1、IDeaS V5 ForecastingIDeaS V5 Forecasting-IDeaS V5 ForecastingAgenda Revenue Management Cycle Forecasting Volume Transient (True Demand) Group Patterns No-Show & Wash RateIDeaS V5 ForecastingWhat We Need to Know Last Room Value What is it? True Demand What is it?IDeaS V5 ForecastingThe IDeaS V5 Syste
2、m Is a tool that is equipped to make decisions on your behalf It uses. Data from your Reservation System Inputs from the Revenue Management TeamShare information about your business with the system -e.g. Special Events, Rate Configuration - so that the decisions IDeaS makes are in line with your bus
3、iness and business objectivesIDeaS V5 ForecastingRevenue Optimization Overview Successful Revenue Optimization has several components: Revenue Optimization Culture Alignment between the departments in the hotel setting strategies (sales and marketing) and the hotel operations (reservations and front
4、 desk) Revenue Optimization Tools IDeaS V5 Suitable Controls in the reservations systems and channels that are used to sell rooms IDeaS V5 ForecastingThe Revenue Optimization CycleForecastMonitorOptimizeControl$ Profits $ Revenues $User InteractionDataIDeaS V5 ForecastingThe Revenue Optimization Cyc
5、le When working with Revenue Optimization, the four main components of the Revenue Optimization Cycle must be accounted for The Cycle provides a framework to structure your work around ForecastMonitorOptimizeControl$ Profits $ Revenues $IDeaS V5 ForecastingForecastingTo produce optimal revenue Contr
6、ols, a Forecast of the volumes and values of each type of demand is required for OptimizationIDeaS V5 ForecastingThe Revenue Optimization Cycle$ Profits $ Revenues $VolumesValuesForecastMonitorOptimizeControl$ Profits $ Revenues $No Show & WashIDeaS V5 ForecastingForecasting Lets look at forecasting
7、, the first step on the road to maximizing revenueIDeaS V5 ForecastingWhy Forecast Demand? IDeaS V5 Controls require the calculation of the Last Room Value in order to ensure that optimal decisions can be applied The forecasting on which they are based must include reliable estimates of demand at al
8、l points along the booking curve by arrival date, segment, rate and length of stayIDeaS V5 ForecastingDemand ForecastingVolumeIDeaS V5 ForecastingDemand ForecastingDemand forecasting must: Be derived from accurately measured historical demand Perform at a sufficiently detailed level to allow for rev
9、enue management decisions and controls Involve separate analysis of any periods of abnormal demand Involve separate analysis of the risk of cancellations and “no-shows”IDeaS V5 ForecastingDemand ForecastingDemand forecasting must: Accurately measure the uncertainty associated with each forecast Ensu
10、re that all reliable historical patterns are correctly identified, and Ensure that all historical data points: Can be explained by identified patterns; or Are associated with identified special events; or otherwise Constitute truly random “noise” or volatility Normal Patterns - Keep Special Events -
11、 Keep Noise DiscardIDeaS V5 ForecastingDemand Forecasting Demand Forecasting must be handled differently for: Transient Business Group BusinessIDeaS V5 ForecastingDemand Forecasting: Transient vs Group Transient Demand Capture History Volumes Segmentation Patterns Estimate True Demand from Historica
12、l Observed Demand Forecast True Demand for Future Group Demand Capture History Volumes Segmentation Patterns Consolidate Historical Observed Demand Forecast and Distribute Consolidated Demand for the FutureWell be covering these both in detail in the following slidesIDeaS V5 ForecastingTransient Dem
13、and ForecastingTo Forecast Transient Occupancy, IDeaS Must Determine the True DemandIDeaS V5 ForecastingWhat does Transient History Look Like?Historical Demand nearly every dayClear day-of-week PatternsClear Seasonal PatternsIDeaS V5 ForecastingTransient HistoryIDeaS V5 ForecastingTransient History
14、From Transient History IDeaS: Captures Volume and Value of Business by Segment Looks for and extracts Stable PatternsIDeaS V5 ForecastingTrue DemandIDeaS V5 Forecasting2003001004005000SunMonTueWedThuFriSat0SunMonTueWedThuFriSatWhat Is True Demand?Historical Occupancy DataRoomsObserved DemandTrue Dem
15、andDemand not limited by the hotels capacityIDeaS V5 ForecastingEstimating True DemandHow to Estimate True DemandIDeaS V5 ForecastingWhat Constrains Demand? The main reasons why demand is constrained Hotel Capacity The hotel was full at the time the booking request was made Length-of-Stay Controls C
16、ontrols deployed did not allow the reservation request to be made Rate Controls The reservation request could not be made as the rate was restricted for the arrival date and length of stay The following is a conceptual example using Hotel Capacity as the constraintIDeaS V5 ForecastingEstimating True
17、 Demand IDeaS detects when rooms sold are not at or near hotel capacity IDeaS extracts patterns IDeaS detects when rooms sold are at or near hotel capacity IDeaS extracts patterns0100200300400MTWT0100200300400MTWTCapacityCapacityIn this example, a pattern relates Monday to the other nightsNot Constr
18、ainedConstrainedIDeaS V5 Forecasting0100200300400MTWTFrom the Two Patterns IDeaS extrapolates the True Demand by using patterns from dates that were not constrained Data around constrained dates serve as checks and balances to the extrapolation of True Demand0100200300400MTWT0100200300400MTWTCapacit
19、yCapacityNot ConstrainedConstrainedTrue DemandThe IDeaS System uses proprietary algorithms & techniques for determining True Demand IDeaS V5 ForecastingTransient Forecast Summary Once weve determined the True Demand for Transient business, were ready to make the forecast Lets pause and bring Group f
20、orecasts up to the same pointIDeaS V5 ForecastingForecasting Group DemandIDeaS V5 ForecastingWhat does Group History look like?Many days with no demand at allErratic day-of-week PatternsErratic Seasonal PatternsIDeaS V5 ForecastingLets look at the differencesIDeaS V5 ForecastingGroup HistoryIDeaS V5
21、 ForecastingGroup History When IDeaS Captures Group History Captures Volume and Revenue of Business By Segment Looks for and Extracts Stable PatternsIDeaS V5 ForecastingGroup vs Transient Group business characteristics differ from Transient business Patterns are not as stable or consistent There are
22、 concentrated points of significant Observed Demand separated by periods with little Observed Demand Group Demand usually has significantly greater day-to-day volatility than Transient DemandIDeaS V5 ForecastingGroup Forecast Summary Group Demand is predictable over longer periods of time The Group
23、Forecasts estimate volume over periods of time and distribute it by day of weekIDeaS V5 ForecastingForecasting PatternsWhich we keep referring toIDeaS V5 ForecastingForecasting Normal PatternsFor forecasting, IDeaS detects Normal Patterns For example data is analysed for the following: Day of Week (
24、DOW) Length of Stay (LOS) Cancellation pattern No-Show pattern In addition to the above, there are many other patterns which the system uses to analyse demandIDeaS V5 ForecastingPatternsFrom the analysis, IDeaS V5 builds a model from which forecasts are producedIDeaS V5 ForecastingDynamic PatternsWh
25、y Patterns? A Pattern is a set of associated values or numbers Examples: Booking Pattern, Length of Stay Pattern, Day of Week Pattern A Pattern can be used to predict the probability of a specific outcomeIDeaS V5 ForecastingDynamic PatternsAccuracy of Patterns Patterns are accurate when constructed
26、from similar underlying data Examples: Booking Patterns differ for leisure and business segments The accuracy of a Pattern affects the accuracy of the subsequent forecasts Patterns have a quantifiable amount of uncertaintyIDeaS V5 ForecastingModel for City Center Business HotelModel ForDay-of-Week P
27、atternLength-of-Stay PatternCancellation RateNo-show RateBooking PatternTransient CorporateTransient Leisure2.5 days18%3.5%70% 10 days priorPeaks midweek1.9 days12%2%75% 20 days priorPeaks weekendIDeaS V5 Forecasting100Length of Stay Patterns from Summary Data IDeaS uses an algorithm to extract LOS
28、information from Summary Data Example 1 LOS = 100 X 1-Night LOS Example 2 LOS = 40 X 1-Night LOS 60 X 2-Night LOS (or longer)ExArrivalsLOSDepartures#1 10010040#2?IDeaS V5 ForecastingModel Creation Normal Patterns All of this goes into creating models of how your business normally behaves These model
29、s are then used when forecasting demand for your hotelIDeaS V5 ForecastingModel Creation Are the v5i Models static or do they develop in line with market trends ? All aspects of the v5i Forecasting Calibration contains booking pattern recognition capabilities and these characteristics of your demand
30、 are updated on a regular basis Some of the characteristics that are changed in line wit your market conditions are :IDeaS V5 ForecastingModel Development Volume forecasts Booking Paces - ACFDemo_BC_1.xls DOW Patterns No Show and Cancellation Rates Group Wash Rates Demand UncertaintyIDeaS V5 Forecas
31、tingIDeaS V5 Self-Learning-10001002003004005000369121518212427303336394245485154576063666972757881DAYSROOMS SOLDFORECASTSOLDFORECAST minus SOLDV5i LEARNING CURVEIDeaS V5 ForecastingModel Creation - Special EventsWhat about Exceptional and Special Events? Statistically Significantly different: Occupa
32、ncy Patterns Segment Patterns DOW Patterns Booking Pace or Lead Time Patterns No Show/Cancellation PatternsIDeaS V5 Forecasting051015202530354045Number of Nights this OccurredNumber of Rooms SoldNormal & Special Event PatternsNumber of Rooms OccupiedSpecial or Exceptional EventsNormal Patterns300 Ro
33、om Hotel1 Market Segment60708090100110120130140150160170180190200210220230240250260270280290300IDeaS V5 ForecastingNormal & Special Event PatternsDays to ArrivalSpecial Event Booking PaceNormal Pattern Booking Pace300150001590Number of Rooms45IDeaS V5 Forecasting01020304050607080NormalSpecialNormalS
34、pecialNormalSpecialNormalSpecialRate BucketsNumber of RoomsRack Rack DiscountCorporatePackageNormal & Special Event PatternsMarket SegmentationIDeaS V5 ForecastingNormal Patterns and Special Events The hotel must define Special Events for IDeaS V5 Special Events can either be set to Impact or not Im
35、pact forecasts These Special Events provide the patterns for Special and Exceptional EventsIDeaS V5 ForecastingUncertaintyAll Demand Has UncertaintyIDeaS V5 ForecastingProbability & Uncertainty in ForecastingUncertainty in Forecasts Uncertainty plays a fundamental role in the IDeaS V5 decision-makin
36、g process, because the impact of uncertainty can be significant.IDeaS V5 ForecastingUncertainty & ForecastingWhat is uncertainty ? For an example, lets look at an everyday occurrence in the hotel to describe uncertainty -checkout time.IDeaS V5 ForecastingUncertainty & ForecastingLets say we watch a
37、front desk for checkout times . . .IDeaS V5 ForecastingUncertainty & ForecastingWe may get a data sample that looks something like this.IDeaS V5 ForecastingUncertainty & ForecastingIf we turn our clipboard sideways, it begins to look like a normal standard curve.Now lets look at that curve with more
38、 information . . . IDeaS V5 ForecastingThe mean or average checkout time is 10:00If the uncertainty is +/-one hour, we can look at one hour on either side of the average.6:007:008:0012:0013:0014:0010:00Uncertainty & ForecastingTimeNumber of RoomsWhat is uncertainty ?9:0011:0034%34%68%10:009:0011:00A
39、ll the values between 9:00 and 11:00 make up 68% of the business we tracked.Check-out Time DistributionIDeaS V5 ForecastingIf the uncertainty is +/-two hours, we can look at two hours on either side of the average.6:007:008:0012:0013:0014:0010:00Uncertainty & ForecastingTimeNumber of RoomsCheck-out
40、Time Distribution9:0011:0013%13%34%34%94%8:0012:00All the values between 8:00 and 12:00 make up 94% of the business we tracked.What is uncertainty?IDeaS V5 ForecastingIf the uncertainty is +/-three hours, we can look at three hours on either side of the average.6:007:008:0012:0013:0014:0010:00Uncert
41、ainty & ForecastingTimeNumber of RoomsWhat is uncertainty?9:0011:0013%3%13%3%34%34%99.7%7:0013:00All the business between 7:00 and 13:00 makes up 99.7% of the business we tracked.Check-out Time DistributionIDeaS V5 ForecastingLets look at three cases where the uncertainty of the demand has an impact
42、 on revenue management decisions . . .Forecasting, ContinuedImpact of Uncertainty on Revenue Management DecisionsIDeaS V5 ForecastingThe hotel has:Forecasting, Continued 300 Rooms Group Rate of $120 Unlimited group demand Transient Rate of $150IDeaS V5 ForecastingForecasting, continuedCase 1. Transi
43、ent Demand = 200 +/- 0Case 2. Transient Demand = 200 +/- 10Case 3. Transient Demand = 200 +/- 50IDeaS V5 ForecastingThe 201st room will never be soldTransient Demand200 Rooms +/- 0$0$150$10$20$30$40$50$60$70$80$90$100$110$120$130$140050100150200250300Group Demand300+ roomsGroup Rate$120 per roomTran
44、sient Demand200 +/- 0Transient Rate$150 per roomThe 200th room will always be soldBest SolutionKeep 200 Rooms for Transient DemandGive 100 Roomsto GroupsMarginal Room RevenueIDeaS V5 ForecastingForecasting, continuedCase 1. Transient Demand = 200 +/- 0Case 2. Transient Demand = 200 +/- 10Case 3. Tra
45、nsient Demand = 200 +/- 50IDeaS V5 Forecasting18017021022023019020016024034%13%3%Number of RoomsLevel of DemandWith this level of demand uncertainty, what revenue can I expectto make from the 200th room if held for the Transient Demand?50% of the time the 200th room will be sold $15050% of $150 = $7
46、5$75.00Transient Demand200 Rooms +/- 1050%The Expected Revenue from the 200th roomif held for the Transient Demand is $75.00IDeaS V5 Forecasting180170210220190200Number of Rooms230160240Level of DemandWith this level of demand uncertainty, what revenue can I expectto make from the 210th room if held
47、 for the Transient Demand?16% of the time the 210th room will be sold $15016% of $150 = $24.00Transient Demand200 Rooms +/- 1016%$75.0013%3%$24.00The Expected Revenue from the 210th roomif held for the Transient Demand is $24.00IDeaS V5 Forecasting180170210220190200Number of Rooms230160240Level of D
48、emandWith this level of demand uncertainty, what revenue can I expectto make from the 190th room if held for the Transient Demand?84% of the time the 190th room will be sold $15084% of $150 = $126$126.00Transient Demand200 Rooms +/- 1084%34%13%3%34%The Expected Revenue from the 190th roomif held for
49、 the Transient Demand is $126.00$75.00$24.00IDeaS V5 Forecasting180170210220190200Number of Rooms230160240Level of DemandWith this level of demand uncertainty, what revenue can I expectto make from the 220th room if held for the Transient Demand?3% of the time the 220th room will be sold $1503% of $
50、150 = $4.50$126.00Transient Demand200 Rooms +/- 10The Expected Revenue from the 220th roomif held for the Transient Demand is $4.50$75.003%$24.00$4.50IDeaS V5 Forecasting180170210220190200Number of Rooms230160240Level of DemandWith this level of demand uncertainty, what revenue can I expectto make f
