1. The Customer Experience
Nigel H.M. Wilson
Professor of Civil & Environmental Engineering
MIT
email: nhmw@mit.edu
1
BRT Workshop: Experiences and Challenges
September 2013

2. Outline
• The changing environment and customer expectations
• Customer Information Strategies
• Recent Research
• OD Matrix Estimation
• Measuring Service Reliability
• Role for Customer Surveys
• Customer Classification
• Summary and Prospects
2
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

3. The Changing Environment and
Customer Expectations
• Many customers expect a personal relationship with
service providers, e.g., Amazon
• Information technology advances provide raised
expectations and new opportunities
• Wireless communications raise expectations for good
real-time information
• Rising incomes result in more choice riders and fewer
captive riders
• Finance for capital and operations remains a challenge
3
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

4. Evolution of Customer Information
• Operator view Customer view
• Static Dynamic
• Pre-trip and at stop/station En route
• Generic customer Specific customer
• Information "pull" Information "push"
4
4
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

5. Enabling Technologies
• AVL provides current vehicle locations
• Automated scheduling systems make service plan accessible
• Google Transit standard formats provide universal trip
planning
• GPS- and WIFI cell phones provide current customer location
• AFC provides database on individual trip-making
• Wireless communication/Internet apps
5
5
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

6. State of Research/Knowledge in CI
• Pre-trip journey planner systems widely deployed but with
limited functionality in terms of recognizing individual
preferences (e.g., Google Transit)
• Next vehicle arrival times at stops/stations well developed
and increasingly widely deployed
• both often strongly reliant on veracity of service schedules
• ineffective in dealing with disrupted service
• Real-time mobile phone information
• open data
• many new apps, some great, some not so great
• Google's entry may be game-changer in the long run
6
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

7. Example of Well-Designed Mobile Web
App: NextBus.com/webkit
• First finds your location
• Lists all services and nearest stops for each within 1/4
mile radius
• Scrolls to show next two vehicles for each service in each
direction
• www.nextbus.com/webkit
7
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

8. Emerging Possibilities
• Exception-based CI based on stated and revealed
individual preferences, typical individual trip-making, and
current AVL data
• Integration of AFC and CI functions through payment-
capable cell phones
• Can CI actually attract more customers?
• multi-modal trip planner/navigation systems
8
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

9. Medium-term Vision
Transit becomes a virtual presence on mobile devices:
• Transit is information-intensive mobility service
• Cell phone is a mobile information device, a perfect match
• People (will) have their lives on their smart phones
• Single device for payment and information
• “Station in your pocket”: no need to restrict countdown clocks, status
updates, trip guides to stations or fixed devices
• Lifestyle services: guaranteed connections, in-station navigation, bus
stop finder, transit validation, rendezvous, …
9
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

10. Recent Research
• OD Matrix Estimation
• Measuring Service Reliability
• Roles for Customer Surveys
• Customer Classification
10
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

11. OD Matrix Estimation
Objective:
• Estimate passenger OD matrix at the network level using AFC
and AVL data
• Multimodal passenger flows
• AFC characteristics
• Open (entry fare control only)
• Closed (entry+exit fare control)
• Hybrid
11
Source:
"Intermodal Passenger Flows on London’s Public Transport Network: Automated Inference of Full Passenger Journeys Using Fare-
Transaction and Vehicle-Location Data. Jason Gordon, MST Thesis, MIT (September 2012).
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

12. Journey 1
1. Enter East Croydon NR station, 7:46
2 & 3. Out-of-station interchange to Central Line at
Shepherds Bush, 8:30
4. Exit LU at White City, 8:35
5. Board 72 bus at Westway, 8:36
6. Alight 72 bus at Hammersmith Hospital, 8:42
Journey 2
7. Board bus 7 at Hammersmith Hospital, 16:17
8. Alight bus 7 at Latymer Upper School, 16:19
9. Board bus 220 at Cavell House, 16:21
10. Alight bus 220 at White City Station, 16:24
11. Enter LU at Wood Lane, 16:25
12 & 13. Out-of-station interchange from Circle or
Hammersmith & City to District or Piccadilly,
16:40
14. Exit LU at Parsons Green, 16:56
12

13. Trip-Chaining Method for OD Inference
• Infer start and end of each trip segment for individual
AFC cards
• Link trip segments into complete (one-way) journeys
• Integrate individual journeys to form seed OD matrix by
time period
• Expand to full OD matrix using available control totals
• station entries and/or exits for rail
• passenger entries and/or exits by stop, trip, or period for bus
13
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

14. Summary Information on London
Application
• Oyster fare transactions/day:
• Rail (Underground, Overground, National Rail): 6 million (entry & exit)
• Bus: 6 million (entry only)
• For bus:
• Origin inference rate: 96%
• Destination inference rate: 77%
• For full public transport network:
• 76% of all fare transactions are included in the seed matrix
• Computationally feasible (30 mins on Intel PC for full London
OD Matrix for entire day, including both seed matrix and
scaling)
14
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

15. Reliability Metrics
• Goal: characterize transit service reliability from
passenger's perspective
• Application: London rail services
• entry and exit fare transactions
• train tracking data
• Application: London bus services
• typically high frequency
• entry fare transactions only
15
Sources:
"Service Reliability Measurement Framework using Smart Card Data: Application to the London Underground." David Uniman, MST Thesis, MIT (2009)
"Automatic Data for Applied Railway Management: Passenger Demand, Service Quality Measurement, and Tactical Planning on the London
Overground Network." Michael Frumin, MST Thesis, MIT (2010)
"Applications of Automatic Vehicle Location Systems Towards Improving Service Reliability and Operations Planning in London." Joseph Ehrlich, MST
Thesis, MIT (2010)
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

16. Excess Journey Time (EJT)
16

17. Example: Reliability Metrics - Rail
High Frequency Service
• use tap-in and tap-out times to measure actual station-station journey
times
• characterize journey time distribution measures such as Reliability
Buffer Time, RBT (at O-D level):
17
RBT = Additional time a passenger must budget to arrive on time for most of their
trips (≈ 95% of the time)
50th perc.
% of Journeys
Travel Time
95th perc.
RBT
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

18. Line Level ERBT
18
Victoria Line, AM Peak, 2007
TravelTime(min)
February November
NB
(5.74)
SB
(10.74)
NB
(6.54)
SB
(7.38)
12.00
10.00
8.00
6.00
4.00
2.00
0.00
Excess RBT
Baseline RBT
4.18 5.524.185.52
1.56
5.22
2.36
1.86
Period-Direction
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

19. Reliability Metrics: Bus
Challenge to measure passenger journey time because:
• no tap-off, just tap-on
• tap-on occurs after wait at stop, but wait is an important part of
journey time
Strategy:
• trip-chaining to infer destination for all possible boardings
• AVL to estimate:
• average passenger wait time (based on assumed passenger arrival
process)
• actual in-vehicle time
19
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

20. Role for Customer Surveys
• Agencies/operators have traditionally relied on customer
surveys for data on:
• multi-modal trip-making
• demographics
• attitudes and perceptions
• Surveys provide the base for travel demand modeling
• Surveys will remain important, but can they be more cost-
effective and reliable?
• Research in London compared Oyster records with LTDS
(Household survey) responses for approximately 4,000
individuals in 2011-2012
20
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

21. Concerns with Household Surveys
• Expensive and usually conducted infrequently
• Public Transport trips may not be fully captured
• Gathering representative data is becoming more difficult
• Large journey sample over multiple days is desired for
public transport planning purposes
• Relies on respondent’s memory
21
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

22. Summary of Matching Specific LTDS and
Oyster (OR) Journey Stages
• 46% of LTDS stages had matching OR Stages
• 51% of OR Stages had matching LTDS Stages
Source: "Utilizing Automatically Collected Smart Card Data to Enhance Travel Demand Surveys." Laura Riegel,
MST Thesis, MIT (June 2013)
22
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

23. LTDS vs. Oyster Stages for People with
Weekday Travel Days
23
Avg. OR on
All Captured
Weekdays
Avg. OR on
All Possible
Weekdays
LTDS on
Travel Day
OR on
Travel Day
LTDSorOysterStages
20
15
10
5
0

24. Variability of PT Travel
• The surveyed travel day is not representative of all days:
• the single day overestimates typical PT use overall
• underestimates the intensity of PT use on the days it is used
• People who used PT in the survey used it only about half the
time (over a four week period), leading to an overestimate of
typical PT use.
• The reported frequency of use is much higher than actual PT
use and may not be the most accurate way to scale up
reported travel day responses
24
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

25. Recommendations
• It is difficult to combine survey and AFC data after the survey
• AFC records could be used during the interview with a card
reader and tablet to enhance the survey process
• AFC records over two weeks (or other time period) could be
used to supplement questions regarding PT frequency of use
• A customer panel could be created to understand variability in
travel behavior over time
• OD matrix estimation and trip chaining could be used to
calculate exact trip attributes (start time, duration speeds)
25
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

26. Online Customer Survey Strategy
• Aim was to demonstrate the potential of online surveys
to gather detailed and representative information from
public transport customers identified through Oyster
records
• Application was to understand customer behavior in
multi-route corridors
Source: "Bus Use Behavior in Multi-Route Corridors." Cecilia Viggiano, MST Thesis, MIT (2013)
26
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

27. Online Customer Survey Strategy
• Survey e-mailed to about 52,000 registered Oyster Card
holders who had used the routes of interest in the prior
two weeks
• Incentive was an iPad awarded to a random respondent
• Response rate of 18% yielded over 9,400 responses
Source: "Bus Use Behavior in Multi-Route Corridors." Cecilia Viggiano, MST Thesis, MIT (2013)
27
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

28. Customer Classification Research
Aims:
• identify homogeneous groups of passengers through analysis
of Oyster records
• investigate the representativeness of registered Oyster Card
holders
• understand the attrition over time of individual Oyster cards
28
Source: "Classification of London’s Public Transport Users Using Smart Card Data." Meisy Ortega,
MST Thesis, MIT (2013)
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

29. Methodology
• Identify Oyster Card clusters based on a number of explanatory
variables:
 Temporal characteristics
• Travel Frequency  No. travel days and trips per day
• Journey Start Time  First and last journeys of the day
 Spatial characteristics
• Origin Frequency  No. of different first and last origins of the day
• Travel Distance  Maximum and minimum distance traveled
 Activity Pattern characteristics
• Activity Duration  Main and shortest activity of the day
 Mode Choices  No. of bus-only and rail-only days
 Sociodemographic  Travelcard or Special Discount
(Freedom, Student/Child, Staff)
• Clustering process based on identifying homogenous groups of travelers
29
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

30. Travel Frequency
30
• London Oyster data for 1-7 October, 2012
• Number of days a card was used over a week
• Many cards are used only one day per week
• Bimodal distribution:
• 1 day a week
• 5 days a week
• Similar usage patterns in Santiago, Chile and
Kochi City, Japan
Source: http://www.coordinaciontransantiago.cl
Number of Days
%ofOysterCards
Pay as You Go Period Pass
24
22
20
18
16
14
12
10
8
6
4
2
0
1 2 3 4 5 6 7
Number of Days
Number of Days
25
20
15
10
5
0
1 2 3 4 5 6 7
%ofOysterCards
Santiago, June 2010
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

31. Activity Patterns
• London weekday activity direction
• Main activity: Activity of the day
with the longest duration.
• Two peaks: 1- 3 and 7-9 hours.
• Shortest activity: Activity of the day
with the shortest duration (If user
has only one activity, main and
shortest activity are the same)
• Clear peak at one hour.
31
Activity: Refers to actions users perform between journeys.
Activity duration: Time lapsed between a tap-out and the subsequent tap-in.
%ofOysterCardsbeingobservedduringweekdays
12
10
8
6
4
2
0
Activity Duration (hours)
Main Activity Shortest Activity
0.5-1.0
1.0-1.5
1.5-2.0
2.0-2.5
2.5-3.0
3.0-3.5
3.5-4.0
4.0-4.5
4.5-5.0
5.0-5.5
5.5-6.0
6.0-6.5
6.5-7.0
7.0-7.5
7.5-8.0
8.0-8.5
8.5-9.0
9.0-9.5
9.5-10.0
10.0-10.5
10.5-11.0
11.0-11.5
11.5-12.0
12.0-12.5
12.5-13.0
13.0-13.5
13.5-14.0
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

32. Passenger Groups
32
Cluster Frequency Start Times Mode
Type of
Card
RegularUsers
1. Everyday regular
users
7 days
w: 8:30 – 19:30
we: 9:30 – 18:15
Mixed Travelcard
2. All week regular
users
6 days
w:10:30 – 16:30
we: 13:30 – 17:00
Mixed
Mix PAYG/
Travelcard
3. Weekday rail
regular users
5 weekdays 7:30 – 15:30 Rail Travelcard
4. Weekday bus
regular users
5 weekdays 9:30 – 16:00 Bus
Child bus
pass
OccasionalUsers
5. All week occasional
users
3 days 15:30 – 18:00 Mixed PAYG
6. Weekday bus
occasional users
2 weekdays 13:00 – 15:30 Bus PAYG
7. Weekend occasional
users
2 weekend days 17:30 – 20:30 Mixed PAYG
8. Weekday rail
occasional users
1 weekday 13:00 - 14:00 Rail PAYG
Exclusive
Commuters
Non-
Exclusive
Commuters
Non-
Commuter
Residents
Leisure
Travelers
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

33. Visitor Travel Patterns Cluster
RegularUsers
1. Everyday regular
users
Non-Exclusive
Commuters
2. All week regular
users
3. Weekday rail
regular users
Exclusive
Commuters
4. Weekday bus
regular users
OccasionalUsers
5. All week
occasional users
Non-Commuter
Residents
6. Weekday bus
occasional users
7. Weekend
occasional users
Leisure
Travelers
8. Weekday rail
occasional users
• Visitor Oyster Card analysis (April 2012)
• High number of short-to-medium duration activities
• Trips start during off-peak periods
• Activities focused in Central London
• Long walking trips between public transport trips
• High number of rail trips
• Leisure traveler groups  similar behavior to
Visitor Oyster Card holders
• Possible identification of visitors (not holding VOC)
%ofEachSample
Activity Duration (hours)
0.5-1.0
1.0-1.5
1.5-2.0
2.0-2.5
2.5-3.0
3.0-3.5
3.5-4.0
4.0-4.5
4.5-5.0
5.0-5.5
5.5-6.0
6.0-6.5
6.5-7.0
7.0-7.5
7.5-8.0
8.0-8.5
8.5-9.0
9.0-9.5
9.5-10.0
10.0-10.5
10.5-11.0
11.0-11.5
11.5-12.0
12.0-12.5
12.5-13.0
13.0-13.5
13.5-14.0
12
10
8
6
4
2
0
Visitor Non-Visitor
04:00-04:59
05:00-05:59
06:00-06:59
07:00-07:59
08:00-08:59
09:00-09:59
10:00-10:59
11:00-11:59
12:00-12:59
13:00-13:59
14:00-14:59
15:00-15:59
16:00-16:59
17:00-17:59
18:00-18:59
19:00-19:59
20:00-20:59
21:00-21:59
22:00-22:59
23:00-23:59
Start Time
%ofEachSample
12
11
10
9
8
7
6
5
4
3
2
1
0
Visitor Non-Visitor
Visitor Oyster Card Cluster Distribution
Type of Cluster
Regular
Cluster 1
Cluster 5
Cluster 2
Cluster 6
Cluster 3
Cluster 7
Cluster 4
Cluster 8
%ofVisitorOysterCards
80
70
60
50
40
30
20
10
0 Occasional

34. Registered Users
• Registered users are distributed
differently among clusters
• Regular user clusters have higher
percentage of registered cards
• Representative characteristics in
each cluster, but more similarity
with regular users behavior
Cluster1
Cluster2
Cluster3
Cluster4
Cluster5
Cluster6
Cluster7
Cluster8
%ofeachclusterofOysterCards
70
60
50
40
30
20
10
0
Cluster 1 First and Last Journey Start Times
Start Time
Relative%ofOysterCards
14
12
10
8
6
4
2
0
Registered Total
04:00-04:59
05:00-05:59
06:00-06:59
07:00-07:59
08:00-08:59
09:00-09:59
10:00-10:59
11:00-11:59
12:00-12:59
13:00-13:59
14:00-14:59
15:00-15:59
16:00-16:59
17:00-17:59
18:00-18:59
19:00-19:59
20:00-20:59
21:00-21:59
22:00-22:59
23:00-23:59
Cluster 2 Activity Duration
Activity Duration (hours)
Relative%ofOysterCards
8
7
6
5
4
3
2
1
0
0.5-1.0
1.0-1.5
1.5-2.0
2.0-2.5
2.5-3.0
3.0-3.5
3.5-4.0
4.0-4.5
4.5-5.0
5.0-5.5
5.5-6.0
6.0-6.5
6.5-7.0
7.0-7.5
7.5-8.0
8.0-8.5
8.5-9.0
9.0-9.5
9.5-10.0
10.0-10.5
10.5-11.0
11.0-11.5
11.5-12.0
12.0-12.5
12.5-13.0
13.0-13.5
13.5-14.0
Registered Total
Cluster
RegularUsers
1. Everyday
regular users
Non-Exclusive
Commuters
2. All week
regular users
3. Weekday rail
regular users
Exclusive
Commuters
4. Weekday bus
regular users
OccasionalUsers
5. All week
occasional
users
Non-Commuter
Residents
6. Weekday bus
occasional
users
7. Weekend
occasional
users
LeisureTravelers
8. Weekday rail
occasional
users

35. Oyster Card Attrition Cluster
RegularUsers
1. Everyday regular
users
Non-Exclusive
Commuters
2. All week regular
users
3. Weekday rail
regular users
Exclusive
Commuters
4. Weekday bus
regular users
OccasionalUsers
5. All week
occasional users
Non-Commuter
Residents
6. Weekday bus
occasional users
7. Weekend
occasional users
Leisure
Travelers
8. Weekday rail
occasional users
• Oyster Card attrition
estimated as a function of
active cards in each month
• 2010/2011 Oyster Card data
analysis  active cards
decreased logarithmically.
• Similar attrition rate for
2011/2012 period
• Occasional users have higher
attrition
y = -0.1576 ln(x) + 0.8632
R2 = 0.9685
Apr-10 Jun-10 Sep-10 Oct-10 --- Log Regression
%ofActiveOysterCards
100
90
80
70
60
50
40
30
20
10
0
Number of months after observed week
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Cluster 1
Cluster 2
Cluster 3
Cluster 4
Cluster 5
Cluster 6
Cluster 7
Cluster 8
Total Sample
Log Regression- - - -
100
90
80
70
60
50
40
30
20
10
0
%ofActiveOysterCards
Months
Oct-2011
Nov-2011
Dec-2011
Jan-2012
Feb-2012
Mar-2012
Apr-2012
May-2012
Jun-2012
Jul-2012
Aug-2012
Sep-2012
Oct-2012

36. Findings
• 8 homogenous groups of users with distinctive travel behavior were
found  logical aggregation in 4 groups:
• Exclusive commuters, non exclusive commuters, leisure travelers, and
non-commuter residents
• Visitors similar to occasional user clusters  business and leisure
• Different % of registered card users per cluster. Registered users
travel behavior more similar to regular users behavior.
• Attrition rates decrease over time. Large drop in number of active
cards explained by occasional users behavior
• First step in understanding user attrition
36
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

37. Summary
• Complete Journey OD Estimation practical with ADCS
• foundation for many analyses related to customer experience
• Realistic to assess service reliability for individuals and journeys
• most critical aspect of customer experience
• Home interview surveys can be enhanced with AFC records
• Targeted on-line surveys an efficient alternative to other survey
methods
• Customer classification is critical in understanding the customer
experience
37
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013

38. Appendix
MIT theses used in this presentation
"Service Reliability Measurement Framework using Smart Card Data: Application to the London
Underground." David Uniman, MST Thesis (2009)
"Automatic Data for Applied Railway Management: Passenger Demand, Service Quality
Measurement, and Tactical Planning on the London Overground Network." Michael Frumin, MST
Thesis (2010)
"Applications of Automatic Vehicle Location Systems Towards Improving Service Reliability and
Operations Planning in London." Joseph Ehrlich, MST Thesis (2010)
"Intermodal Passenger Flows on London’s Public Transport Network: Automated Inference of Full
Passenger Journeys Using Fare-Transaction and Vehicle-Location Data." Jason Gordon, MST
Thesis, MIT (2013).
"Utilizing Automatically Collected Smart Card Data to Enhance Travel Demand Surveys." Laura
Riegel, MST Thesis (2013)
"Bus Use Behavior in Multi-Route Corridors." Cecilia Viggiano, MST Thesis (2013)
"Classification of London’s Public Transport Users Using Smart Card Data." Meisy Ortega, MST
Thesis (2013)
"Quantifying the Current and Future Impacts of the MBTA Corporate Pass Program." Dianne
Kamfonik, MST Thesis (2013)
38
Nigel Wilson, BRT Workshop:
Experiences and Challenges
September 2013