute on their boost days to ensure receiving bonus award points.
    • Late users who won rewards in the past week shift about 3 minutes later in morning and afternoon (non-peak) commutes.

Since the conclusion of the CAPRI program, Stanford University developed and is currently hosting a "Commute Club" through its Parking & Transportation Services program. This program offers a variety of incentives to the more than 10,000 members who commute to and from the university. Incentives include up to $300 a year in "Clean Air Cash" or carpool credit for not purchasing a long-term parking permit; free carpools and vanpools, along with reserved parking for carpool vehicles; Zipcar driving credit of up to $102 per year; free folding bicycle rental for 1 week along with subsidized purchase of folding bicycles; and the opportunity to win other prizes through regular drawings.2

For more information, see

Spitsmijden ("Peak Avoidance")

  • Monetary rewards.
  • Online resources.

In the Netherlands, several iterations of what have been called the "Spitsmijden" (or peak avoidance) experiments have been conducted since 2006 to examine the concept of rewarding drivers for driving during off-peak hours rather than at the height of typically congested periods. Three of the earliest projects are described below, although other peak avoidance projects have also been completed recently in the Netherlands. These projects included different types of rewards (e.g., monetary, lottery, gifts, free bikes, etc.), variance in the length of time during which rewards are given, and a reduction of the rewards after a period of time.


The goal of the first experiment in Zoetermeer in 2016 was simply to investigate the behavioral responses of travelers when incentivized by a potential reward. With only 340 participants, the experiment was not intended to solve congestion problems, but rather to validate the hypothesis that rewards could be used to alter driver behavior.

Graph illustrates the number of travelers who traveled before rewarding, when rewarded with 3 Euros (about $3.78) and when rewarded with 7 Euros (about $8.82). The data show that rewarded travelers were much more likely to travel before or after the peak period.Figure 6. Graph. Number of detected travelers from Zoetermeer during morning peak.
Source: Bliemer, Dicke-Ogenia, & Ettema, 2010.

Two different reward types were identified, and participants could choose their preferred reward. The first group opted for a monetary reward (232 participants), which offered €3 to €7 (about $3.78 to about $8.82 in 2006 U.S. dollars) per day for avoiding using a car during the morning peak between 7:30 and 9:30 a.m.. The second group opted to use a smartphone with global positioning system (GPS) that could also provide real-time traffic information (108 participants). These participants were rewarded by being able to keep the smartphone if they avoided the morning peak sufficiently throughout the 10 week period. During the study, participants could not change their route in order to earn a reward, so the only alternatives available were to use different departure times, different travel modes, or choosing not to travel.

Researchers (Bliemer et al., 2015) found that:

  • Among those earning a €3 (about $3.78 in 2006 dollars) reward, 46 percent fewer trips were made by car during the morning peak, 35 percent of all trips made by car were outside peak hours, and 10 percent shifted to other modes. Approximately 67 people, or about 20 percent of the participants, did not use a car during morning peak when offered a €3 reward.
  • Among those earning a €7 (about $8.82 in 2006 dollars) reward, 61 percent fewer trips were made by car during the morning peak, and 44 percent of all trips made by car were outside the peak hours, and 14 percent shifted to other modes. Approximately 88 people, or just over 25 percent of the participants, did not use a car during morning peak when offered a €7 reward. (Bliemer et al., 2015)

Following the positive results, the Dutch Ministry of Transport decided to execute two additional projects in 2007 and 2008 with the goal of influencing traffic conditions during the planned maintenance and renewal of two major bridges, Hollandse Brug east of Amsterdam and Moerdijk Brug south of Rotterdam.

Hollandse Brug (Hollandse Bridge)

For the Hollandse Brug construction zone the goal was to reduce the number of vehicles using the bridge by 1,000 to 1,500 during the 6 a.m. to 10 a.m. morning peak. In addition to the peak avoidance monetary reward, the Ministry of Transport also offered free public transportation and vanpools. Additional measures included alternate route guidance, including dynamic route information panels with travel times for alternative routes and for travel by ferry.

A total of 2,975 participants were recruited for the 12-month project period, with half of participants being recruited prior to the project, and half being recruited after the first 6 month period to increase participation. Participants were offered a reward of €4 (about $5.48 in 2007 U.S. dollars) for avoiding the morning peak period (6:00 a.m. to 10:00 a.m.) by car. Participants could earn an additional €2 (about $2.74 in 2007 U.S. dollars) if they did not travel by car on the bridge all day. Participants of the peak avoidance project traveled an average of 2.1 times per week on the bridge during the peak periods. During the rewarding period, this number decreased to 1.3 times per week, which is a 40 percent decrease in the number of trips. (Bliemer et al., 2015) Of this 40 percent:

  • Eighteen percent opted to travel outside peak hours.
  • Nine percent chose an alternate route.
  • Six percent chose an alternate mode of transportation.

In sum, the incentive program resulted in a reduction of approximately 1,250 car trips per week for the first half of the year; this is equivalent to a reduction of 250 cars per morning peak, which is 1.5 percent of the total flow. (Bliemer et al., 2015) During the second half of the year, this reduction rose to 425 cars per morning peak, about 2.6 percent of the total flow, but still significantly less than the 1,000 to 1,500 vehicle reduction goal.

Moerdijk Brug (Moerdijk Bridge)

The Moerdijk Brug project conducted from April to July of 2008 aimed to avoid significant increases in the congestion on southbound lanes due to road work during the evening peak period (3 p.m. to 7 p.m.). In total 2,703 people participated. For this experiment, participants could earn €4 per day (about $5.88 in 2008 U.S. dollars) or a maximum of €20 per week (about $29.40 in 2008 U.S. dollars). Out of all participants, 66 percent of the travelers indicated that they changed their behavior. Of the 1,784 participants who did change their behavior (Bliemer et al., 2015):

  • Twenty-eight percent chose alternate routes.
  • Fifteen percent changed their departure time.
  • Six percent worked from home more often.
  • Five percent carpooled or used an alternate mode.

Among those who changed their departure time, 37 percent chose to depart later and 19 percent chose to depart earlier. On average, people departed 95 minutes earlier or 87 minutes later.

Follow-on Experiments

Since 2009, several additional experiments in the same vein—with the same rules of participation regarding eligibility, travel modes, and reimbursement—have been conducted, with the most recent being the Spitsmijden A2 Nederweert-Eindhoven project in 2017. With more than 2,000 participants (SmartWays, 2018), the Spitzmijden A2 initiative was responsible for a total of 145,571 rush hour avoidances and 746 fewer cars on the road during peak traffic each day, significantly more than the goal of 680 fewer vehicles per day. (Rijkswaterstaat, 2018) Three-quarters of participants indicated that they got used to their new way of traveling and will continue to choose to avoid driving during peak traffic. Together, they continue to realize 650 rush hour avoidances per day.

The financing for the peak avoidance research projects in the Netherlands has changed since the early Spitsmijden projects were conducted. Businesses are typically now providing the rewards instead of the government directly financing the projects.

A 2017 report commissioned by the Municipality of Amsterdam (Stemerding and Mateboer, 2017) points to the general success of these efforts to date and suggests the city continue to identify opportunities to incentivize peak period travel reductions by focusing on:

  • Changing the travel behavior of students.
  • Working with large businesses to institute flexible work hours, working from home, and changes in shift start and end times (for hourly workers).
  • Targeting employees among smaller employers within the city center for future incentive initiatives.

For more information, see Rewarding for Avoiding the Peak Period: A Synthesis of Three Studies in the Netherlands by Bliemer, Michiel, Matthijs Dicke-Ogenia, and Dick Ettema. Available at: PEAK_PERIOD_A_SYNTHESIS_OF_THREE_STUDIES_IN_THE_NETHERLANDS.

Bay Area Rapid Transit (BART) Perks

  • Point-based rewards.
  • Raffle rewards.

The San Francisco Bay Area Rapid Transit (BART) system has seen a rapid increase in ridership thanks to population and job growth in the area. From 2004 to 2016 alone, the system saw a 40 percent increase in overall ridership and a 75 percent increase in ridership in the Transbay corridor connecting the San Francisco financial district to the East Bay cities through the underwater Transbay Tube. BART's maximum train car load target is 117 riders, but train car loads of about 140 riders are seen during rush hour.

In 2016, BART and the San Francisco County Transportation Authority (SFCTA) rolled out the BART Perks program, aimed at nudging a small percentage of peak hour, peak direction Transbay Tube riders to hours before or after the peak rush.

Image from a promotional flyer encouraging riders to sign up for BART Perks.
Figure 7. Screenshot. Promotional flyer for BART Perks.
Source: BART, n.d.

BART Perks was a 6-month test program that provided incentives to riders to shift travel time from the peak morning rush hour to the shoulder hours occurring before and after the peak— specifically, riders were encouraged to travel during 6:30 a.m. – 7:30 a.m. and 8:30 – 9:30 a.m.. The program had nearly 18,000 participants out of the approximately 26,000 riders who travel during peak hour. (SFCTA et al., 2018) Of those surveyed who did not participate, lack of schedule flexibility was the most cited barrier to participation.

Participants enrolled in the program through a mobile-friendly website and provided their Clipper Card ID number, allowing BART and SFCTA to provide awards based on frequency, timing, and length of trips, as well as to observe trends and effectiveness of the program. BART and SFCTA monitored rider trends in the 6 months leading up to the program period and 4 months after the program ended.

During the BART Perks program period, Transbay peak hour travel demand decreased 10.9 percent, while the overall BART system travel demand decreased by 9.6 percent. (SFCTA et al., 2018) Participants were more likely to travel during one of the bonus hours if it was close to their typical departure time and their schedule wouldn't be altered greatly. Transbay travelers were less likely to shift into the bonus hours, perhaps due to many of those participants already traveling outside the peak hour, and many long distance commuters shifted more than other commuters, typically opting for the hour before peak rush. Two-thirds of participants shifted their travel time to bonus hours at least twice per week.

Image of a subway stop in the bay area. Source: Joe Potato, iStock Editorial/Getty Images Plus

Shifting a trip to the bonus hours rewarded participants with points that they could use on the BART Perks reward generator. The points were played in a "chutes and ladders" type game, where participants could win from zero to $100. On average, participants earned about $2 per month.

Upon completion of the 6-month program period, BART saw continued effects on travel behavior: 35 percent of peak hour trips cut during the program continued to happen outside of the peak hour in the four months following the program. Program participants were satisfied overall with BART Perks, with many participants stating their desire for higher levels of rewards, better rewards, or more opportunities to earn rewards, perhaps during a longer bonus period.

This program demonstrated a successful effort to shift departure times from peak hours. In the future, the Perks program would consider working more closely with employers to learn about barriers to shifting employee start times, therefore allowing later departure times from transit riders. While many participants felt comfortable arriving at work earlier, many did not feel that arriving at work late was an option. The BART Perks program model could also serve as a tool to achieve multiple objectives for a transit agency in the future. While this particular program was aimed at shifting morning rush travel demand, a Perks-type program could be used to reward frequent riders or encourage travel in lower ridership times of the day or week.

For more information, see Lessons from Perks: Evaluation Findings from the BART Perks Test Program at:

You may need the Adobe® Reader® to view the PDFs on this page.

Office of Operations