1Thomas Paulus

CITT CENTRUM FÜR INFORMATIONS-TECHNOLOGIE TRANSFER GMBH

EP-TS Use Case Workgroup

Use Case:

Fraud Management in the Retail Domain

DEBS2009

2Thomas Paulus

AGENDA

Definition of Shrinkage Fraud Management nowadays Characteristics of the IT-Infrastructure Event Processing for Fraud Management Architecture of a prototype implementation

3Thomas Paulus

USE CASE: FRAUD MANAGEMENT IN THE RETAIL DOMAIN DEFINITION OF THE TERM „SHRINKAGE“

Shrinkage (also known as loss or shortage) can be defined as loss of stock caused by one or a combination of:

Shrinkage

Employee Fraud

Shoplifting and Customer Fraud

Administrative Error

Supplier Fraud

Theft Discount fraud …

Theft Return fraud …

Misspelling Accounting error …

Delivery error Accounting error …

4Thomas Paulus

The Global Retail Theft Barometer 2008:

Source: http://www.retailresearch.org/theft_barometer/index.php

USE CASE: FRAUD MANAGEMENT IN THE RETAIL DOMAIN FIGURES ON GLOBAL SHRINKAGE

Total global shrinkage: > 100 billion $

Only ~ 50% of stock loss is considered to be „known“

Expenses for loss prevention: >1 billion $

Sectors differ greatly, loss calculation differs greatly

About 1.34% of the overall retail salesAll Retail IndustryCustomer

Staff

Suppliers /Service

Organisation

5Thomas Paulus

AGENDA

Definition of Shrinkage Fraud Management nowadays Characteristics of the IT-Infrastructure Event Processing for Fraud Management Architecture of a prototype implementation

6Thomas Paulus

USE CASE: FRAUD MANAGEMENT IN THE RETAIL DOMAIN METHODS EMPLOYED AGAINST SHRINKAGE

PROCEDURES AND ROUTINES Annual stock loss awareness campaign Company-wide stock loss refresher training Customer returns & refund controls (operator and customer database) Damaged goods resale controls Employees exit searches Hot product identification Hot product management Hot products routine counting Security newsletter Internal key control Patrol routes for employees (red routes) Point of sale information or data checks Random till cash checks Rigorous delivery checking procedures Shelf replenishment techniques Induction training for new employees Unique till operator PIN numbers ‘Watertight’ product monitoring procedures

PEOPLE AND PROCESSES Anonymous phone line Civil recovery Covert surveillance of customers or employees Employee awareness and training Employee stock loss training and education Employee incentives—discount purchase schemes Employee incentives—stock loss bonus schemes Employee integrity checks External compliance monitoring External security/loss prevention function External stock audit function Internal compliance monitoring Internal security/loss prevention function Internal stock audit function Random checks on distribution centre picking accuracy Store detectives Test purchasing (mystery shopper) Uniformed security guards

EQUIPMENT AND TECHNOLOGY Automated ordering processes Cash protection tactics and equipment (both cash offices and tills) Company-wide stock loss awareness posters Dummy display cards in place of high-risk products E.A.S. hard tagging (recycled) E.A.S. soft tagging (disposable) E.A.S. source tagging (either disposable or recycled) Employee purchasing arrangements Employee panic alarms Employee uniforms without pockets Intruder alarm systems Non-active CCTV Point-of-sale camera monitoring Protector display cases applied by retail outlets R.F.I.D. intelligent tags on pallets, cases or items (radio frequency) Replenishment equipment to support techniques Secure lockers for employees Security-sealed containers/shippers Shoplifting and theft policy posters for customers and staff Specialist anti-theft display equipment

DESIGN AND LAYOUT Appropriate product location strategies Designing-out blind spots Designing-out crime programme Distribution centre secure storage Employees entry/exit access control External security—fences, anti-ram raid, roll shutters Risk-based design and layouts Robust anti-theft packaging Single direction product flow Supply chain and logistics network design

All these methodscan’t avoid

TODAYthe high levelof shrinkage!

7Thomas Paulus

USE CASE: FRAUD MANAGEMENT IN THE RETAIL DOMAIN REASONS FOR INEFFICIENT FRAUD MANAGEMENT

Why are thecurrent

methodsinefficient?

Autonomous Systems

Post processingof recorded data

Missing or not usedstandards

Manual fraud detection

Independent loss prevention systems Video surveillance Electronic article surveillance

Fraud is detected after it happens Too late for interventions

Individual implementations Hard to compare and measure

Efficiency depends on the quality of the staff Mistakes can be made

8Thomas Paulus

USE CASE: FRAUD MANAGEMENT IN THE RETAIL DOMAIN KNOWN STANDARDS AND PRODUCTS

Firstapproaches

to solvethe problems

Semantics and Standards

Infrastructures

ARTS NEAR (Notification Event Architecture for Retail) ARTS Video Analytics ARTS POSlog (formerly IXRetail) ARTS SOA Blueprint for Retail ...

IBM: Store Integration Framework Microsoft: Smarter Retailing Architecture Wincor Nixdorf: Store Communication Framework ...

9Thomas Paulus

AGENDA

Definition of Shrinkage Fraud Management nowadays Characteristics of the IT-Infrastructure Event Processing for Fraud Management Architecture of a prototype implementation

10Thomas Paulus

USE CASE: FRAUD MANAGEMENT IN THE RETAIL DOMAIN TYPICAL IT LANDSCAPE IN THE RETAIL DOMAIN

Heterogeneous

Independent Components

No common data standards between components

No common data standards between enterprises

Characteristics:

11Thomas Paulus

USE CASE: FRAUD MANAGEMENT IN THE RETAIL DOMAIN TYPICAL IT LANDSCAPE IN THE RETAIL DOMAIN

Sales-Data: x sales per shop y sales per hour z items per sale

Customer-Data: Customer Counter Credit / Debit Cards Loyalty programs

Security-Data: Electronic Article

Surveillance data RFID-tracking Customer Position

tracking Video Surveillance data

…...

High amount of business data:

12Thomas Paulus

AGENDA

Definition of Shrinkage Fraud Management nowadays Characteristics of the IT-Infrastructure Event Processing for Fraud Management Architecture of a prototype implementation

13Thomas Paulus

USE CASE: FRAUD MANAGEMENT IN THE RETAIL DOMAIN EVENT PROCESSING FOR FRAUD MANAGEMENT?

Current situation

Pro: Best Practice Already available Accepted technology

Contra: Still high amount of shrinkage

Possible solution based onevent processing technologies

Pro: Real-time fraud detection Pro-active fraud management

Contra: No implementations available Expenses Rare knowledge

Why should event processing technologies be used for fraud management in the retail domain?

14Thomas Paulus

USE CASE: FRAUD MANAGEMENT IN THE RETAIL DOMAIN POSSIBLE FRAUD SCENARIOS

Possible Fraud Scenarios and corresponding actions

A customer wants to return an already returned item

A customer returns very frequently items

A customer returns an item with a differing serial number

A customer wants to return a not sold item

Pattern detection

Abort the return process

Start a fraud management process

Call the police

Modify the customer properties

etc.

Scenario Action

15Thomas Paulus

AGENDA

Definition of Shrinkage Fraud Management nowadays Characteristics of the IT-Infrastructure Event Processing for Fraud Management Architecture of a prototype implementation

16Thomas Paulus

USE CASE: FRAUD MANAGEMENT IN THE RETAIL DOMAIN CONCEPT OF A PROTOTYPE IMPLEMENTATION

Fraud Management

Possible architecture for future Fraud Management

Process Layer

Application Layer

POS CRM …

Complex Event

Processing

BPM WorkflowEngine

Define Recognize Act

17Thomas Paulus

USE CASE: FRAUD MANAGEMENT IN THE RETAIL DOMAIN NEXT STEPS

Next steps for the Fraud Management System:

Implementation of the prototype BPM Suite: jCOM1

http://www.jcom1.com

CEP Engine: RTM (Realtime Monitoring) http://www.realtime-monitoring.de