<<

utdallas.edu/~metin Page 1 . Çakanyıldırım 2018 Updated Updated in Fall resourcestoprepare this document for teaching/training. various used Strategy Tactics Çakanyıldırım – – If you find any inaccuracies, please contact [email protected] for corrections. for [email protected] contact please inaccuracies, any find youIf Net Contribution Net Levels of Decisions History

  Outline  To usethisin your owncourse/training, pleaseobtainTo permission from Prof.

Prof.Metin Revenue Management utdallas.edu/~metin Page 2 of until 2006. load is 70.4%. bankruptcy break-even ofsales. 08 have filed for modes while the the while 1 cent per $ 137 carriers transportation 72.4% of seats cancellations. fill fill is about 3% higher in 2007- in higher 3% about is 99 the industry earned earned industry 99 the Domestic/international vary. 90- UnitedContinentalHoldings.com AA+USAir speed train High typically » » » 95-99 (thebest industry’s years5 ever) airlinesearned 3.5 cents on eachdollar

prices Invest into alternative Downsize Merge. Utilization The US average for all industries is around 6 cents. Over

– – – – – – perhaps due to flight flight to due perhaps

Gas Carriers Over Since deregulationin 78,

sales:

Some U.S. industry U.S.Some observations airline

 

  utdallas.edu/~metin Page 3 ,etc). rentalcars when supply is fixed when is supply adjust the demand! the adjust – is a solution: fixed supply: ., control the average price by changing the mix customers. changingmix of controlthe., average the price by Low willingness to pay. i.e High willingness to pay » » » Limit Limit the numberof tickets sold at a low price, If adjusting supply is impossible impossible is supply adjusting If into customers Segment Doctor’savailability forappointments. Advertising time number (limited of time slots). bandwidth. Telecommunications program. MBA the of Size industries (fixed number of seats, rooms, rooms, seats, of number (fixed industries Travel – – – – – – – – Revenue management management Revenue Examples of of Examples

 Matching supply to to supply demand Matching utdallas.edu/~metin Page 4 that have differentthat have WTP. demand. demand. anagement anagement low. segments segments is customers customers buying economy seat). leisure is soldin advance of an additional sale charged to different be always served. Capacity prices cannot (e.g., business & same same unit of capacity (e.g., ) can be used to deliver variable cost of storable, degrading/unavailable units. Different - The : The Non segments segments All customers All customers Is revenue management for incoming MBA class revenueIs possible? management for incoming MBA capacity margins: margins: is not illegal or morally irresponsible to discriminate the customers. the to discriminate irresponsible morally or illegal is not side - gross gross side In contrast, moving curtain between economy and business sections on domestic flights of small small of flights on domestic sections business and economy between curtain moving contrast, In airlines. European It -

» » imited capacity: imited

L Costly adjustment of sold capacity: Perishable capacity: High Homogenous segmentation: Customer services to different

– – – – – –

Supply

Demand uitable Contexts for Revenue M Revenue for Contexts uitable

 S utdallas.edu/~metin Page 5 A phonograph phonograph cover A became operational became ://www.youtube.com/watch?v=07LFBydGjaM Read in American Way Magazine of Read Way Magazine in American https some flights: some Southwest LA flights. flights. LA - of Airline Travel in the USA the in Travel Airline of except for for except assengers in the as assengers picture on the right. rofessionals: At the beginning At the rofessionals: beginning of career his Frank Sinatra is to said but could not confirm independently. but Airlines American could not independently. confirm P P play on NYC See M. Cobb attendant M. flight See – – –

Air travel was an exciting experience but now became a but now became torture experience was travel an Air exciting endured, to be Flights used to feature good service, even live music by music live even service, good to feature used Flights 1950’s, 20% of Americans have ever flown on an an on airplane. flown ever have Americans of 20% 1950’s, system) (reservation Sabre 1964,

History

is taken from Barry Smith’s AGIFORS presentation. AGIFORS Smith’s Barry from taken is

This picture of manual reservation system reservation picture manual system of This

  utdallas.edu/~metin Page 6 DC for $70. This $70. DC for  years inflation. with 4% years Dallas Board corresponds tocorresponds $700 over 60 In 1950s, OW (one way) (one way) ticket 1950s, OW In Good Service & High Cost Then Cost High Service & Good Auronautics Controlled by Civil

AGIFORS presentation. AGIFORS

This is from Barry Smith’s Smith’s Barry from is This

Deregulation in 1978 removed the restrictions on schedules and fares scheduleson restrictions the removed 1978 in Deregulation Published Fares Fares Published utdallas.edu/~metin Page 7 LA - was established circa 1980wasestablished Chicago and Orleans New - Ultimate Ultimate SuperSaver : : PeopleExpress PeopleExpress on the verge of bankruptcy of verge the on butter routes: Newark - PeopleExpress and - costcarrier - 1984 by serving underserved markets: passengers Leisure markets: underserved serving by 1984 PeopleExpress Donald Burr, CEO of of CEO Burr, Donald charged 70% less than major airlines American’scounterattack – American vs. American We We had great people, tremendous value, terrific growth. We did a lot of things right. But we didn’t get our hands around the and automation issues. Discount fare if reserving weeks2 before departure stayingand over a Saturday night. Restrictions on the number of seats available under Ultimate Super Saver program Fast growth over 1980- over growth Fast attackedIn 1984, breadAA’s PeopleExpress – – – – – – September 1985, 1985, September Leisure passengers used SuperUltimate Saver while Business paidtravelers the fullfare. January 1985, 1985, January After deregulation, low After deregulation,     utdallas.edu/~metin Page 8 1986 Optimizer, 1979 1980 System, 1982 System, 1981 Allocation Modeling Modeling Allocation System, Report, OptimizationModeling AllocationDecision City Analysis City Allocation : Dynamic Inventory Inventory Dynamic : : Super Super : : Multiclass: : Discount: : City City :

SCARS DINAMO DADS CARS MOMS companies came companies came afterwards.

Airlines have been in the RM game for a long while. Hotels and Car Rental Rental Car and Hotels while. a long for RM game the in been have Airlines      American’s Experience American’s 1980’s Experience RM in with utdallas.edu/~metin . Page 9 United Northwest Delta Western Jet Value AirTran Southwest AA US Air Continental America West TWA 2013 in US. February 14 issue of WSJ of 14 issue February US.in 2011 2010 2007 2005 merger horizontalmerger linesare thicker. - 2000 S. McCartney, 2013. How the AMR Deal Might Affect Airline Service Airline Affect AMR Might the Deal 2013. How S. McCartney, 1995 1990 Time isaxis not drawn to scale. PostTime Based on the data in in on data the Based 1985 1978 AA Deregulation Delta United US Air

Western AirTran Value Jet Value

Northwest Southwest Continental Trans World Trans

America West Airline the MergersAirline Wake in of Low Profits utdallas.edu/~metin

Page 10 . - - - - - % Spirit 3.3 3.3 Spirit Other JetBlue 23.4 JetBlue - - - - - 6.8 32.7 13.7 in US. February 14 issue of WSJ of 14 issue February US.in South west % 692 1289 1239 4.2 3.7 5.2 5.9 5.5 3.9 21.1 % Delta ofAircrafts 898+624=1,522 # - - - - 1.7 3.0 46.7 18.3 13.8 36.7 37.2 % United 24.9+13.8=38.7 % 2.2+89.9 3.1+75.5 3.9+47.0 83.6+02.7 68.1+02.3 35.9+02.8 19.0+02.3 18.0+01.4 2012 Revenue2012in $B AA+US Air AA+US Air S. McCartney, 2013. How the AMR Deal Might Affect Airline Service Airline Affect AMR Might the Deal 2013. How S. McCartney, 473,062 352,904 685,666 472,913 459,844 705,721 693,485 557,302 Southwest Delta United AA+US Air seats / wk Departing More than 50% share in CLT (Charlotte, NC), DFW, PHL, MIA, PHX. MIA, PHL, DFW, NC), (Charlotte, in CLT share 50% than More Based on the data in in on data the Based 3,111 4,791 3,809 6,053 3,535 7,541 4,905 3,498 week / Flights

JFK CLT PHL MIA PHX LAX ORD DFW

Airport US Airlines & AA when Merge Air US Airlines US utdallas.edu/~metin Page 11 Okla. Okla. .” provider.” provider.” expected to stay system system - travel functions already raised staffing staffing airport raised [were] 23 satellite offices, satellite offices, 23 in 2010 and in 2010 and made the system from a command center near 17 using American’s Oct 17 flight American’s using program” 14 and 14 and [were] reservations Oct 27.” 27.” Oct and retiree- [American] : problems integration - Oct 17 Sat morning morning Sat integration date Oct 17 Corp. AAdvantage  There also .” big switch [was] run switch big the transition period, transitionthe period, Airways Airways up in 2005, and linked America Airlines West centers went live on Oct “American also stopped offering a US Airways Airways schedule beyond “American US a stopped also offering American.” Oct. 17, forcing passengers to shift bookings to On Oct. 17, US Airways Airways operates fewer ~200 flights to its three Oct. On 17, US less than usual) to lighten load the on the hubs (11% system. “For “American’s “[United and Continental] merged Continental]“[United and merged transition in early 2012, disrupted passengers employees and alike for weeks “US executed poorly transition on their two years IT later revenue cargo, Airways “The US2015:Before March accounting, employee and into folded frequent was flier plan Airways’ 2015: “US March the Americanlarger July: “Americanstarted … transferring reservations made on Airways for travel USon or after codes and to its Sabre more 20%.”than by levels headquarters... its switched switched to American versions” Tulsa, reservations database in Sabre’s at including one The clock the until around open – – – – – – – – arlier mergers suffered from suffered mergers Earlier AA’s gradual transition . 18   Oct . issue of WSJ . Oct 15 Weekend . American Airlines’ Reservations Test: Few Glitches in American Switchover in Glitches Few Smooth Travel This This Travel Smooth issue WSJ of Based on Carey. 2015. S. S. Carey. 2015. Carey. S. ended on Oct 29, 2015

AA and US Air Merger started in 2013, 2013, in started Merger Air and US AA

with a victorious email victorious email announcement to passengers with a

-20152013 over American = US + Air American AirlineMergers: utdallas.edu/~metin Page 12 - of - city lead to different could could , etc. , points points frequent flyers , and same and same starting children , stayover , , see next Travel page: agentperson in Travel vs.agent Same Same round trip route but different starting points can lead to senior citizens , Saturday night origin needed to fly the passengers from of from town out fly origin to passengers the needed Same Same round trip route - of rates for Hawaii residentsHawaii forrates - Customer Segmentation Customer based segmentation Kamaaina book earlier Any Any logic here? This may be a relic of the past as it does not make much sense in the age of . » hannel Revenue Management Strategy Management Revenue software vs. Airline’s website. Theme Parks:Theme Disney’s resident residentsOrlando forrates Cruise Lines: City Hotels: Origin based segmentation: C Government employees Government International pricing: They acceptThey or prefer Finer segmentation is also possible among Leisure and Business Passengers dependingon flexibility,sensitivity, etc. scheduleprice prices in different countries. different prices in different countries. Leisure passengers highly are price sensitive They haveThey more flexibility in departure and arrival times – – – – – – – – – – – – In addition to these, hotels, theme parks and cruise lines can have have can lines cruise and parks theme hotels, these, to addition In Othercustomer airlineinclude segments Leisure Passengers vs. Business Passengers origin based pricing based origin    utdallas.edu/~metin

Page 13

Business Customers Business Leisure Customers Leisure 800 Air Air .com .com Large Agent Travel Travel Travel Travel Opaque Frequent Number lines.com agent.com Corporate Distressed Flyer Program Flyer Worldspan System Global Distribution Sabre, Amadeus, Galileo, Galileo, Amadeus,

System Reservation Computerized Reservation and Distribution System Distribution and Reservation utdallas.edu/~metin Page 14 that customers customers (seats,rooms,rental cars) -paying low . the number of products products of number the products are sold to high-paying number number of products to sold . or in lower classes for a fare class is the number of products reserved for for reserved products of number the is class a fare for Booking Control Booking ) for a fare class is the maximum number of products that can can that products of number maximum the is class fare a for y ) b thatclass ensure that limited or higherclasses To ensure that enough To Booking Limits Levels Protection Limits & Booking – – soldto customer classes Protectionlevels Protectionlevel ( Booking limit ( Bookinglimit in be sold Booking control: Controlling Controlling control: Booking class      utdallas.edu/~metin Page 15 - paying paying rooms is rooms customers 50 rooms the booking booking the customers 200-50=150 limit for low for limit high-paying protected for for protected Capacity 200 rooms 106 can 106 can serve both class class R and Q. The rooms are not physically The rooms are not physically allocated to classes. Allocation Allocation virtual. is Room number classes. - 200 of rooms rooms of level customers protection paying and high-paying -paying maximum number number maximum 50 paying high-paying paying customers? customers? No! -paying . . is This customers and high-paying -paying , is the is the , customers and high-paying -paying low , is the number of number , is the customers and high-paying -paying the is the , customers -paying paying customers. -paying Effectively, the booking limit for these is the capacity. capacity. is the these for limit booking the Effectively, , low , is the number of rooms reserved reserved rooms of is number , the customers high-paying low booking limit booking low . customers paying paying customers? customers? No! high-paying for limit paying customers -paying reserved for low levels to

150 low for the the for limits high-

rotection level level for low rotection ne for the the ne for customers. customers. rooms the capacity. maximum number of rooms to sell to low to sell of rooms number maximum p the one for one one for o to sell booking for the another

2 Fare Classes at at a Hotel Classes 2 Fare Booking Limit vs. Protection Level Level Protection vs. Limit Booking

paying customers: high-paying for level protection the and customers paying room. Limits and levels for nested classes. for nestedand levelsLimits room. protection booking – – – – – – 0

Only two numbers above differ from the capacity: booking limit for the low the for limit booking capacity: the from differ above two numbers Only

The The

The The

2 classes: low same for thepayingcustomers and high-paying utdallas.edu/~metin Page 16 3 . b 322 2 paying customers. b 1 b 232, C= nested limits nested limits betterare 223, B= 1. only. = is where 2 A can be illustratedcan be with at least classes.3 class Class 3 Class 3 Class Class Class 2 Class Class 2 Class 2 Class class {2} {2} class → → → → → → 2 price] 2 price] single 3 2 classes. Sequence Nestedness lass 2 Class 2 Class Class Class Class 3 Class 2 Class → → → → → → 2 price] paying customers have priority over low- - 2 3 and class 3 rejected 2 price]+[Class 3 price] 3 price]+[Class 3 price]+[Class Class 2 Class 2 Class Class2 price]+[Class 3 price] Class Class Class 2 Class 3 Class 2*[Class limit limit is set for a ======accepted B C a class and all lower class anda class 2 Revenue=[Class Revenue = [ , Multiple Fare Classes Fare Multiple Nested Booking Booking Nested Limits capacity=2; capacity=2; Revenue = limit for class {3} is 1; booking limit for for limit is 1; booking {3} class for limit Sold capacity=2; Revenue=[Class Sold capacity=2; Revenue=[Class Sold Soldcapacity=2; Revenue = [Class2 price]+[Class 3 price] Class 2 acceptedClass Sold Sold capacity=2; » » » » » » » Request Arrival Sequence C Arrival Request Booking Booking Sequence A Arrival Request Sequence B Arrival Request Request Arrival Sequence Arrival Request Booking limit for class {3} is 1; booking limit for classes {2,3} is 2. {2,3} classes for limit is 1; booking {3} class for limit Booking Sequence A Arrival Request Sequence Arrival Request – – – – – – – – Partitioned limits Partitioned(Allotments): limits A Nested limits: Nested limits: limit A is for

Allotment Allotment violates RM’s first principle: high Soldcapacity and revenues are the same except for

request arrivals{Class 2, Class 2, Class 3} in 3 sequences: A request arrivals{Class 2, Class 2, Class 3} in 3 sequences:

3

Supposeclasses are 1,2,3, and 1 is the highest class.

  utdallas.edu/~metin Page 17 3 b 2 b 1 b 332 = 323, C 323, = In general, nested limits are better. . 3 3 233, B 233, A, B, C Class 3 Class Class Class 2 Class Class 3 Class 2 Class price] Class Class 2 price] 2 price] 2 price] → → → → → 2 3 lass 3 Class Class 2 Class 3 Class Class Class Class 3→ Class Sequences → → → → → price]+[Class 3 3 price]+[Class 2 3 price]+[Class 3 in all Class 2 Class Class 2→ Class Class Class Class Class 3 Class Class 3 Class 3 Class =[Class 2 price]+[Class 3 price] ======B C A ): Revenue = [ Revenue = [Class Revenue = [ Sequence Sequence Another Set of Arrivals Set of Arrivals Another capacity=2; Revenue = Arrival Arrival Arrival Arrival Soldcapacity=2; Revenue=[Class 2 price]+[Class 3 price] Sold capacity=2; Revenue = [Class 3 price]+[Class Sold capacity=2; Sold Soldcapacity=2; Sold capacity=2; » » » » » » Request Arrival Sequence B Arrival Request Sequence C Arrival Request Booking limit for class {3} is 1; booking limit for class {2} is {2} 1. class for limit is 1; booking {3} class for limit Booking Request Request Request Sequence Arrival Request Booking limit for class {3} is 1; booking limit for classes {2,3} is 2. {2,3} classes for limit is 1; booking {3} class for limit Booking Sequence A Arrival Request – – – – – – – – capacity and revenues are the same

Partitioned limits Partitioned(Allotments limits Nested limits: Nested limits:

On this page and the previous one, we have considered all arrivals involving 3 requests ofClass 2 and 3 all andpossible sequences. Nested limits are better ininstance one neverand worse. The other 3 request arrivals involving Classes 2 and 3 include only one class (e.g., 222 or 333) so they they 333) so 222 or (e.g., class one only 3 include 2 and Classes involving arrivals 3 request other The result in the same revenues under bothnested limits & partitioned limits. Sold

3 request arrivals{Class 2, Class 3, Class 3} in 3 sequences: A= request3 arrivals{Class sequences:in2, 3 ClassClass 3, 3} Multiple Fare Classes: Nested Booking Limits with with Limits Booking Nested Classes: Fare Multiple

  utdallas.edu/~metin Page 18 , 0 limit for class 2 3 class and for limit 0 0 limit for class 2 3 class and for limit 0 8. 2 2 >12. ≤ ≤ ≤ 6 accepted . customers high-paying rejecting 0+ 12+6 . , 12+6 12+6 Nested Booking Limits, Limits, Booking Nested 1 is the highest class. & and the booking limit for classes {2} is 12. {2} classes for limit booking the and and the booking limit for classes {2,3} is 20. {2,3} classes for limit booking the and A limit limit A is set singlea for class only. 2 customers, 8 limit for class 3; class for limit 8 8 limit for class 3; class for limit 8 ≤ paying customers while customers -paying 6 more 3 class is accepted more more 2 class is rejected more 6 6 0+ customers, 2 3 customers, class or more 12 to 12 class 6 6 more class 2 or 3 customers are are 2 customers 3 class or 6 more for accepting low accepting A booking limit bookinglimit A is setclass for a all andlower classes. Suppose classes 1,2,3Suppose are Request for for Request If class 2: < 0+0 2: class If class 3: If for Request     This is This Booking limit for class {3} is 8 {3} class for limit Booking 12 class 2 customers, to selling After for Requests Booking limit for class {3} is 8 {3} class for limit Booking selling After Requests » » » » » » »

Another Example with with Batch Example Arrivals Another

Partitioned limits Partitioned(Allotments): limits Nested limits: Nested limits:

Booking limits should be nested for batch arrivals as as well arrivals for batch be nested should limits Booking Batch arrivalsof customers to the system, reservationwhen they togethertravel Multiple Fare Classes: Classes: Fare Multiple utdallas.edu/~metin Page 19 6 5 b 4 b 3 b 2 b 1 b b class class 6. 1, 2, 3, 4, 5 or classes. Allocation is virtual. For For virtual. is Allocation classes. example, number room 206 can serve Rooms are Rooms are not physically allocated to it 5 bottles of of bottles , 2 show these these show Newspaper 1,3}, queens or a king). These services are are services These king). a or queens available classes classes whose available differ prices The rest is the same as same Class is the 5. rest The products: Exactly the same room. Show Show room. same the Exactly products: (2 at arrival, daily daily at arrival, The rest is the same 1. as same Class the is rest The The rest is the same 4. as same Class is the rest The The rest is the same 3. as same Class is the rest The Fruits in available classes={1,3}, show 3 classes={1,3}, in available , - in available classes={ -in available The rest is the same 2. as same Class is the rest The choice of of bed choice Internet everyday, everyday, days days before check days days before check , show show , differentiated minimally 2 2 in available classes={1,2,3,4,5}, show classes={1,2,3,4,5}, -in available check 30 days before show 5 classes={1,3,5}, -in available check 10 days before in available classes={1,3,5}, show these classes={1,3,5}, -in available check 10 days before show 6 classes={1,2,3,4,5,6}, -in available check 60 days before in available classes={1,2,3,4,5,6}, show show these classes={1,2,3,4,5,6}, -in available check 60 days before show these classes={1,2,3,4,5}, -in available check 30 days before No choice of bed type; type; bed of 6: No choice Class Internet at $5/day. $5/day. at 2: Internet Class fruits; 3: No Class newspaper; 4: No Class $3/bottle; at 5: Water Class Class 1: Free 1: Free Class water water size. to room are the that not specific -ons add just What What are Classesat Fare Hotel? a ...... and ask customers to choose customers ask and When When

only the minimum priced available class and ask customers to buy to buy customers ask and class available priced minimum the only Do not even minimally differentiate differentiate Do not even minimally Prices: $250; $220; $190; $160; $130; $100. $190; $160; $130; $100. $250; Prices: $220; are Prices $30. of differential price the justify cannot services extra of Cost customers. the to segment used and based, value but based cost not Consider a standard room, not a deluxe or a suit. a or deluxe a not room, standard a Consider

     Physically Physically but minimally differentiating the product

Different distributiondifferent channelsusecan selling practices:  utdallas.edu/~metin

Page 20 } 1 . { 𝑖𝑖 ↔ ,2} . 1 1 { = and = 0 𝑖𝑖 , 0 1 ,2,3} y b 1 … { , 𝑛𝑛 2 > , < is the lowest class. 1 ↔ 1 2 1} 𝑛𝑛 𝑖𝑖 b - = = = ,.,n Protection levels Protection 1 { 𝑛𝑛 , … , ,.,n} 𝑖𝑖 1 < …. < < { > …. > y = 1 2 - - 𝑗𝑗 n n b < > y Protection Levels & Protection 1 -1 n - n n n b Class Class Class 1 Class Class 2 Class 3 Class Class Class < > y is the highest class, while 𝑗𝑗 while class, highest the is n 1 y n+1 b = 𝑗𝑗 . = 0 𝑛𝑛 𝑦𝑦 , protectsfuturereservations or classes 𝑗𝑗 = … i 1 limits bookings for classes limits bookings {1,.,n} , + , 0 , 2 𝑛𝑛 , 𝑏𝑏 1 Multiple Fare Classes Fare Multiple = = 0 {2,.,n} ; 𝑗𝑗 𝑛𝑛 for class 𝑦𝑦 𝑖𝑖 for class i = 𝑖𝑖 1 nestedness 𝑏𝑏 {3,.,n} =

Booking limitsBooking

Capacity Capacity

1,.,n}

-

Becauseof

– Nested Booking Limits Booking Nested Limits {n Booking limit 𝑏𝑏 Protection level 𝑦𝑦 Classes indexed by

n}

 { utdallas.edu/~metin Page 21 } 1 0 y b {} {1,.,n } 2 } 1 1 b y { {2,.,n . } ] 3 1 2 ,2} b y 1 − { {3,.,n 𝑖𝑖 :

1

}

𝒚𝒚 𝒃𝒃

𝟏𝟏 − 𝒊𝒊 4 3 𝒊𝒊 y b ,2,3}

1

{4,.,n

{ capacity = + . 𝑖𝑖 𝑏𝑏 Protection Levels Protection − ]. . ] 1 1 2 -2} 1 𝑏𝑏 𝑛𝑛 n- n- : ,.,n y b -1,.,n} 𝑖𝑖 − + = 1 [ 𝑖𝑖 { 𝑖𝑖 {n : : 𝑏𝑏 1 1 − 1 -1} n − n- n} b 𝑖𝑖 { y ,.,n : 1 { 1 n capacity n+1 {} ,.,n} y b 1 = { protects future reservations classesor [ 1 − 1 𝑖𝑖 𝑦𝑦 − [Protecting classes : n 𝑖𝑖 -1 n 1 ≡ protects classes limits limits bookings for classes ] − ] ] i Class Class Class 1 Class Class 2 Class 3 Class 𝑖𝑖 𝑛𝑛 𝑛𝑛 : : : Class Class 𝑖𝑖 𝑖𝑖 1 [ [ for class 1 for − Booking Limits 𝑖𝑖 for class over over 𝑖𝑖 i 𝑏𝑏 b = protected imiting classes imiting [L

– Number Number Protection level 𝑦𝑦 Booking limit Booking limit

    Capacity utdallas.edu/~metin Page 22 Levels 0 1 y b 2 1 y b Protection Protection 2 3 y b and and 4 3 b y n-2 n-1 y b n n-1 b

y n n+1 y b

n -1 Capacity n Booking Limits Limits Booking Class Class Class 1 Class 2 Class 3 Class Class Class Summary: utdallas.edu/~metin Page 23 This approach is pursued in This approach is the appendix. Altering booking limits is confusing. 0 1 0 y b 100 limits 1 2 Booking y b 73 27 for class above 2 and class for Reduce booking limit Reduce booking limit 3 2 y b 12 88 3 4 y 4 b 96 levels 4 5 Protection y 0 b 100 6 5 0 b y 100 Protect Protect less Classfor 2 and above Example: Increasing Bookings Increasing Example: 5 Class 3 Class Class Class Class 1 Class Class 4 Class 5 Requested for Class 2 Class for 5 Requested utdallas.edu/~metin Page 24 . . Both Both . 1..5) to classes (applies reject reject the request . . . . =100 =100 1 , 0) b =93 ≥ 0, accept the request. request. the accept 0, ≥ =93 1 1 =86≥ 0, =86≥ , 1 -x , 0, 0,, 0, 0) 1 1 -x , 1 1 , 0, 0,, 0) 4, 0). , 1 -B 5 1 , -B , (1 b 1 3, 3, 3, 0, 0) 0, 3, 3, )+(3, b + )+(1 )+(5 0) , 0, 0, 0) for the request. the for 0) 0, 0, , =(12, 10, 4, 1, 0) . 0) 4, 1, 10, =(12, 5 , =90 ≥ 0, accept the request. request. the accept 0, ≥ =90 5 1 ( -x =66 ≥ 0 and ≥ =66 7, 7, 1,7, 1, 0 7, 1 2 =59 ≥ 0 and ≥ =59 5, 5, 0, 0, ) = 2 -x -B for the request. for the request. for the request. 5 (0, 0,0, 0,(0, 0 2 1 ) = (100, 12, 73, (100, ) = =95≥0, so accept the request. the accept so =95≥0, -x x 1 5 b 2 )= (6, 6, 0, 0, 0 -B , b -x 2 (2, 0, 0, 0, 0) 0) 0, 0, 0, (2, 4 -B 1 , , 0) for the, request. x 2 4 b + 1 ) = (0, 0, 0, 0, 0) , b 5 , 3 , =73 (applies to classes 2..5) and and to classes 2..5) (applies =73 x 1 B 3 2 , =94 ≥ 0, accept the request. request. the accept 0, ≥ =94 , , 0, 0, 0) for the request. the for 0) 0, 0, , , b b 1 2 4 1 1 , x -x 2 , , B , 1 =11 ≥ 0 , b ≥ =11 0 and ≥ =63 b 1 1 , 1 =4 ≥ 0 , b ≥ =4 3 2 (3, 3, 3, 0, 0) (4, 4, 4, 4, 0) (2, 0, 0, 0, 0) 3 3 = (6, 6, 0, 0, 0)=( 0, 0, 6, (6, = =(7, 7, 1, 1, 0 0)=( 1, 4, 10, =(10, (5, 5, 0, 0, 0, 0)= 5, 0, =(5, x , -B 1 1 -x -x B B B B B -x = ( = = = = ( 3 2 =68 ≥ 0 and and 0 ≥ =68 b 3 b , 2 2 x -B -B -B -x =( 3 2 =(10, 10, 4, 1, 0). 1, 4, 10, =(10, B 3 2 b b , B =(10, 10, 4, 1, 0) 0) 1, 4, 10, =(10, b 1 , B ; Let x ; Let x ; Let x ; Let ; Let x ; Let ; Let x=( ; Let =( B =3 ≥ 0 , b ≥ =3 0 , b ≥ =8 =-1<0 =67 ≥ 0 and and 0 ≥ =67 4 3 4 2 -x -x -x -x 4 3 4 2 =(88, 63, 8, 3, 0) ≥ 0 , accept the request. request. the accept , 0 ≥ 0) 3, 8, 63, -x=(88, -B -B -B -B Keeping Booking Limits Constant Limits Booking Keeping 4 3 4 2 -B b b b b b Example: Increasing Bookings while while Bookings Increasing Example: Set the bookings to B bookings the Set Since Keep as the bookings Since Since Set to bookings the current Set Set to bookings the current Since Set to bookings the current Set Set to bookings the current Since Class 2 bookings are restricted by by restricted are 2 bookings Class have enough room have room as enough

            2 requested for class 1; Let 4 requested class 4 for 4 requested 3 requested class 3 for 3 requested 1 requested class 4 for 1 requested 1 requested class 2 for 1 requested Current bookings= Current class 2 for 5 requested Suppose Booking limits= Booking Suppose

 

     utdallas.edu/~metin Page 25 for the request. )=(12, 10, 4, 1, 0). 5 B , 4 B , 3 B , (30, 30, 0, 0, 0, 0) 0, 30, (30, 2 B , 1 ,) = 5 B x ( , 4 x B= , 3 x , 2 x 4, 0). , 1 (20, 20, 0, 0, 0). 0, 0, 20, (20, x = =(10, 10, 10, 0, 0). 0, 10, 10, x=(10, x=( x =(62, 60, 4, 1, . 1, 0) 4, 60, =(62, 0) 0, 0, 20, (20, =(42, 40, 4, 1, . 1, 0) 4, 40, =(42, 0) 0, 0, 30, (30, + + . reject fails, 0 ≥ 0) 3, -1, ; Let ; Let ; Let . reject fails, 0 ≥ 0) 3, -2, )=(100, 12, 73, )=(100, , 5 class 2 class 2 b , 4 -x=(29, 4, -x=(28, 3, =(58, 33,8, 0≥ 0) 3, , accept. -x=(58, 13,8, 0≥ 0) 3, , accept. -x=(38, b Keeping Booking Limits Constant Limits Booking Keeping -B =(68, 66, 10, 1, . 1, 0) 10, 66, =(68, =(62, 60, 4, 4, . 1, 0) 60, =(62, , -B -B -B b 3 b b b =(74, 72, 10, 1, 0)+(1, 1, 1, 0, 0)=(75, 73, 11, 1, 0). 1, 11, 73, 1, 0, 0)=(75, 1, 1, 0)+(1, 10, 72, =(74, =(68, 66, 10, 1, 0)+(6, 6, 0, 0, 0)=(74, 72, 10, 1, 0). 1, 10, 72, 0, 0, 0)=(74, 6, 1, 0)+(6, 10, 66, =(68, =(62, 60, 4, 1, 0)+(6, 6, 6, 0, 0)=(68, 66, 10, 1, . 1, 0) 10, 66, 0)=(68, 6, 6, 0, 4, 1, 0)+(6, 60, =(62, =(42, 40, 4, 1, 0) 0) 1, 4, 40, =(42, =(12, 10, 4, 1, 0) 0) 1, 4, 10, =(12, B B =(25, 0, 1,0, 3, accept.-x=(25,≥ 0, 0) =(26, 1, 2,1, 3, accept.-x=(26,≥ 0, 0) =(32, 7, 2,7, 3, accept.-x=(32,≥ 0, 0) b Example: Increasing Bookings while while Bookings Increasing Example: B B B B B , -B -B -B 2 b b b b If Set If Set Set Since Keep Since Keep If Since Set Since Set , 1 b

             

=( 1 for1 class 3, x=(1, 1, 1, 0, 0). 6 for6 class 2, x=(6, 6, 0, 0, 0). 3 for3 class 3, x=(3, 3, 3, 0, 0). 6 for6 class 3, x=(6, 6, 6, 0, 0). 10 requested for class 3; Let 20 requested for 20 requested 30 requested for 30 requested b

 

 

 

  utdallas.edu/~metin Page 26 =0. 5 B =1; =1; class: Fifth 5 -B 4 B )=(75, 73, 11, 1, 0). 5 B =10; Fourth class: class: Fourth =10; 4 Only the first class can accept from now on. , 4 -B B 3 , B 3 B , 2 B , 1 B ( . . . B= =62; Third class: class: Third =62; 3 -B 4, 0). 2 B . reject fails, 0 ≥ -1) 1, 0,-1,0 3, fails,≥ 0)reject 1,-1,0 3, fails,≥ 0)reject -1, 0, 2, 0,-1,0 2, fails,≥ 0)reject )=(100, 12, 73, )=(100, , 5 , 1,, 3, 0), all low fare classes are closed. b =27; =27; class: Second , 1, 3, 0), all classes0), 1, 3, are , closed. 0 , 2 4 0 =(24, -x=(24, -x=(24, =(24, -x=(24, -x=(24, b -B , Keeping Booking Limits Constant Limits Booking Keeping 1 =(75, 73, 11, 1, . 1, 0) 11, 73, =(75, =(75, 73, 11, 1, . 1, 0) 11, 73, =(75, . 1, 0) 11, 73, =(75, . 1, 0) 11, 73, =(75, , -B -B -B -B 0 25, 3 b b b b =(75, 73, 11, 1, 0)+(25, 0, 0, 0, 0)=(100, 73, 11, 1, 1, 0). 11, 73, 0, 0)=(100, 0, 0, 1, 0)+(25, 11, 73, =(75, B B B B =(0, 0, 1,≥ 0,0) 3,-x=(0, accept. b Example: Increasing Bookings while while Bookings Increasing Example: =( =( B , -B B B 2 b - - b b b If Set Since Keep Since Keep Since Keep Since Keep , 1 b

First class: B class: First          

=( Recover bookings Recover Since 25 for class x=(25,1, class 25 for 0, 0, 0, 0). 1 for1 class 2, x=(1, 1, 0, 0, 0). 1 for1 class 3, x=(1, 1, 1, 0, 0). 1 for1class 4, x=(1, 1, 1, 1, 0). 1 for1class 5, x=(1, 1, 1, 1, 1). Since b

 

 

     utdallas.edu/~metin Page 27 . m to x; , 0, …, 0]. equal B+ and do , …, m , …, } m > 0 classes) 1 nonzero x=[ of : -B i 1 are b } . cancellation = a number is given.is elements i b length denotes 0 first (of

m< only requests for class vector a m is whose

x { Reject the request. the Reject Remaining_capacity . Booking limit , Accept the request and B:= and set request the , b≥B+x Accept else If Suppose Suppose := 0 Both B and b are vectors of length n. of length b vectors B and are Both EndWhile {Initialization} B { {Iterative step} While Booking Algorithm with Booking Limits Booking with Algorithm Booking utdallas.edu/~metin Page 28 for MBA; SCM, Finance, Energy… Finance, SCM, Fewer Fewer for Master. Course Degrees Many SOM OM, Many airlines. . RASM=Net yield * Load factor. * Load yield . RASM=Net spoke airlines ; airlines spoke point - - -to Increase by booking more. by more. Increase booking classes available seat seat mile available -3 mile. Increase by booking less. booking by Increase mile. Itineraries hub-and for Many point for few A Passenger Airline Passenger Seat Service 1 of resources. resources. of passenger : Revenue/Cost per per : Revenue/Cost : Utilization ASM /C Revenue per Resources in product a Number Number of types Products Resource unit Resource types Tactical Management Revenue Tactical RASM night. room available per Revenue REVPAR: Load factor Load Yield: – – – – Resources: Units of capacity managed by a supplier supplier a by managed capacity of Units Resources: to purchase. want customers What Products: customers. of to a segment allocated resources or Group class: Fare Periodic checking/updates: Daily checking of expected vs. realized demand, when 15 or fewer days until days fewer or 15 when demand, realized vs. expected of checking Daily checking/updates: Periodic expected. from significantly differs realized if limit booking Update days. more if frequently Less departure. change. Aircraft 2010; in Dallas in game NBA updates: -star all -driven Event

– – – – – Some terminology Some Tactical RM calculatingis and updating booking limits.

  utdallas.edu/~metin Page 29 shows. on the ship for the same number of days. days. of number same the for ship the on remains remains everybody everybody in only on Sundays and a minimum 7 days. at of least stay a and minimum on Sundays only -in & airlines. airlines. depart on a certain day on day a depart certain -point Exception: Some jazz clubs run multiple sittings on a single night. Some patrons stay for 2 sittings in a row. row. a in sittings 2 for stay Some patrons night. single a multipleException: sittings on Somerun jazzclubs to  - Point Cruises is the event the of time ending and start the lines, cruise to similar are events sport and Concert everybody. same for Med Resorts accepts-Med check Club » » » » Executing Management Revenue Executing Resorts, cruise lines, events are not overbooked. In these industries, there is no viable customer. overbooked to an of serving method airlineSome tickets are refundableparts and ofcustomer’s hotel costs can be recovered overbooking. use hotels and airlines both of these, Because up. show not does one when Network management for is not relevant management Network – – – Overbooking: Booking extra to compensate for no- for compensate to extra Booking Overbooking: Capacity Allocation: Booking for various fare classes. fare various for Booking Allocation: Capacity or of an network airline the flight across Booking management: Network days multiple ata hotel.    utdallas.edu/~metin

Page 30

Commerce & in Research Operations Commerce & - Planning, Marketing & Distribution”, INTERFACES 31:2 INTERFACES & Marketing Distribution”, Planning, Airline Smith, Guenther, Rao & (2001)“E Ratliff & Rao Guenther, Smith, Systematic Systematic View Revenue of Management utdallas.edu/~metin Page 31 Moderate Moderate Low Low low Very High High Importance Moderate Incremental Cost − Fees fuel, food, passenger fees Price+Ancillary Commissions Commission, Commission, processing, wear/tear wear/tear Extra sessions, TAs, Commissions, cleaning, wear/tear Commissions, cleaning, wear/tear Incremental Cost Incremental cleaning food, Commission, fuel handling, Repositioning, =

Casinos Events Rental cars Rental University Hotels Container shipping Passenger airlines Cruise lines

Relative importance of incremental costs depends on the industry the on depends costs of incremental importance Relative the net contribution. contribution. net the Incremental costs andIncremental feeobtained can from changeancillary products/services

 Net Contribution Net utdallas.edu/~metin Page 32 What toWhat expectin the future? Incremental Cost − Fees Importance Low Low-Moderate Moderate Moderate Moderate Moderate-High High high Very -free sales , internet, phone fees phone internet, , -phone, duty Price+Ancillary minibar parking merchandise, = fee, library fee, library fee beverage,

Ancillary fees Ancillary Food, beverage, ear Food, Activity Food, expediting Call before delivery, Insurance, gasoline sales on-board Gambling, losses Gambling Food, beverage, Ancillary fees becoming more important for passenger airlines. airlines. passenger for important more fees becoming Ancillary

dependent Relative Importance Fees. of Ancillary Importance Relative -dependent

Industry

Net Contribution Net

University Hotels Container shipping Rental cars Cruise lines Casinos Events Passenger airlines utdallas.edu/~metin Page 33 . reference price.) – minutes are integers, captured by by AA but not captured are integers, in minutes data (Expedia, Travelocityand Priceline) provide cheaper according to Oct 14 2015 Informs DFW Chapter presentation by to by Oct according Jim 14 DFW Diamond presentation Chapter 2015 Informs http://lastminute.travelocity.com Because of fareease visibility, of among communicationcustomer Recent Trends: Recent Practice RM in out, you will see that Expedia charges about 20% premium as distribution/commission fee. fee. you -out, will see as 20% charges about premium that Expedia distribution/commission 4. Enter 4. Enter name/credit card Whatinfo. exactly did buy? you 1. Choose location: Downtown San Diego & Harbor Island. 1. Choose location: San Diego Downtown & Island. Harbor 2. Choose star level: more. or 3.5 stars 3. Name price: your per $70 night. (You are prompted with a median price $187 per of night » » » » time departure and at most 14 minute late arrival per flight; binary data reported to FAA, DOT. to FAA, reported data binary flight; per arrival late 14 minute at and most departure -time On A14: D0, customer; per time arrival customer’s each in Delay the outside company, reported Go to priceline.com bid for a hotel in San Diego for three nights, starting on coming Saturday. Saturday. on coming starting nights, three for Diego San in a hotel bid for Go to priceline.com Attempting to book a hotel in San Diego: Google search for “cheap hotels San Diego” takes you to you takes San Diego” hotels “cheap for search Google Diego: in San a hotel to book Attempting to rate a pays wholesale Expedia hotel. the not Expedia pay and Expedia with a hotel book You Expedia.com. the after insist you If you insist. unless customers to rate wholesale the reveal not will hotel The hotel. the check Yapta.com: Price drop alerts help you to time the purchase. purchase. the to time you help alerts drop Price Yapta.com: Case in point: Strategic customers wait for discounting season. A higher percentage of merchandise is merchandise of percentage A higher season. discounting for wait customers retail Strategic in point: Case discounts. sold Can discounting. to resorting year remain every in the without business through retailers is trying. Dallas) Street, Main 1618 Square, Marcus (One Marcus Neiman minute deals: See deals: -minute

– – – – – –

Customer Customer specific satisfaction measures as opposed to timelydeparture and arrival Last Opaque products: Hidden product attributes. Online travel agents/intermediaries distribution channels. Strategic customer: the amount of purchases.their groups, customers groups, customers are better informed than before. ever mayTheytheoptimize timing,

  

  utdallas.edu/~metin Page 34 When When Ticket Buy a to Source: S. McCartney. The Best Day to Buy Airline Tickets. WSJ The Middle Seat Column, Oct 22, 2014. 2014. 22, Oct Column, Seat Middle The WSJ Tickets. Airline Buy to Day Best The McCartney. S. Source: utdallas.edu/~metin Page 35 (2005) 6 Chapter up reading: Phillips - Strategy Tactics

– – Follow Net Contribution Net History Levels of Decisions

    Summary utdallas.edu/~metin Page 36 0 1 0 y b 100 limits 1 2 Booking y b 73 Affected 27 3 2 y b 12 88 3 4 y 4 b 96 levels 4 5 Protection y Affected 0 b 100 5 Fare Classes 5 Fare 6 5 0 b y 100 5 Class 3 Class Class Class Class 1 Class Class 4 Class Example: Decreasing Limits & Levels & Limits Decreasing Example: 5 Requested for Class 2 Class for 5 Requested utdallas.edu/~metin Page 37 0 0 94 90 27 27 67 63 8 82 82 12 4 3 87 90 0 0 90 94 Request Rejected 0 0 94 90 2 5 2 5 Protection Levels Protection Class Class Class Class 1 Class Class 3 Class Class 1 Class Class 3 Class 4 for Class 4 Class 4 for 1 for Class 4 Class 1 for and and 5 Classes 0 0 93 95 27 27 66 68 82 83 11 12 3 90 4 91 Booking Limits 0 0 93 95 0 0 93 95

2 5 5 2

Class Class Class 1 Class Class Class Class Class 3 Class Class 1 Class Class 4 Class Class 4 Class Example: Example: 3 for Class 3 Class 3 for 1 Class for utdallas.edu/~metin Page 38 0 0 88 38 25 25 63 13 8 8 80 30 3 3 85 35 0 0 88 38 Request Rejected 0 0 88 38 4 4 2 5 1 5 1 2 Protection Levels Protection Class Class Class Class Class Class Class Class Class Class Class Class 3 Class 30 for Class 30 Class for 10 for Class 10 Class for 3 and and 5 Classes 0 0 90 58 27 25 63 33 8 8 50 82 3 3 55 87 Booking Limits 0 0 58 90 0 0 58 90

4 4 5 5 1 2 2

Class Class Class Class Class Class Class Class Class Class Class Class Class 3 Class Class 3 Class Example: Example: 2 for Class 1 Class 2 for 20 for Class 20 Class for utdallas.edu/~metin Page 39 0 0 32 32 7 25 25 7 -6 2 30 30 8 2 2 30 30 -6} min{3,8 0 0 32 32 operations 6,30} - max{35 Order preservingOrder 0 0 32 32 4 4 5 1 2 5 1 Protection Levels Protection Class Class Class Class Class Class Class Class Class Class Class Class 3 Class Class 3 Class 6 for Class 2 Class 6 for and and 5 Classes 0 0 38 32 25 25 13 7 2 8 30 30 3 2 30 35 Booking Limits 0 0 32 38 Request Rejected 0 0 32 38

4 4 5 5 1 2 1 2

Class Class Class Class Class Class Class Class Class Class Class Class Class Class Example: Example: 6 for Class 3 Class 6 for 3 for Class 3 Class 3 for utdallas.edu/~metin Page 40 0 26 25 1 1 25 1 25 0 26 0 26 4 1 2 5 Protection Levels Protection Class Class fare classes-fare are closed. Class Class Class Class 1 for Class 3 Class 1 for and and All low All class. first the only on accept now From bookings. accepted by bookings accumulating Recover preserve to operations preserving order make must We Why to not keep These are cumbersome. . nestedness and bookings? constant limits increase See booking the body. main in the algorithm 5 Classes 0 0 26 25 -6 25 25 0 7 1 0 25 25 -6} min{2,7 6,25} - max{30 0 1 25 25 } Booking Limits 1 0 0 26 25 min{2, 6,25} - max{30 0 0 26 25 4 5 1 2

4 5 1 2

Class Class Class Class Class Class Class Class Class Class Class Class Class Class Class 3 Class

Class 3 Class Example: Example: