38
The Customer Experience Nigel H.M. Wilson Professor of Civil & Environmental Engineering MIT email: [email protected] 1

BRT Workshop - The Customer Experience

  • Upload
    embarq

  • View
    234

  • Download
    1

Embed Size (px)

DESCRIPTION

O Centro de Excelência em BRT Across Latitudes and Cultures (ALC-BRT CoE) promoveu o Bus Rapid Transit (BRT) Workshop: Experiences and Challenges (Workshop BRT: Experiências e Desafios) dia 12/07/2013, no Rio de Janeiro. O curso foi organizado pela EMBARQ Brasil, com patrocínio da Fetranspor e da VREF (Volvo Research and Education Foundations).

Citation preview

Page 1: BRT Workshop - The Customer Experience

1

The Customer Experience

Nigel H.M. WilsonProfessor of Civil & Environmental Engineering

MIT

email: [email protected]

Page 2: BRT Workshop - The Customer Experience

2

Outline

• The changing environment and customer expectations• Agency/Operator Functions• Customer Information Strategies• Recent Research

• Measuring Service Reliability• Role for Customer Surveys• Customer Classification

• Summary and Prospects

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 3: BRT Workshop - The Customer Experience

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

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 4: BRT Workshop - The Customer Experience

Key Transit Agency/Operator Functions

A. Off-Line Functions

• Service and Operations Planning (SOP)• Network and route design• Frequency setting and timetable development• Vehicle and crew scheduling

• Performance Measurement (PM)• Measures of operator performance against SOP

• Measures of customer experience

4Nigel Wilson, MIT

Rio de Janeiro, July 2013

Page 5: BRT Workshop - The Customer Experience

Key Transit Agency/Operator Functions

B. Real-Time Functions

• Service and Operations Control and Management (SOCM)• Dealing with deviations from SOP, both minor and major• Dealing with unexpected changes in demand

• Customer Information (CI)• Information on routes, trip times, vehicle arrival times, etc.

• Both static (based on SOP) and dynamic (based on SOP and SOCM)

5Nigel Wilson, MIT

Rio de Janeiro, July 2013

Page 6: BRT Workshop - The Customer Experience

Key Functions

6

Off-line Functions

Real-time Functions

Supply Demand

Customer

Information (CI)Service Management

(SOCM)

Service and Operations

Planning (SOP)

ADCSADCS

Performance Measurement (PM)

System Monitoring, Analysis, and Prediction

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 7: BRT Workshop - The Customer Experience

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"

77

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 8: BRT Workshop - The Customer Experience

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

88

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 9: BRT Workshop - The Customer Experience

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

9Nigel Wilson, MIT

Rio de Janeiro, July 2013

Page 10: BRT Workshop - The Customer Experience

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

10Nigel Wilson, MIT

Rio de Janeiro, July 2013

Page 11: BRT Workshop - The Customer Experience

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

11Nigel Wilson, MIT

Rio de Janeiro, July 2013

Page 12: BRT Workshop - The Customer Experience

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, …

12Nigel Wilson, MIT

Rio de Janeiro, July 2013

Page 13: BRT Workshop - The Customer Experience

13

Recent Research

• Measuring Service Reliability

• Roles for Customer Surveys

• Customer Classification

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 14: BRT Workshop - The Customer Experience

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

14

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, MIT Rio de Janeiro, July 2013

Page 15: BRT Workshop - The Customer Experience

Excess Journey Time (EJT)

15Nigel Wilson, MIT

Rio de Janeiro, July 2013

Page 16: BRT Workshop - The Customer Experience

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):

16

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 Time95th perc.

RBT

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 17: BRT Workshop - The Customer Experience

Line Level ERBT

17

Victoria Line, AM Peak, 2007

Trav

el T

ime

(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, MIT Rio de Janeiro, July 2013

Page 18: BRT Workshop - The Customer Experience

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

18Nigel Wilson, MIT

Rio de Janeiro, July 2013

Page 19: BRT Workshop - The Customer Experience

19

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

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 20: BRT Workshop - The Customer Experience

20

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

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 21: BRT Workshop - The Customer Experience

21

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)

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 22: BRT Workshop - The Customer Experience

22

LTDS vs. Oyster Stages for People with Weekday Travel Days

Avg. OR on All Captured

Weekdays

Avg. OR on All Possible Weekdays

LTDS on Travel Day

OR on Travel Day

LTD

S or

Oys

ter S

tage

s

20

15

10

5

0

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 23: BRT Workshop - The Customer Experience

23

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

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 24: BRT Workshop - The Customer Experience

24

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)

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 25: BRT Workshop - The Customer Experience

25

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)

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 26: BRT Workshop - The Customer Experience

26

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)

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 27: BRT Workshop - The Customer Experience

27

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

Source: "Classification of London’s Public Transport Users Using Smart Card Data." Meisy Ortega, MST Thesis, MIT (2013)

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 28: BRT Workshop - The Customer Experience

28

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 travelersNigel Wilson, MIT Rio de Janeiro, July 2013

Page 29: BRT Workshop - The Customer Experience

29

Travel Frequency

• 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

% o

f Oys

ter C

ards

Pay as You Go Period Pass242220181614121086420

1 2 3 4 5 6 7Number of Days

Number of Days

25

20

15

10

5

01 2 3 4 5 6 7

% o

f Oys

ter C

ards

Santiago, June 2010

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 30: BRT Workshop - The Customer Experience

30

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.

Activity: Refers to actions users perform between journeys.

Activity duration: Time lapsed between a tap-out and the subsequent tap-in.

% o

f Oys

ter C

ards

bei

ng o

bser

ved

durin

g w

eekd

ays 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.

510

.5-1

1.0

11.0

-11.

511

.5-1

2.0

12.0

-12.

512

.5-1

3.0

13.0

-13.

513

.5-1

4.0

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 31: BRT Workshop - The Customer Experience

31

Passenger Groups

Cluster Frequency Start Times Mode Type of Card

Regular Users

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

Occasional Users

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

LeisureTravelers

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 32: BRT Workshop - The Customer Experience

Visitor Travel Patterns Cluster

Regular Users

1. Everyday regular users

Non-

Exclusiv

e Commuters

2. All week regular users

3. Weekday rail regular users

Exclusiv

e Commuters

4. Weekday bus regular users

Occasional Users

5. All week occasional users

Non-

Commuter

Resident

s

6. Weekday bus occasional users

7. Weekend occasional users Leis

ure Travelers8. 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)

Visitor Oyster Card Cluster Distribution

Type of ClusterOccasional Regular

Cluster 1

Cluster 5

Cluster 2

Cluster 6Cluster 3

Cluster 7

Cluster 4

Cluster 8

% o

f Vis

itor O

yste

r Car

ds

80

70

60

50

40

30

20

10

0

% o

f Eac

h Sa

mpl

e

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.

510

.5-1

1.0

11.0

-11.

511

.5-1

2.0

12.0

-12.

512

.5-1

3.0

13.0

-13.

513

.5-1

4.0

12

10

8

6

4

2

0

Visitor Non-Visitor

04:0

0-04

:59

05:0

0-05

:59

06:0

0-06

:59

07:0

0-07

:59

08:0

0-08

:59

09:0

0-09

:59

10:0

0-10

:59

11:0

0-11

:59

12:0

0-12

:59

13:0

0-13

:59

14:0

0-14

:59

15:0

0-15

:59

16:0

0-16

:59

17:0

0-17

:59

18:0

0-18

:59

19:0

0-19

:59

20:0

0-20

:59

21:0

0-21

:59

22:0

0-22

:59

23:0

0-23

:59

Start Time%

of E

ach

Sam

ple

121110

9876543210

Visitor Non-Visitor

Page 33: BRT Workshop - The Customer Experience

33

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

Clus

ter 1

Clus

ter 2

Clus

ter 3

Clus

ter 4

Clus

ter 5

Clus

ter 6

Clus

ter 7

Clus

ter 8

% o

f eac

h cl

uste

r of O

yste

r Car

ds

70

60

50

40

30

20

10

0

Cluster 1 First and Last Journey Start Times

Start Time

Rela

tive

% o

f Oys

ter C

ards 14

12

10

8

6

4

2

0

04:0

0-04

:59

05:0

0-05

:59

06:0

0-06

:59

07:0

0-07

:59

08:0

0-08

:59

09:0

0-09

:59

10:0

0-10

:59

11:0

0-11

:59

12:0

0-12

:59

13:0

0-13

:59

14:0

0-14

:59

15:0

0-15

:59

16:0

0-16

:59

17:0

0-17

:59

18:0

0-18

:59

19:0

0-19

:59

20:0

0-20

:59

21:0

0-21

:59

22:0

0-22

:59

23:0

0-23

:59

Registered Total

Cluster 2 Activity Duration

Activity Duration (hours)

Rela

tive

% o

f Oys

ter C

ards

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.

510

.5-1

1.0

11.0

-11.

511

.5-1

2.0

12.0

-12.

512

.5-1

3.0

13.0

-13.

513

.5-1

4.0

Registered Total

Cluster

Regular Users

1. Everyday regular users

Non-Exclusiv

e Commuters

2. All week regular users

3. Weekday rail regular users

Exclusiv

e Commuters

4. Weekday bus regular users

Occasional Users

5. All week occasional users

Non-Commuter

Resident

s

6. Weekday bus occasional users

7. Weekend occasional users Leis

ure Travelers8. Weekday rail

occasional users

Page 34: BRT Workshop - The Customer Experience

34

Oyster Card Attrition Cluster

Regular Users

1. Everyday regular users

Non-

Exclusive Com

muters

2. All week regular users

3. Weekday rail regular users

Exclusive Com

muters

4. Weekday bus regular users

Occasional Users

5. All week occasional users

Non-Com

muter 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) + 0.8632R2 = 0.9685

Apr-10 Jun-10 Sep-10 Oct-10 --- Log Regression

% o

f Acti

ve O

yste

r Car

ds

100

90

80

70

60

50

40

30

20

10

0

Number of months after observed week0 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- - - -

100908070605040302010

0

% o

f Acti

ve O

yste

r Car

ds

Months

Oct

-201

1

Nov

-201

1

Dec

-201

1

Jan-

2012

Feb-

2012

Mar

-201

2

Apr-

2012

May

-201

2

Jun-

2012

Jul-2

012

Aug-

2012

Sep-

2012

Oct

-201

2

Page 35: BRT Workshop - The Customer Experience

35

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

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 36: BRT Workshop - The Customer Experience

36

Summary

• 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

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 37: BRT Workshop - The Customer Experience

37

Prospects

Panel data combined with full journey OD estimation and journey time provides the basis for extensive customer experience and behavior analysis including:

• understanding impacts of changes in service and price

• understanding customer attraction, retention, and attrition

• informing "information push" customer information strategies

• documenting the impacts of marketing and promotional strategies

Nigel Wilson, MIT Rio de Janeiro, July 2013

Page 38: BRT Workshop - The Customer Experience

38

AppendixMIT 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)

"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)

Nigel Wilson, MIT Rio de Janeiro, July 2013