M I T I n t e r n a t i o n a l C e n t e r f o r A i r T r a n s p o r t a t i o nM I T I n t e r n a t i o n a l C e n t e r f o r A i r T r a n s p o r t a t i o n

MIT MIT ICAT ICAT

WILLINGNESS TO PAY AND COMPETITIVE REVENUE MANAGEMENT

Peter P. Belobaba

With help from the MIT PODS Team:Maital Dar, Valenrina Soo, Thierry Vanhaverbeke

12th PROS Conference Houston, TX

March 5-8, 2006

MIT MIT ICAT ICAT

2

Traditional RM Systems Are Struggling Under Fare Simplification

Simplified fare structures characterized byOne-way fares with little or no product differentiation, priced at different fare levels Existing RM systems employed to control number of seats sold at each fare level

But RM systems were developed for restricted faresAssumed independent fare class demands, because restrictions kept full-fare passengers from buying lower faresTime series forecasting models used to predict future demand based on historical bookings in each fare classGiven independent demand forecasts, top-down protection for highest classes, extra seats made available to lowest class

MIT MIT ICAT ICAT

3

30,000

35,000

40,000

45,000

50,000

55,000

60,000

65,000

FullyRestricted

Remove AP Remove SatNight Min Stay

Remove AllRestr, Keep

AP

Remove AllRestr and AP

Revenue Impacts of Fare Simplification with Traditional RM Models

-0.5

%

-29.6

%

-45%

-16.8

%

MIT MIT ICAT ICAT

4

Traditional RM Models “Spiral Down”without Product Differentiation

0

10

20

30

40

50

60

70

80

90

100

Fully Restricted Remove AP Remove SatNight Min Stay

Remove AllRestr, Keep AP

Remove All Restrand AP

FC 6

FC 5FC 4

FC 3FC 2

FC 1

81.6

87.8

79.8 82.7

88.1

MIT MIT ICAT ICAT

5

Existing Airline RM Systems Need to be Modified for Changing Fare Structures

Without modification, these RM systems will not maximize revenues in less restricted fare structures

Unless demand forecasts are adjusted to reflect potential sell-up, high-fare demand will be consistently under-forecastOptimizer then under-protects, allowing more “spiral down”

Current RM system limitations are negatively affecting airline revenues

Existing systems, left unadjusted, generate high load factors but do not increase yieldsMany legacy carriers are using “rule-based” RM practices

RM forecasting models must be changed to reflect passenger willingness to pay (WTP)

MIT MIT ICAT ICAT

6

Forecasting by Willingness to Pay

With undifferentiated fares, forecasting must focus on demand by willingness to pay higher fares

Approach is to “transform” all historical bookings at different fares to maximum demand potential at lowest fare

Total demand potential at lowest fare converted to demand forecasts by fare level for future flights

Based on estimates of sell-up potential/WTP

But, most simplified fare structures still retain some product differentiation

Lowest fares can have more several cancel/change restrictionsHigher fares can offer full flexibility and additional amenities

MIT MIT ICAT ICAT

7

Hybrid Forecasting For Partially Differentiated Fare Structures

• Generate separate forecasts for price (“priceable”) and product (“yieldable”) oriented demand

A passenger is price-oriented if the next lower class from the one booked is closedA passenger is product-oriented if the next lower class from the one booked was open.

• Combine standard forecasts and WTP forecasts for input to RM optimizers

For product-oriented demand, bookings are treated as a historical data for the given class, and standard time series forecasting applied.For price-oriented demand, forecasts by WTP based on expected sell-up behavior

MIT MIT ICAT ICAT

8

PODS Simulations in Network D

Simplified Fare Structure 6 fare classesCompressed fare ratio of 4.1

2 competing hub airlines40 Spoke Cities252 legs482 OD markets

Class AP Min Stay

Chg Fee

Non Ref

1 0 0 0 0

2 3 0 1 0

3 7 0 1 1

4 14 0 1 1

5 14 0 1 1

6 21 0 1 1

Class 1 2 3 4 5 6Average Fare 412.85 293.34 179.01 153.03 127.05 101.06

MIT MIT ICAT ICAT

9

PODS “Network D”

H1(41)

19

28

H2(42)

43

2

1

10

98

7

6

5

15

1716

14

12

11 22

21

20

18

27

26

25

24

23

33

3231

30

29

39

38

37

36

3534

40

13

Traffic Flows

MIT MIT ICAT ICAT

10

Baseline Results: Standard RM Models

Standard pickup (moving average) forecaster, booking curve unconstraining and EMSRb optimizer

Revenue is $1,124,782 ; Yield is $0.1061

Load factor is 86.45%

Loads by Fare Class

5.6 7.410.2

13.8

25.923.7

0.0

5.0

10.0

15.0

20.0

25.0

30.0

1 2 3 4 5 6

Fare Class

MIT MIT ICAT ICAT

11

Spiral Down is Evident in Standard Forecasts

Forecast + Bookings by Fare Class

0

20

40

60

80

100

120

140

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

FC1 FC2 FC3 FC4 FC5 FC6

Very High Initial Forecasts in Classes 5 + 6

MIT MIT ICAT ICAT

12

Airline 1 Implements Hybrid Forecasting

Loads by Fare Class

0.05.0

10.015.020.025.030.0

1 2 3 4 5 6Fare Class

Hybrid Standard

Airline 1

Revenue

Yield

Loads

∆ from Trad RM

Airline 2 ∆ from Trad RM

$1,163,408 + 3.4% $1,118,299 - 0.2%

$0.1107 + $0.005 $0.1016 - $0.001

85.70% - 0.75 86.40% + 0.26

MIT MIT ICAT ICAT

13

Both Competitors Use Hybrid Forecasting

Airline 1

Revenue

Yield

Loads

∆ from Trad RM

Airline 2 ∆ from Trad RM

$1,158,167 + 2.97% $1,152,222 + 3.03%

$0.1098 + $0.004 $0.1055 + $0.004

85.95% - 0.5 85.72% - 0.68

Loads by Fare Class

0.05.0

10.015.020.025.030.0

1 2 3 4 5 6Fare Class

Hybrid Standard

MIT MIT ICAT ICAT

14

WTP Approach Brings Forecasts Back into Line

Forecast + Bookings by Fare Class

0

20

40

60

80

100

120

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

FC1 FC2 FC3 FC4 FC5 FC6

Higher Forecasts

Classes 1-3

MIT MIT ICAT ICAT

15

Distorted Forecasts Affect RM Performance

Spiral-down leads to high forecasts in lower classesAnd, in turn, forecasts that are too low in higher classes

In EMSR-based leg RM, at least the lowest class demand is rejected when demand is high

Not revenue maximizing, but some benefit from booking limits

Recent PODS simulations illustrate impacts of distorted forecasts on O+D method performance:

Mismatch between independent path/class demands assumed by network optimizer and reality of passenger sell-upNetwork optimizers over-protect for forecast low-class connecting traffic, leading to distorted bid prices or displacement costs

MIT MIT ICAT ICAT

16

Impacts on O+D Revenue Gains Under Simplified Fare Structure

Revenue performance of O+D methods is affected:Standard forecasting assumes path/class independenceIncorrect forecasts fed into network optimizers (LP, ProBP)Network optimization methods more affected than Heuristic BP

Rev

enue

Gai

n ov

er E

MSR

b

0.74%0.74%

0.43%0.43%

1.93%1.93%

0.00%

0.25%

0.50%

0.75%

1.00%

1.25%

1.50%

1.75%

2.00%

DAVN PRO BP H BP

MIT MIT ICAT ICAT

17

All RM Methods Benefit From Hybrid Forecasting

Use of Hybrid Forecasting with WTP component improves RM revenue gains by 2-4%

O+D RM Methods once again outperform EMSRb leg controls by 1% or more

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

4.5%

2.6%2.6%

3.3%3.3%

2.9%2.9%1.7%1.7%

0.74%0.43%

1.93%

2.6% 4.0% 3.3% 3.7%

Rev

enue

Gai

n ov

er E

MSR

b

EMSRb DAVN ProBP H BP

MIT MIT ICAT ICAT

18

Simulations in Larger Network R4 Competitors with 4 Hubs

All methods benefit from Hybrid ForecastingProBP again is most affected by simplified fare structure, but benefits most from use of Hybrid Forecasts based on WTP

0.0%

0.5%

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

EMSRb DAVN ProBP H BP

1.3%1.3%

0.8%0.8%

2.3%2.3%

1.9%1.9%

1.51%

0.65%0.44%

1.3% 2.4% 3.0% 2.4%

Rev

enue

Gai

n ov

er E

MSR

b

MIT MIT ICAT ICAT

19

Existing RM Forecasts Are Inadequate for Simplified Fare Structures

Existing RM systems need to be modifiedMismatch between RM model assumptions and fare structures

Price/product hybrid forecasting increases revenuesCompared to use of standard RM forecasting methodsGains come from higher forecasts in upper/middle classes, increasing protection and helping to reduce “spiral down”

Modified forecasters require estimates of passenger WTP by time to departure for each flight

Approach is to forecast maximum demand potential at lowest fare, and convert into “partitioned” forecasts for each fare class

But, WTP forecasting is much more difficult…

MIT MIT ICAT ICAT

20

Standard RM Forecasts Assume Independent Demands by Class

Demand observable in each DCP in any

fare classFare class

DCP6

5

4

3

2

1

1 2 3 4 5 6 7 8

Current DCP

Product-oriented demand observed for any open class by DCP

MIT MIT ICAT ICAT

21

Forecasts of Demand to Come by Class Used as Inputs by Optimizer

Demand to come in any

fare classFare class

DCP6

5

4

3

2

1

1 2 3 4 5 6 7 8

Current time-frame

Demand to come by class assumed to be independent of what classes are open or closed.

MIT MIT ICAT ICAT

22

Price-Oriented Demand for Undifferentiated Fares

Fare class

On a single flight departure, bookings in each class observed only when lower class was closed down.

DCP

Historical information obtained when j was the lowest open class

1 2 3 4 5 6 7 8

1

2

3

4

5

6

MIT MIT ICAT ICAT

23

Sample of Price-Oriented Demand over Multiple Departures

Fare class

With information about class closures and observed bookings, we can estimate WTP and sell-up

But we can’t estimate what we’ve never seen!

DCP1 2 3 4 5 6 7 8

1

2

3

4

5

6

MIT MIT ICAT ICAT

24

Factors Affecting WTP Estimates

Time to departureBusiness travelers with higher WTP book closer to departure date

Peak vs. off-peak periodsHigher demand periods tend to have higher WTP

Market characteristicsLimited competition and low capacity (relative to demand) means a higher WTP among consumers

RM seat availability – your own and your competitors’Weak RM controls on previous flights mean historical bookings don’t reflect higher WTP of consumersHigh availability of low-fare seats on competitor makes it difficult (impossible) to observe high-fare bookings

MIT MIT ICAT ICAT

25

Willingness to Pay Relative to Lowest Fare Changes over the Booking Process

0.00

1.00

2.00

3.00

4.00

5.00

6.00

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

HIGH

MEDIUM

LOW

BOOKING PERIODS (DCPs)

MIT MIT ICAT ICAT

26

WTP Estimates are Critical to Effective RM under Simplified Fares

Nothing simple about RM models required by these simplified fare structures

Traditional RM methods do not maximize revenuesWithout forecasting modifications, even O+D control gains are affected

New approaches to “hybrid” forecasting of price- vs. product-oriented demand show good potential

Significant revenue gains over standard forecasting methods

Recent PODS research shows potential for conditional WTP estimates:

Requires separate estimation of WTP for each scenario of class closure, for own airline and potentially competitor(s)