1 OPSM 405 Service Management Class 15: Yield management: introduction Koç University Zeynep Aksin [email protected].

1

OPSM 405 Service Management

Class 15:

Yield management: introduction

Koç University

Zeynep [email protected]

2

Service Delivery SystemCustomer Demand

Limited Capacity

Fundamental Problem:

Variable Usage

Services cannot be produced in advance and stored for later consumption; they must be produced at the time of consumption.

3

Matching supply and demand in services Management options

• reject demand• inventory excess demand (queueing)• modulate capacity (add facililities, scheduling,

resource allocation)• modulate demand (pricing, yield management)

Primary considerations• return on assets• operating costs• revenue losses (opportunity costs)• service levels

4

Successful implementations

American Airlines

– $1.4B additional revenue over three-year period

– “Selling the right capacity to the right customer at the right price”

Hertz

– 1-5% revenue increase annually ($10-50M per year) Marriott Hotels

– $25-35M additional revenue in 1991 Royal Caribbean Cruise Line

– $20M+ additional revenue per year

Source: Arthur D. Little, 1992

5

When is this strategy appropriate?

Limited flexibility in supply Variable/uncertain demand Price flexibility/segmentation possible Available data Examples:

– airlines– hotels/resorts/theme parks– car/equipment rental – cruise ships– freight shipping– theater/performing arts– broadcasting (TV, radio,etc.)– utilities (elec., telecom)

1970s

1980s

1990s

6

Yield Management System

Reservation System

Forecasting

Overbooking Levels

Discount Allocation

current demandcancellations

cancellation rate estimates

overbooking levels

futuredemandestimates

fare class allocations

7

What is revenue/yield management?Two Perspectives:

1) A Market Segmentation Strategy (capture consumer surplus)

p

q

p

q

p0 p2p1

Create separate“fare products”

Intelligently allocate fixedcapacity to products

NOTE: Segmentation may make sense even with static allocation! Segmentation can also provide value (e.g. cancellation option)

8

Segmentation/product design

Ideally, we would like to discriminate (sort) customers based on their actual willingness-to-pay (reservation price).

$0.00

$50.00

$100.00

$150.00

$200.00

$250.00

C4 C2 C3 C1 C5

Cust. Res. PriceC1 $120C2 $180C3 $167C4 $230C5 $ 45

=====$742

ConsumerSurplus = $742

But willingness-to-pay is usually unobservable!

Ex:

9

So we try to find a variable that is correlated with willingness-to-pay (a “sorting mechanism”)

Cust. Res. Price Adv. Purchase?C1 $120 YESC2 $180 NOC3 $167 YESC4 $230 NOC5 $ 45 YES

Create two produce (advance/late purchase) with two prices: Adv: $100 Late: $150

$0.00

$25.00

$50.00

$75.00

$100.00

$125.00

$150.00

C4 C2 C3 C1 C5

ConsumerSurplus = $500

10

Sorting mechanisms Time of purchase/usage

– advanced/spot purchase– day-of-week/season

Purchase restrictions– cancellation options– minimum term– Saturday night stay

Purchase volume (individual vs. group) Duration of usage (single night/weekly rate) Customer affiliation

– corporate– contract user

Finding a good sorting mechanism is an art and requires a certain amount of trial and error.

11

2) Matching Price to Demand (peak-load pricing)

High Low

Discount

Full Fare

Demand

Pricexx

xxxxxxxx

xxxxxxxxxx

x

Create a small number of“price points”

Allocate more capacityto low price points if demandis weak; allocate more capacityto high price points if demand isstrong

12

Example: Using capacity controls for peak load pricing

Capacity = 100 seats

Demand Rev. Demand Rev.

$50 fare 30 $1,500 150 $5,000

$25 fare 80 $2,000 20 $500$75 fare 2 $150 80 $6,000

$2,150 $6,500

Off-Peak Day Peak Day

SinglePrice

TwoPrices

13

Example

2 FlightsCapacity = 3 seats$800

$700

$400

$300

$200

Ex: 5 customers with different valuations

NOTE: We usually cannot observe these valuations in practice

8:00 AM 1:00 PM

14

Best single price: $700Revenue: 2 x $700 = $1400

Maximum obtainable revenue$800 + $700 + $400 + $300 +$ 200 = $2400

Only 58% of maximum achieved!

$700

$400

$800

$300

$200

priced out

$700 $700

15

Discrimination via a “sorting mechanism”

$800

$700

$400

$300

$200

Customers returning by Saturday

Customers staying a Saturday

A trait that is correlated with willingness to pay allows for discrimination

- Saturday night stay - Advance purchase req. - Distribution channel (e.g. internet)

16

Price discrimination: SA stay: $400 No SA stay: $700Revenue: 2 x $700 + 1 x $400 = $1800

Maximum revenue$2400

Now 75% of maximum achieved!

$700

$400

$800

$300

$200

priced out

$700 $400 $700 $400

17

$400

Implement dynamic pricing

Capacity-controlled fares can be used to dynamically adjust the “effective price” of each departure.

$700

$800

$300$200

$700

$400

8:00 AM 1:00 PM

$700

$400

Priced outWe would like to price the empty flight to attract more traffic! How?

18

$400

$700

$800

$300

$200

$700

$400

$200

8:00 AM 1:00 PM

$700

$400

$200 XNo seats available

Revenue = 2 x $700 + 1 x $400 + 2 x $200 = $2200

2 seats available

92% of maximum!

Capacity-controlled deep discount

19

Example summary:2 Flights3 Seats each

1) Single price $1400 (+0%)

2) Two prices w/ sortingmechanism $1800 (+29%)

3) Two prices w/ sortingmech. & capacity-controlleddeep discount $2200 (+57%)

Forecasting demand

Data requirements– high-level of detail (origin-destination, fare-class, day-of-week,

departure time)– quantities tracked

• demand for each rate category/fare-class/departure• cancellation rates• no-show rates/ go-show rates

– daily processing Forecasting issues

– seasonalities– trends– special events

Good forecasting and accurate data are essential

21

Announcement

Midterm exam on Wednesday March 26 Will start at 12:30 sharp and end at 13:59 All topics until revenue management