car park revene managment

© 2015 GrayMatter All Rights Reserved

AA+ (CPRM) Car Park Revenue Management

© 2015 GrayMatter All Rights Reserved 2

• AA+ CPRM Implement s a demand driven dynamic pricing system for all the car parks in Airport– Car park prices will be updated every day for next X Days in advance

• Dynamic pricing will result in – Increase in the car parking revenue

– Increase in occupancy levels in all the car parks managed by Airport

• Pricing decisions should adhere to business rules and constraints

• System should monitor demand and decision accuracies and recalibrate when demand predications deviate significantly from actuals– Recalibration can be a mix of automatic and manual tuning of models and

algorithms

AA+ CPRM Business Objectives


• AA+ CPRM is an extension to AA+ car park analytics framework for revenue optimization

• AA+ CPRM is built on advanced statistical models with several linear and non linear variables for occupancy predictions

• AA+ CPRM system predict and recommends future price every day for few days ahead based on historical data analytics, algorithms and business rules

• AA+ CPRM provides user interface to control and override recommendation as needed

• AA+ CPRM provides ongoing basis monitoring and remodeling of algorithms for model optimization

Introduction to AA+ CPRM


• Changing demand (supply and demand curves)– Competitor pricing and other actions

– External factors (weather, etc.)

• Demand management– Trying to make as much money from selling a limited capacity as possible

– Stimulate demand when it is low

• Hedging– Customers are rewarded with discounts for committing early

– Limit the effect of, e.g., bad weather

• Willingness to pay of different market segments– Customer differentiation and price discrimination

AA+ CPRM built on following Concepts


• A demand model is a mathematical or pattern based function of independent attributes

• Several algorithms are used to infer this function– Triple Exponential Series

– Auto Regressive Integrated Moving Average

– Seasonality & Growth Models

– Linear and Non-Linear Regression

– Pattern Based Models

• Selection of the algorithm is done as part of model validation using “accuracy” of prediction and “uncertainty” in predictions– Accuracy measures how close predicted demand is with actuals

– Uncertainty measures spread in the prediction error

• Sometime ensemble models are also deployed. Model selection is driven by the data

AA+ CPRM Model Overview


CRRM Solution Architecture Overview


Data Requirements


Data Needed

PURPOSE MODELLING DECISION MONITORING

Transactions Completed Transactions for past 24 months

Incremental Completed TransactionsandIn-Park TransactionsCars which have arrived in the carpark but not yet departed

Both Completed and In-Park Transactionsof Recent past weeks (2-4)

Reservations Completed Reservationsfor past 24 monthsCars have either arrived and departed or booking cancelled or Noshow

Incremental Completed Reservationsand On-Books ReservationsThese are reservations where arrival date is in future

Both Completed and On-Books Reservationsof Recent past weeks (2-4)


• Transaction ID• Carpark• Date and Time of Arrival• Date and Time of Departure• Type of Arrival: PreBook or DriveIn• Parking Charges

– Split charges based on components like Parking, Tax, Commissions

• Discount/Promotion Coupon, if any– Employee Discount – Free Parking

• Customer Type– Employee/Staff Vehicle– Maintenance Vehicle– General Vehicle

• Booked Date and Time of Arrival• Booked Date and Time of Departure• Booked parking rate

Completed Transaction Data


• Transaction ID

• Carpark

• Date and Time of Arrival

• Type of Arrival: PreBook or DriveIn

• Prorated Parking Charges– Split charges based on components like Parking, Tax, Commissions

• Discount/Promotion Coupon, if any– Employee Discount

– Free Parking

• Customer Type– Employee/Staff Vehicle

– Maintenance Vehicle

– General Vehicle

• Booked Date and Time of Arrival

• Booked Date and Time of Departure

• Booked parking rate

In-Park Transaction Data


• Reservation ID• Carpark• Date and Time of Reservation• Date and Time of Cancellation• Planned Date and Time of Arrival• Planned Date and Time of Departure• Discount/Promotion Coupon, if any

– Employee Discount – Free Parking

• Customer Type– Employee/Staff Vehicle– Maintenance Vehicle– General Vehicle

• Number of parking slots needed• Parking Charges, split if any• Status

– No Show, Used as planned, Early Exit, Early Arrival, Delayed Departure, Delayed Arrival

Completed Reservation Data


• Reservation ID

• Carpark

• Date and Time of Reservation

• Planned Date and Time of Arrival

• Planned Date and Time of Departure

• Discount/Promotion Coupon, if any– Employee Discount

– Free Parking

• Customer Type– Employee/Staff Vehicle

– Maintenance Vehicle

– General Vehicle

• Number of parking slots needed

• Parking Charges, split if any

On-Books Reservation Data


• In addition to the transaction and reservation data, we will also need carpark meta data which includes– Carp park name

– Type of carpark

• Multilevel

• Valet/Drop off

• Park & Ride

– Capacity

– Number of hours of stay classified as Short Stay

• This data will be captured once and an interface will be provided to add/delete/modify information

Meta Data


High Level Solution Approach


High Level Solution ApproachA

A+

War

eho

use

Res

erva

tio

ns

Historical DataExtraction Descriptive

AnalyticsData

CleaningModelling

Use

r In

terf

ace

and

Das

hb

oar

ds

IncrementalData - Daily

MODEL

Prediction Optimization

Pre

dic

tio

n

&

Dec

isio

ns

Decision upload to reservation system(s)

Are Predictions

and Decisions “Good”

Recalibrate Model

No Yes

BusinessRules

Modelling, Daily Decisions, Monitoring

Patterns –booking / Arrival

Special Events

ConstrainingOutliers


• Three stage solution

• Historical data processing– Uses past 2+ years of historical transactions and reservations

– Data cleaning – outlier detection, identifying constrained periods & special events

– Customize demand prediction model for ADM

• Daily Decision Optimization– Uses recent transactions & reservations

– Using business constraints, predictions generates pricing decisions

• Monitoring Process– Monitors patterns and predictions in a predefined frequency

– Significant deviations results in alerts and re-modelling activities

High Level Solution Approach


• Data cleaning and imputation of missing data

• Identify outliers and associate reasons with outliers– Low occupancy with maintenance of car park

– Low occupancy with airport closure due to weather

– High occupancy with holidays

• Extract of useful insights/ patterns from the Car Park data, like– Special Events, Price Elasticity, Seasonality, Cancellation rates

Historical Data Analysis


• Car park demand is a function of several attributes like,– Seasonality (Week of Year (WOY), Day of Week (DOW), Special Events- Holidays, Extended

Weekends, Conference, ...– Length Of Stay (LOS)– Type of carpark, Price– PAX volume , Competition

• Using statistical testing techniques, extent of the impact of each of the attributes on the demand is inferred– Time Series– Pattern based– Regression

• Historical price changes are used to extract price elasticity and cross elasticity functions

• In order to reduce uncertainty in the demand predictions, demand groups having similar demand functions are made– Clustering algorithms are used to find these demand groups

• Demand models are build for each demand group. Group demand is the distributed to lower granular element

Prediction – Car park Demand Model


• Dynamic pricing comes in two scenarios– Demand exceeds the carpark capacity

• Optimizer should not allow lower rates/discounts to be offered

– Demands are very low compared to capacity

• Optimizer should attract more demand by reducing the rates/offering more discount

• Optimizer will scan price and demand space and select a price for each carpark & LOS to achieve the objectives within the business constraints

• Objectives for the optimizer can be to maximize revenues and occupancy levels across all the carparks

Car Park Optimizer


• Quality of decisions depends on the forecasts which in turn is a function of various patterns and models

• System will monitor Occupancy Levels, Revenue per slot and predictions v/s actual arrivals by LOS on a weekly basis

• Alerts are raised when monitored metrics crosses a threshold

• Pattern monitoring and model re-validations are carried out once a quarter

Monitoring


• User Interface to evaluate the decision provided by system– Workflow approval process to override any decision

• Configuration master data management, capture competition and other external data

• Business Rule Configuration UI

• System Monitoring UI

User Interfaces


• Data Connector to upload data from decision database to car park reservation system

Connector to reservation Systems


Thank You

car park revene managment

Documents