Inventory Network Management System Revenue Shares - PowerPoint

Document Sample
Inventory Network Management System Revenue Shares - PowerPoint Powered By Docstoc
					Integration of Pricing and Revenue
Management for a Future without
Booking Classes
AGIFORS
Reservation and Yield Management Study Group
Annual Meeting, Bangkok, May 8 - 11, 2001

Natascha Jung      Senior Operations Research Specialist
Klaus Weber        Senior Scientific Analyst
 Presentation
 on behalf of my colleague Natascha




Natascha
                                                                             ...
                                                                             and her
                                                                             boss




Chart 2        AGIFORS
               Reservation and Yield Management Study Group Annual Meeting
               May 8 - 11, 2001
Agenda


       Motivation



       Booking Classes



       Pricing & Revenue Management



       Integrated Approach to Pricing & Revenue Management



       Outlook / Conclusions



Chart 3              AGIFORS
                     Reservation and Yield Management Study Group Annual Meeting
                     May 8 - 11, 2001
Motivation
The Customer

         Typical Travel Agency Situation?
                                                                                 Do you have any
                                                                                     I see, ...
                           I wouldNo, ... book
                                      like to                                  idea when and for
                                                                               Are you lookinghow
                                     That’s
                                    would prefer a
                           Yes,aI flight from
                                not important!                                 a you would like to
                                                                                 last minute offer?
                             three leg itinerary
                                 Bangkok to                                             fly?
                             I just like to travel
                             via Frankfurt and
                              Rio de Janeiro.
                                in M class ...
                                    Madrid.
                           I mean, from BKK to
                            FRA in M, and then
                              H class to MAD.
                            But for transatlantic
                                flights I prefer
                                   class Q!!

                                                                                                  travel agent



           customer
Chart 4                AGIFORS
                       Reservation and Yield Management Study Group Annual Meeting
                       May 8 - 11, 2001
Motivation
The Customer (cont.)

         Typical Travel Agency Situation?
                                                                           If you don’t mind
                                                                              understand.
                                                                        DoI you have any
                      I would like to book                                    whenrisk a
                                                                      idea You need of
                                                                                the and how
                        I am going to take
                                                                                       N
                                                                         comfortable seat
                                                                        you would like to
                            a inOh, no!
                        part flight from
                                  the annual                                 thrombosis
                                                                        with power supply
                                                                                  fly?
                            Bangkok Nobel
                         meeting of to and
                         I just like to sit
                                   Great!
                         Rio de Janeiro.
                          think about a new                                  I your laptop.
                                                                          for can offer you
                          laureates which
                                 I take it!
                                                                          a discounted fare
                         startsformula.
                                 next month
                                                                                 in Eco?
                      and will take 8 days.




                                                                                               travel agent



           customer
Chart 5                  AGIFORS
                         Reservation and Yield Management Study Group Annual Meeting
                         May 8 - 11, 2001
Motivation
The Customer (cont.)

         Resume
              Customer does not care about booking classes!
              Customer demand depends on
                    Travel day
                    Prices (offered at ticketing day)
                                                                                        fares
                    Service
                    Restrictions (minimum stay, day application, ...)

           Demand is controlled by fares, i.e. customer needs!

         Revenue management systems
              Forecast is estimated on basis of historical demand
              Booking histories are stored w.r.t. booking class!

!!! Contradiction !!!

Chart 6                   AGIFORS
                          Reservation and Yield Management Study Group Annual Meeting
                          May 8 - 11, 2001
Agenda


       Motivation



       Booking Classes



       Pricing & Revenue Management



       Integrated Approach to Pricing & Revenue Management



       Outlook / Conclusions



Chart 7              AGIFORS
                     Reservation and Yield Management Study Group Annual Meeting
                     May 8 - 11, 2001
Booking Classes
Historical Review

          Historical Development of Revenue Management at Lufthansa

              1955                  Central Reservation


              1965                  Overbooking


              1972                  Currency Control                       1972 Littlewood’s rule


              1977                  Segment Pricing                                Booking classes


              1989                  Fare Mix / Control                               1989 EMSRb


              1994                  Network Management / Bid Pricing

Chart 8              AGIFORS
                     Reservation and Yield Management Study Group Annual Meeting
                     May 8 - 11, 2001
      Booking Classes
      Why Booking Classes?

   Revenue management point of view
          Booking classes reflect
           ‘somehow’ the customers’
           will to pay
          booking classes are the basis for
           demand forecast
          Reservation systems use them to
           indicate availability

However, there are other means to
          reflect customer will to pay (fares)
          compute customer demand (customer choice model)

         We do not truly need booking classes but we are used to use them.

      Chart 9            AGIFORS
                         Reservation and Yield Management Study Group Annual Meeting
                         May 8 - 11, 2001
                Booking Classes
                Why Booking Classes? (cont.)


                       Pricing point of view
                                  Booking classes are not important at all
                                  Only fares are important!
                                  Fares are richer in information than booking classes


                                                                                Ticketing   Ticketing   Travel    Travel    Minimum   Maximum    Advance     Day
Origin   Destination   Carrier   Booking Cl.   Fare Basis Amount Currency        from        to         from      to        Stay      Stay       Purchase    Application
BER      BKK           XY        C             CRT          5990 DEM            01JAN01     open        01JAN01   open      -         -          -           -
BER      BKK           XX        C             CRT          5650 DEM            01JAN01     open        01JAN01   open      -         -          -           -
BER      BKK           XY        U             ULXAP7DE     1550 DEM            15MAR01     20JUN00     15APR01   30JUN01   SU              30           7         1234
BER      BKK           XX        V             VLXAP7DE     1500 DEM            20MAR01     25JUN00     10APR01   15JUN01   SU              30           7         1234
BER      BKK           XY        U             ULWAP7DE     1750 DEM            15MAR01     20JUN00     15APR01   30JUN01   SU              30           7           567
BER      BKK           XX        V             VLWAP7DE     1650 DEM            20MAR01     25JUN00     10APR01   15JUN01   SU              30           7           567




               Chart 10                            AGIFORS
                                                   Reservation and Yield Management Study Group Annual Meeting
                                                   May 8 - 11, 2001
               Booking Classes
               Problems with Booking Classes

                      Booking classes are inaccurate, e.g.
                                                                                Ticketing   Ticketing   Travel    Travel    Minimum   Maximum    Advance     Day
Origin   Destination   Carrier   Booking Cl.   Fare Basis Amount Currency        from        to         from      to        Stay      Stay       Purchase    Application
BER      BKK           XY        U             ULXAP7DE     1550 DEM            15MAR01     20JUN00     15APR01   30JUN01   SU              30           7          1234
BER      BKK           XY        U             UDUMMY       4990 DEM            01JAN01     open        01JAN01   open      -         -          -           -
BER      BKK           XX        V             VLXAP7DE     1550 DEM            15MAR01     20JUN00     15APR01   30JUN01   SU              30           7          1234
BER      BKK           XX        V             VDUMMY       4750 DEM            01JAN01     open        01JAN01   open      -         -          -           -


                       Quite different fares in same booking class
                            Distortion of average revenue of booking class




                      Another indicator
                                 Booking class harmonisation in airline alliances usually
                                  increases number of booking classes


               Chart 11                            AGIFORS
                                                   Reservation and Yield Management Study Group Annual Meeting
                                                   May 8 - 11, 2001
Booking Classes
Problems with Booking Classes (cont.)

      RM systems forecast on historical data
 !!! Meanwhile fares might have been changed !!!



              Different demand

              Different average values of corresponding booking classes




Chart 12               AGIFORS
                       Reservation and Yield Management Study Group Annual Meeting
                       May 8 - 11, 2001
Booking Classes
Booking Classes - pros and cons

            Booking classes                                                  Booking classes
                 pros                                                              cons
     History building for demand                           Bid price control requires
      forecast                                               calculation of bid prices
      (generally possible for fares                         Booking class protects
      as well, however: too many                             not necessary
      fares)                                                Booking classes only indirectly
     Booking class corresponds                              used for BP calculation
      to class value (‘somehow’)
     Booking class can be identified
      with restrictions (‘somehow’)                         Inexactness of booking classes
                                                            restricts opportunity of
     ‘approximation’
                                                             additional revenue gain
     inexact, but works
                                                             ... in 0.1 % range (which are
                                                             millions of € or $)

Chart 13            AGIFORS
                    Reservation and Yield Management Study Group Annual Meeting
                    May 8 - 11, 2001
Booking Classes
Booking Classes - pros and cons (cont.)

           Booking classes                                                 Booking classes
                pros                                                             cons
                                                              ‘customer pricing’
                                                               (priceline.com, tallyman.de)
                                                               shows
                                                          Booking classes are not
                                                           necessary

                                                         !!! Approach not generally
                                                             applicable to airlines !!!
                                                              restrictions (not refundable) are
                                                               necessary to avoid down selling
                                                              not possible to sell ‘open tickets’



Chart 14          AGIFORS
                  Reservation and Yield Management Study Group Annual Meeting
                  May 8 - 11, 2001
Agenda


      Motivation



      Booking Classes



      Pricing & Revenue Management



      Integrated Approach to Pricing & Revenue Management



      Outlook / Conclusions



Chart 15            AGIFORS
                    Reservation and Yield Management Study Group Annual Meeting
                    May 8 - 11, 2001
Pricing & Revenue Management
Overview: Pricing System - Price Elasticity Model


  Trigger      Price Elasticity Model                                                       Output

                                                      Fare
                                                     Analyser




                  Competitor                                                    Traveller
                   Reaction                   Market Share                     Preference    Demand
       Fare                                                                     Analyser
                   Analyser                      Model                                         Shift
      Change



                                                  Market
                                                Stimulation
                                                   Model




Chart 16         AGIFORS
                 Reservation and Yield Management Study Group Annual Meeting
                 May 8 - 11, 2001
Pricing & Revenue Management
Overview: Revenue Management System


                   O&D Forecast Engine                                                      GDS
           Forecast Building


                           Forecast                                                         GDS
                         Interface DB




                                                                                Inventory




                                                                                            ...
      Demand Forecasts

                       O&D Optimiser
                                                                                            GDS
                                                     Control
                                                   Parameters
                           Optimiser
                         Interface DB                                                       GDS

Chart 17                 AGIFORS
                         Reservation and Yield Management Study Group Annual Meeting
                         May 8 - 11, 2001
Pricing & Revenue Management
Comparison

               Pricing                                            Revenue Management
                                 Target level
                          (What shall be controlled?)

     Market - OD                                                 ODIF POS
      origin destination                                           origin destination itinerary fare
      Fares given on this level.                                   class point of sale
                                                                   Booking values given on this level.
 customer oriented                                           Value / revenue oriented
 coarse level                                                fine level
                            Target quantities
                    (Which quantities are focused on?)
     Market size                                                 ODIF POS demand
     Market share                                                ODIF POS values
     Fares                                                       marginal seats revenues
                                                                  bid prices

Chart 18              AGIFORS
                      Reservation and Yield Management Study Group Annual Meeting
                      May 8 - 11, 2001
          Pricing & Revenue Management
          Comparison (cont.)

                                    Pricing                            Revenue Management
                                                       Target quantities
                                               (Which quantities are focused on?)

               Market size                                                                    ODIF POS demand
               Market share                                                                   ODIF POS values
               Fares                                                                          marginal seats revenues
                                                                                               bid prices

Trigger        Price Elasticity Model                              Output
               Linking both spheres means
                                         Fare
                                        Analyser

                                 replace booking classes in RM sphere by fares (properly)
                     Competitor                        Traveller
   Price
  Change
                       
                      Reaction
                      Analyser
                                  integrate both spheres
                                    Market Share
                                       Model
                                                    Demand
                                                      Preference
                                                       Analyser
                                                      Shift



                                       Market
                                     Stimulation
                                        Model




          Chart 19                                 AGIFORS
                                                   Reservation and Yield Management Study Group Annual Meeting
                                                   May 8 - 11, 2001
Pricing & Revenue Management
Requirements for Integration

           Current Demand Forecast

Booking histories                                             Current booking level

     Booking classes                                              Booking classes
             PNR




                                                                                     PNR
 ”COPY PNR_ - RT IN AMADEUS FOR MODIFICATION.
      1.WEBER/KLAUSMR(001) 2.JUNG/NATASCHAMRS(001)
 017 X3.SSR OTHS LH PLS DAPO TO CNFM LH 729 24 OR 25 APR
        (017)
 ...
     LH 721 H 24APR01PEKFRA XL2   1025 1425
        NN(009) LL(009) HL(009) PE(009) HL(009) XL(010)
     LH3002 H 24APR01FRAHEL XL2   1720 2050
        NN(009) LL(009) HL(009) PE(009) HL(009) XL(010)
 011 05APR01 0906 998 MUCRM1A /05090651/SHALH0100//SHA/LH/A/CN//GS/SHAGXLH
 ...
 015 10APR01 0215 998 MUCRM1A /10021609/SHALH0100//SHA/LH/A/CN//GS/SHAGXLH
     LH 729 H 10APR01PVGFRA XX2   1110 1625
        NN(004) LL(004) HK(004) LK(004) HK(004) XX(015)
     LH3002 H 10APR01FRAHEL XX2   1720 2050
        NN(004) LL(004) HK(004) LK(004) HK(004) XX(015)

Chart 20               AGIFORS
                       Reservation and Yield Management Study Group Annual Meeting
                       May 8 - 11, 2001
Pricing & Revenue Management
Requirements for Integration

               Integrated Demand Forecast

Booking histories                                                  Current booking level

  Fare buckets                                                     Fare buckets
= fares clustered with respect to                                  ??? Source ???
       value / revenue
       restrictions                                                                             In the future?
”COPY PNR_ - RT IN AMADEUS FOR MODIFICATION.




                                                                                          fare
           1.WEBER/KLAUSMR(001) 2.JUNG/NATASCHAMRS(001)                                          ... together with
 Why buckets? DAPO TO CNFM LH 729 24 OR 25 APR




                                                                                                                     (inexact)
017 X3.SSR OTHS LH PLS




                                                                                                               fare
            (017)                                                                                other usefull,
 ...          Optimisation requires
     LH 721 H 24APR01PEKFRA XL2   1025 1425                                                      quantities,
           competition among
        NN(009) LL(009) HL(009) PE(009) HL(009) XL(010)
     LH3002 H 24APR01FRAHEL XL2
                                                                                                 e.g. bid prices
           customers for same1720 2050
                                   seat
        NN(009) LL(009) HL(009) PE(009) HL(009) XL(010)
 011 05APR01 0906not take fares, i.e.
 ...
        Why 998 MUCRM1A /05090651/SHALH0100//SHA/LH/A/CN//GS/SHAGXLH
           one fare per bucket?
                                                                                          class info          Pricing
 015 10APR01 0215 998 MUCRM1A /10021609/SHALH0100//SHA/LH/A/CN//GS/SHAGXLH
     LH Too many buckets 1110 1625
        729 H 10APR01PVGFRA XX2                                                                                 DB
        NN(004) LL(004) HK(004) LK(004) HK(004) XX(015)
     LH3002 H 10APR01FRAHEL XX2   1720 2050
                                                                             restrictions etc.
        NN(004) LL(004) HK(004) LK(004) HK(004) XX(015)

Chart 21                    AGIFORS
                            Reservation and Yield Management Study Group Annual Meeting
                            May 8 - 11, 2001
Pricing & Revenue Management
Requirements for Integration (cont.)


  Bookings                                                                              Fare buckets
   bkg     fare           amount                               value /
   no.                                                         revenue                    av. value
   1       CRT            5990                                 2515                      no. of bkgs.
                                                                                          Example
   2       CRT            5990                                 2515
   3       VLXAP7DE       1500                                 670                     Value value
                                                                                           av.
   4       CRT            5990                                 2515                      no. 1100
                                                                                        900 -of bkgs.
   5       VLWAP7DE       1650                                 705                     restriction
   6       VLWAP7DE       1650                                                          sunday return
                                                               705
                                                                                       advanced purchase
                                                                                           av. value
   ...     ...            ...                                  ...                      14 daysbkgs.
                                                                                          no. of




                                                                                            ...
           For optimisation we need to forecast the number of bookings in
           each fare bucket.
Chart 22              AGIFORS
                      Reservation and Yield Management Study Group Annual Meeting
                      May 8 - 11, 2001
Agenda


      Motivation



      Booking Classes



      Pricing & Revenue Management



      Integrated Approach to Pricing & Revenue Management



      Outlook / Conclusions



Chart 23            AGIFORS
                    Reservation and Yield Management Study Group Annual Meeting
                    May 8 - 11, 2001
Integrated Approach Pricing & RM
Price Elasticity Model

Forecast of demand for each fare bucket - step 1

           Data    Price Elasticity Model                                                          Utility
           sources                                                                                 factors
                                              Alternative
                                               choice
                                      Customer Analyser model
             Internal
            fare data                 multinomial logit model
                                           customer’s choice                                        æ b1 ö
                        Competitor                                                     Traveller     ç ÷
                                                                                                     ç b2 ÷
                         Reaction           depends on
                                                Market Share                          Preference
                         Analyser              amount
                                                   Model                               Analyser
                                                                                                     ç M ÷
             ATPCO
                                               minimum stay                                         ç ÷
                                                                                                     çb ÷
                                               advanced                                             è nø
                                                 purchase
                                                   Market
                                               ...
             MIDT                                     Stimulation
                                                         Model


                                  Separate for business and leisure passengers
Chart 24                AGIFORS
                        Reservation and Yield Management Study Group Annual Meeting
                        May 8 - 11, 2001
Integrated Approach Pricing & RM
Market Size Forecaster

 Forecast of demand for each fare bucket - step 2

           Data                        Forecast                                       Demand
           sources                     basis                                          forecast

             Internal
            fare data
                                           bkg    fare
                                           no.
                                           1      CRT
                                           2      CRT
                                           3      VLXAP7DE
            ATPCO                          4      CRT
                                           5      VLWAP7DE
                                           6      VLWAP7DE
                                           ...    ...                                  Market Size
                                                                                       Forecaster

             MIDT
                                              Current
                                             bookings



Chart 25                AGIFORS
                        Reservation and Yield Management Study Group Annual Meeting
                        May 8 - 11, 2001
         Integrated Approach Pricing & RM
         Fare Buckets Forecast
Forecast of demand for each fare basket - step 3
                       fare               market
 Price Elasticity                         share
                       fare1                         Per ticketing/travel
 Model                                    1
                       fare2                         combination                      av. value
                                          2
                       fare3              3                                         no. of bkgs.
(b1     b2 K bn )      fare4              4
                       fare5              5
                       ...                ...
                                                                                      av. value
                       faren              n                                         no. of bkgs.

                       itinerary forecast forecast
                                 value     error
                                         Per departure                                av. value
                       ODI1      fc_val1   fc_err1
                                         date                                        no. of bkgs.
                       ODI2      fc_val2   fc_err2
      Market Size      ODI3      fc_val3   fc_err3
      Forecaster       ODI4      fc_val4   fc_err4




                                                                                        ...
                       ODI5      fc_val5   fc_err5
                       ...       ...       ...
                       ODIn      fc_valn   fc_errn
                                                                                     Fare buckets

         Chart 26      AGIFORS
                       Reservation and Yield Management Study Group Annual Meeting
                       May 8 - 11, 2001
   Integrated Approach Pricing & RM
   System Overview

  Market Shares                   Market Size Forecast                                                 GDS
Market Size
                                  Price Elasticity Model

                                                                        Bookings                       GDS
                                                                        with Fare
                Forecast                                               Information




                                                                                           Inventory
              Interface DB




                                                                                                       ...
                      Demand Forecasts

        O&D Optimiser                                                                                  GDS

                                                                   Control
                                                                 Parameters
                                   Optimiser
                                                                                                       GDS
                                 Interface DB

   Chart 27                  AGIFORS
                             Reservation and Yield Management Study Group Annual Meeting
                             May 8 - 11, 2001
Integrated Approach Pricing & RM
Demand Changes

           Re-Optimisation

           Previous  New
                                                             Previous Forecast
              Forecast


                                                                                         Current Demand


                                              Learning
            Price Elasticity
                Model




                          Floating dcp                                       Snap shot


Chart 28               AGIFORS
                       Reservation and Yield Management Study Group Annual Meeting
                       May 8 - 11, 2001
Agenda


      Motivation



      Booking Classes



      Pricing & Revenue Management



      Integrated Approach to Pricing & Revenue Management



      Outlook / Conclusions



Chart 29            AGIFORS
                    Reservation and Yield Management Study Group Annual Meeting
                    May 8 - 11, 2001
Outlook / Conclusions
Outlook - Mining PNRs
     Business passengers behave different from leisure passengers
             Separate customer choice models

Problem: Identification of passenger types

     history data                            Fare information available (insufficient?)

     current booking data                            !!           PNR may not contain fare information!

             Mine PNR data to decide to which type passenger belongs to
”COPY PNR_ - RT IN AMADEUS FOR MODIFICATION.
     1.WEBER/KLAUSMR(001) 2.JUNG/NATASCHAMRS(001)
017 X3.SSR OTHS LH PLS DAPO TO CNFM LH 729 24 OR 25 APR
                                                                                                  e.g.
...
       (017)
                                                                                                   pax travels single
    LH 721 H 24APR01PEKFRA XL2   1025 1425
       NN(009) LL(009) HL(009) PE(009) HL(009) XL(010)
                                                                                                     ... or with family
    LH3002 H 24APR01FRAHEL XL2   1720 2050
       NN(009) LL(009) HL(009) PE(009) HL(009) XL(010)
011 05APR01 0906 998 MUCRM1A /05090651/SHALH0100//SHA/LH/A/CN//GS/SHAGXLH
...
                                                                                  PNR              pax is frequent
015 10APR01 0215 998 MUCRM1A /10021609/SHALH0100//SHA/LH/A/CN//GS/SHAGXLH
    LH 729 H 10APR01PVGFRA XX2   1110 1625
       NN(004) LL(004) HK(004) LK(004) HK(004) XX(015)
                                                                                                     traveller
    LH3002 H 10APR01FRAHEL XX2   1720 2050
       NN(004) LL(004) HK(004) LK(004) HK(004) XX(015)                                             ...

Chart 30                            AGIFORS
                                    Reservation and Yield Management Study Group Annual Meeting
                                    May 8 - 11, 2001
Outlook / Conclusions
Summary

     Booking classes
              are inaccurate w.r.t. amount of fares contained
              are not used directly in O&D bid price control

     Booking class-based forecast
              reflects historical customer choice only
              does not concern current market changes

     Booking class-oriented revenue management systems
              are sophisticated (O&D control)
              work well
              may not exhaust the full potential of RM control!!!
           There is need for strong integration of Pricing and RM

Chart 31                AGIFORS
                        Reservation and Yield Management Study Group Annual Meeting
                        May 8 - 11, 2001
Outlook / Conclusions
Summary
     Gap between pricing and revenue management could be closed
              by replacement of booking classes by fare buckets
              by forecast concerning
                   market size
                   market shares

      Integrated approach to pricing and revenue management
         manages demand changes
     Challenges
         Inventory changes
                                                     Interested airlines are
         Information provision
                                                        kindly invited for
     Good experience with                            proving the concept
         pricing              NetLine/Price
         pricing simulation   NetLine/Price:PSM
         pax demand forecast
           for scheduling      NetLine/Plan
Chart 32                  AGIFORS
                          Reservation and Yield Management Study Group Annual Meeting
                          May 8 - 11, 2001
Thank you for your attention!
      Any questions?

				
DOCUMENT INFO
Description: Inventory Network Management System Revenue Shares document sample