Smart Airlines Baggage Management and Improving Customer Experience

Smart Airlines Baggage Management and improving Customer Experience

Carla Romano Director - Industry Big Data, IoT Soluon Oracle

Paresh Sansare Co-CEO / SVP Sales and Projects eBIW

Feb 3, 2017

Copyright © 2017, Oracle and/or its aﬃliates. All rights reserved. | Safe Harbor Statement The following is intended to outline our general product direcon. It is intended for informaon purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or funconality, and should not be relied upon in making purchasing decisions. The development, release, and ming of any features or funconality described for Oracle’s products remains at the sole discreon of Oracle.

1 Baggage Handling Processes and Costs

2 Data Consolidaon

3 IOT, Streaming and Smart Tracking

4 Maximizing Customer Sasfacon

1 Baggage Handling Processes and Costs

2 Data Consolidaon

3 IOT, Streaming and Smart Tracking

4 Maximizing Customer Sasfacon

Chicago Airport Lines this summer, up to 3 hours wait me, to board a domesc ﬂight!

Checked bags. Mishandled (delayed, damaged, or lost) In 2015 * Source SITA

Costs to Airlines Break-down of the mishandled bags

NORMAL NORMAL HANDLING OPERATING

cost / bag = $10 cost / passenger = $185

DELAYED BAG LOST BAG

cost / bag = $100 cost / bag = up to $3300

Baggage Related Sasfacon • Among the two-thirds of passengers who check baggage for their ﬂight, 52% indicate they had to wait 15 minutes or longer to receive their baggage, among whom sasfacon is 711, compared with 751 among those who experience a shorter wait me. • • Sasfacon among passengers who pay for checked baggage has improved steadily during the past ﬁve years to 700 in 2015 from 637 in 2011. • The above data implies that other expanded services can be sold to passengers around baggage, provided it reduces/eliminates wait and has lower delays/damages to the bags.

hp://www.jdpower.com/sites/default/ﬁles/2015057%20NA%20Airline_%20(FINAL).pdf

How Airline Baggage Handling Systems Work

• A baggage-handling system has three main jobs: - Move bags from the check-in area to the departure gate - Move bags from one gate to another during transfers - Move bags from the arrival gate to the baggage-claim area • The measure of a successful baggage-handling system is simple: Can the bags move from point to point as fast as the travelers can? • If the bags move slower, you'll have frustrated travelers waing for bags, or bags failing to make connecng ﬂights on me. • If the bags move too fast, you might have bags making connecng ﬂights that passengers miss (is that important?)

hp://science.howstuﬀworks.com/transport/ﬂight/modern/baggage-handling.htm

Transparency is #1 Aviaon Impediment Data Silos, Situaonal Awareness, Interoperability

IoT can provide shared, real me awareness & coordinaon Historical Data Silos have made for a Dysfunconal Ecosystem

Improvement happens at the Interface between Ecosystem ? Partners

$ AIRCRAFT DATA TECH OPS CREW DATA PASSENGERS AIRPORT DATA

Passengers and Baggage Passenger tracking via PNR, split at Check-In and API and DCS re-join at Baggage Claim Baggage tracking via PNR

New and BRS Transfer Flight Direct

Baggage Customs (Security check, baggage Baggage Arrivals sorng, etc.) Claim Hall

PNR = Passenger Name Record API = Advance Passenger Informaon DCS = Departure Control System BRS = Baggage Reconciliaon System

Parcularly when custody changes

New Transfer Flight Direct

Baggage Customs (Security check, baggage Baggage Arrivals sorng, etc.) Claim Hall

PNR = Passenger Name Record API = Advance Passenger Informaon DCS = Departure Control System BRS = Baggage Reconciliaon System

Copyright © 2017, Oracle and/or its affiliates. All rights reserved. | Baggage Flow and Mis-handling Baggage (Security check, baggage sorng, etc.) Departure Hall Check-in Passenger Direct (Security check, immigraon check, Transfer etc.) Mulple points of failure Resoluon 753 in baggage handling across travel routes • Prevent and reduce mishandling by determining custody of every bag during different phases of the baggage chain • Demonstrate delivery when custody changes Parcularly when custody • Demonstrate acquision when custody changes changes • Reduce the possibility of baggage fraud • Enable excepons to be detected • Provide an accurate inventory of bags on flight departure New • Transfer Speed up reconciliaon and flight readiness Flight Direct • Help measure compliance to SLAs • Exchange events with other airlines as needed • Increase customer sasfacon as mishandling is reduced Baggage Customs (Security check, baggage Baggage Arrivals sorng, etc.) Claim Hall

PNR = Passenger Name Record API = Advance Passenger Informaon DCS = Departure Control System BRS = Baggage Reconciliaon System

1 Baggage Handling Processes and Costs

2 Data Consolidaon

3 IOT, Streaming and Smart Tracking

4 Maximizing Customer Sasfacon

Copyright © 2017, Oracle and/or its affiliates. All rights reserved. | Confidenal – Oracle Internal/Restricted/Highly Restricted 15 Oracle Airline Data Model More Than Just a Data Model • Industry-standard compliant based Enterprise-wide Data Model Oracle Airline Data Model • Contains Logical and Physical Data Models Third Normal Atomic, Dimensional Schema • Industry specific Airlines Measures and KPI • Automac Data Movement Among Layers • Extensive business intelligence metadata • Derived Tables Easily extensible and customizable • Usable within any GDS, CRS/DCS Applicaons • Central repository for atomic level data Foundaon Analyc Presentaon Layer Layer Layer • Pre-built Analyc Models and Reports • Cloud enabled for DBCS

The Oracle Airline Data Model Product Roadmap – by line of business Complete Loyalty & Passenger Customer Business Experience R1 R1

Strategy & Finance & Airline Retail Baggage Planning Operaons R2 R2 R2 R2

Maintenance Flight Fuel Cargo Catering & Engineering Operaons Management R3 R3 R4* R4 * R4 *

* R4 To Be Conﬁrmed Based Upon OADM CAB Priories

• Customer Segmentaon Analysis – FFP – Non-FFP • Customer Loyalty Analysis • Customer Life Time Value Analysis • Frequent Flyer Passenger Predicon

Customer Loyalty Analysis • Idenﬁes factors that may have more inﬂuence on customer loyalty to airlines • Leveraging Support Vector Machine(SVM) algorithm and Decision Tree (DT) • The output from the model is twofold: – The discovered rules provide correlaon between the customer loyalty to Airlines and Customer aributes. – The predicon can be made on current base customer's data for the next month/quarter/year using the model built on historical data.

Copyright © 2017, Oracle and/or its aﬃliates. All rights reserved. | Customer Life Time Value Analysis • Idenfy/predict the customers who are likely to represent the highest value of revenue over their life me • Regression model used to ﬁnd out relaonship between customer informaon and their potenal value • Model can be used to predict future revenue for any new or exisng customer over next 1-5 years (# years can be adjusted in the model) • Leverages Generalized Linear Model (GLM) and SVM algorithm

Copyright © 2017, Oracle and/or its aﬃliates. All rights reserved. | Frequent Flyer Passenger Predicon • Idenfy/predict the Non-FFP(Non Frequent Flyer Passenger) customers who are likely to become a FFP • Classiﬁcaon algorithms, Support Vector Machines (SVM) and Decision Tree (DT) are used in this model • The training data would be mix of Non-FFP passengers and FFP passengers. FFP passengers are those who became FFP from Non-FFP in the last 1 year me period.

Copyright © 2017, Oracle and/or its affiliates. All rights reserved. | Customer Segmentaon Analysis • Groups customers based on customer demographics, flown history, booking paerns, profitability, etc. • Business analysts can further look into each idenfied segment to beer understand each customer group • Clustering rules draw profile of the customers; show most important similar characteriscs of each group. • Leverages K-means clustering algorithm

1 Baggage Handling Processes and Costs

2 Data Consolidaon

3 IOT, Streaming and Smart Tracking

4 Maximizing Customer Sasfacon

Copyright © 2017, Oracle and/or its aﬃliates. All rights reserved. | Conﬁdenal – Oracle Internal/Restricted/Highly Restricted 25 Proof of Concept for Smart Airlines Baggage RFID Tag RFID Reader Real me Capture & Streaming Analycs 1 Big Data Discovery Web Socket for 8 Publishing Oracle IoT “Push” Collecon Decoding Analycs Cloud Service 9 Dashboard, REST API for “Pull” Reports in-memory and Metadata in Historical 26 NOSql” 10

3 Event Coherence 7 Processor Oracle NoSQL 4 Big Data SMS Alert to Customer Smart bags Local Storage 2 with Sensors Property Graph Spaal Reports 6 Post Request for SMS JAVA API to Load DB Crossing Geo Fence with baggage data for 5 with Customer Communicaon Airline Service DW & Deep Learning Property Graph Gate as “Node” & Baggage as “Line GDS Global Distribuon System OADM on DBCS 15 16 Spaal for Bags in Geo Map 17 PSS Passenger Service System Structured DCS Departure Control System System of Baggage Records BRS Reconciliaon Database 12 Instance System Passenger, Travel & LOYALTY Derived Tables 13 14 Oracle Integraon Oracle Integraon Baggage Informaon 11 Cloud Service Cloud Service

Copyright © 2017, Oracle and/or its affiliates. All rights reserved. | Confidenal – Oracle Internal/Restricted/Highly Restricted Copyright © 2017, Oracle and/or its affiliates. All rights reserved. | Copyright © 2017, Oracle and/or its affiliates. All rights reserved. | Copyright © 2017, Oracle and/or its affiliates. All rights reserved. | Real Time Streaming Architecture link Reporting Netapp Layer (Angular JS) Server 1 RFID, Smart Tracker

Flume Kaka Storm Snapshot Server 2 Cluster Push Cluster Cluster Push Pull Engine Apache Server … Ticket Sales Drill

Metadata (NoSQL) Oracle DB NoSQL NoSQL cluster cluster BD SQL Transform (Derived (Base Ext OADM (Java) Tables) BD SQL Tables) tables

Copyright © 2017, Oracle and/or its affiliates. All rights reserved. | Confidenal – Oracle Internal/Restricted/Highly Restricted 30 E Airline logiscs monitor and analyze the flow of bags with real-me streaming analycs

Misrouted or stuck baggage can trigger alerts, and support incidents

Copyright © 2017, Oracle and/or its aﬃliates. All rights reserved. | 31 Logiscs can also follow the golf bag thru the airport, and across the country*, using real-me maps generated from the tracker’s GPS data and Oracle Spaal technologies.

*FAA requires that Smart Tracking devices to be deacvated while onboard the aircra

Data Data Sources Integraon Applicaon & Database Dashboards Layer Layer Bookings & Check-In XML Hosted Services Crew Scheduling Websocket IRROPS Baggage Handling Applicaons Rest API Aircra Scheduling

Weather EDIFACT Database

Airport Informaon AAA

Aircra Maintenance LDAP / IM / ID Plaorms Airlines IRROPS Team

Copyright © 2017, Oracle and/or its aﬃliates. All rights reserved. | IRROPS Dashboard Sample Single view for Flight, Crew and Seat (Capacity, Check-In, Boarded) Can be conﬁgured to show desired data points, in customized report formats

1 Baggage Handling Processes and Costs

2 Data Consolidaon

3 IOT, Streaming and Smart Tracking

4 Maximizing Customer Sasfacon

Social Mobile Online Travel Agencies Call Center Reservaons Customer Life Cycle Oracle Commerce Oracle Service Cloud Customer Social ATG Web Self Web Self Web center/ Web center/Mobile Mobile Applicaon ATG / Endeca Relaonship Engagement Commerce Service/OPA Service RTD App Framework Framework Management Customer Oracle Markeng Cloud Oracle Service Cloud Lifecycle Blue Kai/ Service Knowledge Siebel Loyalty Oracle Sales Management Responsys/SM Cloud for Contact Center Management/OPA Loyalty Analycs Cloud SOA/ESB/Middleware

Passenger Data Management Customer Hub Oracle Big Data Appliance Oracle DB Cloud

GDS Customer Oracle Linux 5.8 with Unbreakable Enterprise Kernel Oracle Airline Data Model Oracle Hotspot Java Virtual Machine Data Historical Enterprise DW Cloudera’s Distribuon including Apache Hadoop 4 PSS Management Oracle R Distribuon Data Mining Oracle NoSQL Database Community Edion Pre-built Analycs Oracle Big Data Appliance Enterprise Manager Plug-In Operaonal Data Store Loyalty

Unstructured Data

Copyright © 2017, Oracle and/or its aﬃliates. All rights reserved. | Oracle Conﬁdenal – Internal/Restricted/Highly Restricted 37 Mark Purchases his trip… Mark is heading to play golf near Miami, he’s planning to use his Smart Tracker to make sure he knows where his golf clubs are at all mes during the trip

Mark packs his bags He includes the Smart Tracker he received from the airlines Loyalty Program in his golf bag. His other bag will use the RFID tracking

Tue, May 31, 9:50am

OA1978 DELAYED. DEP now 1:43pm termA, gate3 LAX ARR 4:50pm term5, gate 53B Connecon OA1986 Delayed Travel… 2:05pm DEP Term 2, gate 23A in Mark checked in for the flight but as he arrives at JEOPARDY the LAX, he receives an alert that the flight is delayed. He’s worried about making his connecng flight in Chicago and contacts a service agent

Copyright © 2017, Oracle and/or its affiliates. All rights reserved. | Mark Makes it to Chicago! Despite the delay, Mark lands in his connecon city 15 minutes prior to final boarding for his flight to Miami Copyright © 2017, Oracle and/or its affiliates. All rights reserved. | Mark, your bag has arrived in Chicago. Mark Makes the Connecon! Ready to board flight 872 to JFK He receives an alert that his golf clubs have arrived in Chicago, but are not in the Miami-bound geo- fenced transit area Copyright © 2017, Oracle and/or its affiliates. All rights reserved. | Instant resoluon Sue also receives an alert about the misdirected bag, and sends

Mark is happy to know exactly where is golf clubs are, and where he can pick them up

Tue, May 31, 9:50am

Get 10% oﬀ from parcipang OA Partners! REPLY “help” to be connected with an Agent for more informaon. Customer Sasfacon help hp:// soclouds2-3.rightnowdemo.c Sue sends Mark om/app/answers/detail/a_id/ some discounts for 262 parcipang partners

Advocacy! Mark Posts his excitement at making it to the Miami.

Copyright © 2017, Oracle and/or its aﬃliates. All rights reserved. | Mark is now an Advocate for Oracle Air! We reduced Mark’s customer eﬀort and simultaneously increased his sasfacon, moving him from merely a fan, to a customer advocate.