Global trade analytics

Delivering business insights to manage risk

and enhance global trade effectiveness

September 2014
Presenters

► Alex Kuperman, Ernst & Young

LLP (moderator)
► Colm Halpin, Ernst & Young
(Ireland)
► Ann McDonald-Hintze,
General Motors
► Muir Macpherson, Ernst & Young
LLP (Quest)

Page 2 Global trade analytics

Global trade challenges

How much duty is paid

globally? Is this an area
Customs where real and
authorities significant savings can
Customer be found?
Exports
Is trade data (which
Where are internal and CRO drives duty costs)
external trade recognized as master
Broker(s)
compliance resources data? Is it accurate and
most needed and how consistent across
should these be Customs Distribution
center locations and countries?
managed? operations
Are distribution centers
How are brokers (DCs) adding to customs
performing from a CFO Export duty costs or losing free
performance, cost and compliance trade agreement (FTA)
compliance perspective? benefits?
Do transfer pricing
ERP (TP) and tax team
Imports understand and
appreciate the duty
Receiving cost and risk impact of
Fulfillment changes to TP model
Supplier COO or TP adjustments?
Manufacturing
Could the business be re-
aligned to increase trade
benefits, i.e., is company
deriving the cost and
Can finance/tax teams access Is supplier trade data Are procurement and Is full advantage taken of competitiveness benefits
quantitative data on customs and correct? Are sales and purchasing teams able to FTAs and customs from supplier-plant and
trade for duty risk controls and delivery consistent with factor in customs duty regimes? Do we have plant-customer
governance? Will OECD require new contract, incoterms costs and savings potential metrics and/or has anyone configuration?
data and reports? for alternative sources? ever tested for leakages?

Page 3 Global trade analytics

Analytics can deliver business insights

To drive better decisions, we must first ask

Thus, our work moves left to right, but our
the right business questions and then seek
thinking must move from right to left.
answers in the data.

Organizations are looking to data analytics to drive actionable insights, improve business
performance, drive better decision making and improve risk management.

Page 4 Global trade analytics

Global trade data analytics can provide
enterprise-wide benefits
Business functions and trade related concerns

Customs and trade compliance Tax and transfer pricing

Duty minimization Broker management

Export control Supply chain security

Technology deployment Supply chain network strategy optimization

Operating model efficiency to support trade Readiness for entering new markets

Supplier and third-party trading strategies Country and specific legal alignment

Defining and prioritizing the objectives is the starting point

Page 5 Global trade analytics

Analytic capabilities

Companies need to prioritize what is most important to improving their global trade operations. As
capabilities advance, the level of complexity and associated time horizon increases.
Capabilities

Framework Report Analyze Predict Synchronize

Strategy ► Analytics strategy aligned to Visibility Management Risk and duty Operational
business imperatives and modeling integration
goals
The ability to mine past The ability to apply The ability to utilize both The ability to align,
data and report on comprehensive analysis past transactional data integrate and
transaction activity to gain a better to proactively identify collaborate,
(value, volume, timing understanding of possible compliance for known or unknown
and costs) program effectiveness, risks prior to the needs, through real-
Information ► Necessary IT framework to risk patterns, occurrence, optimize time scenario planning
infrastructure collect, protect and analyze processes, product duty costs and supply and simulation across
data level costs and risk chain trade-offs an extended network,
► Visualization layer for global resulting in a complete
distribution across different demand, supply and
functions and levels finance model

Data ► New data source integration

management ► Governance program to
(acquisition, improve data quality
quality and ► Change management to
governance) create long-term
sustainability

Reporting and ► Standard and bespoke

analytics analytics
► Performance benchmarks
and Key Performance
Indicators (KPIs)
► Reporting dashboards

Page 6 Global trade analytics

Technology considerations

Performance measurement

This layer will allow

Key considerations executives, management,
analysts, and operations to
► Requires ongoing input Visualization effectively view and use
information to make
from both business and layer decisions.
technology group
Decision and support analytics Decision support/analytics
► The benefit of The tools and processes
“real-time” Management Program Transaction Predictive within this layer will be
standardized and improve
reporting analytics analytics risk
the overall value of the data
modeling within the organization.
► Supply chain
Source and repository technology
visualization Data warehouse
The most complex of
layers, and data feeds will
go through the
Extract/Transform/Load
(E/T/L) process and should
be automated. Data
warehouse/Data marts will
provide effective decision
support systems.

External data sources Enterprise systems Open data

► Brokers ► Enterprise Resource ► Government
► Carriers
Planning (ERP) ► Industry
► Global Trade Management
System (GTMS)
*Note: This framework is a conceptual model only. It is not meant to imply specific recommendations on architecture.

Effective analytics approaches require that each layer be addressed

Page 7 Global trade analytics

A look inside GM’s data analytics program

Page 8 Global trade analytics

The starting point

► In June 2012, GM Global Customs initiated its Global

Customs Business Transformation.
► Recognized the need for consistent, timely, cumulative
global import data
► Developed methodology for obtaining and utilizing data
from multiple business units and import systems
► Identified critical data elements and determined roll-out
cadence starting with largest importing business units
► August 2013, established the Global Customs Data
Warehouse

Page 9 Global trade analytics

Strategic drivers

► Support corporate vision by providing easy to understand

information, insight and metrics to drive the right business
decisions
► Collaborative work with organizations outside Global
Customs require accurate, timely and reliable data
► Use of data analytics to identify potential duty and tax
savings opportunities
► Established Centers of Excellence to enhance operational
efficiency, recognize savings opportunities and identify
potential compliance risk

Page 10 Global trade analytics

Global customs
Entry data warehouse
Of 44 countries of import in total, only 5 are outstanding and those are
expected to be inducted by September 30, 2014.
Data elements
► Importing country/business unit 27 unique data elements, with all value fields also
► Ship from name/DUNS converted and stored in common currency – USD
► Ship to name/DUNS/plant code
► Part #
► Description Data is received through three channels:
► Quantity
► Scheduled extraction from GM’s global import
► Value
system
► Scheduled extraction from other import systems
► Country of origin
► Extractions from broker systems – manually
► Country of export
provided and reformatted for upload
► Free trade agreement status
► Amount of duty paid Records to date > 15.5m
► Mode of transportation
► Category(Prod/P&A/VEH/M&E) Dates of import covering 1 January 2013 to
► Entry/invoice detail date
► Import related taxes and fees

Page 11 Global trade analytics

Global cash duty paid report

Effective duty rate, entered value and duty paid are available at the global and country level within
minutes, versus days.

Page 12 Global trade analytics

Year-over-year analysis

Various views of the data can provide visibility into our business.

Page 13 Global trade analytics

Import line count report by category

Page 14 Global trade analytics

Duty savings opportunity identification

Page 15 Global trade analytics

Duties paid by global purchasing teams

Tailoring the information for the “consumer” is critical.

Page 16 Global trade analytics

The role of the data miner

► Relational database knowledge is essential to join import

data with other GM customs data, such as FTA eligibility
and Harmonized Tariff Schedule (HTS).
► Master reporting tools – Tableau, MS Excel
► Assist in design and implementation of interim approaches
for non-standard data
► Extract data and build reports, tailored to various
audiences
► Reports
► Data

► Metrics

► There is a huge benefit if these individuals understand the

data and can offer ideas and solutions.

Page 17 Global trade analytics

Lessons learned

► Resources to mine and analyze provide a level of

operational visibility critical to our business success.
► Ensure appropriate systems/storage/data mining
capabilities are available and utilized.
► Be mindful that collected data adds value, is useful and is
readily understood by the users.
► Get the word out that the data is available, encourage
other organizations to take an interest and be invested.
► Demonstrate the value of the data collection,
dissemination and utilization process.

Page 18 Global trade analytics

Duty modeling using enterprise data

Page 19 Global trade analytics

Why an enterprise approach?

► Global trade costs and risks Common challenges:

vary enormously from country ► How much duty do we pay by country?
► How much do we save?
to country. ► What trade compliance resources/tools
are needed and where?
► Duty rates (levels and variability) ► Where would we likely find “duty
and classification complexity leakages?”
► Are our sourcing and distribution supply
► Available inbound/outbound trade chains duty optimized?
► How can we assess and manage
agreements, customs regimes business changes that have customs
duty cost or compliance implications?
► Customs and trade environment Conversely, can we factor in customs
duty costs to business change decisions?
(pitfalls and non-compliance
costs)
► Volume and complexity of the
business

Page 20 Global trade analytics

Using enterprise data to create duty models

► Flows can be mapped from ERP or similar data, e.g.,

purchasing database.
► Source data can include non-customs elements, e.g., sku,
product group, business unit division etc.
► Outputs (by country and business unit etc.) can include:
► Duty costs, savings and trade compliance risk heat maps
► Stack ranked duty leakage and optimization opportunities
► Predictive duty data, e.g., for scenarios, change management
► Other users (in the organization) can include supply chain,
procurement, tax, finance, planning, legal and internal
audit.

Page 21 Global trade analytics

Enterprise data
Examples of results
► Visual duty rates model
► These models show duty rates by flow lane for each
material/product.
► Color-coding highlights sub-optimizations
► Alternative flows (Available (AVL) and non-AVL) can be
toggled.
► “Scenario” source and manufacturing countries can
be added by material/product, e.g., China, Vietnam.
► Visualization can be adjusted, e.g., from linear (as
shown) to filled map, according to user group.
► Filters can include or Business Unit (BU), PG. Models
can point to specifics of the FTAs and preferences.
► Duty cost and optimization model
► This model builds in the $ value of the flows (again, Global / Regional / Loca(All)

color coding represents duty efficiency). [Dbl click country to drill down] [Dbl click code to see all rates]
Active Supply Base

► Model calculates the gross duty (red), net duty (if available Material or Product Country Name HS6 Local Code Used Optimization Statutory Rate Rate as % Basis Values UoM CN TH BR CL IN MA TN US ZA
per WT Val
FTA savings are claimed) and excess duty (additional Soya oil Australia 1507.90 AU-1507.90.00.13 MFN 5.0% 5.0% MFN Rate as % 5.0% 5.0% 5.0% 5.0% 5.0% 5.0% 5.0% 5.0% 5.0%
Volume 10,107,181 15,160,768
savings that could be realized) Excess v. Best Pref 5,053,590 7,580,385
PREF 4.0% 4.0% DCS Rate as % 4.0% 4.0% 4.0% 4.0% 4.0% 4.0% 4.0%
► Highest realizable FTA savings and FTA opportunities Volume
Excess v. Best Pref
10,107,181
4,042,872
15,160,768
6,064,308
stack ranked to check “leakages” and “optimizations” BESTPREF Free 0.0% AANZFTA Rate as %
Volume
0%
15,160,768

Flows with highest net duties can be checked for

Excess v. Best Pref
► ACI-FTA Rate as % 0%
Volume
regime and tariff suspension opportunities, duty driver Excess v. Best Pref
AUS Rate as % 0%
planning (valuation). Volume
Excess v. Best Pref

► Duty model supports focused savings projects (and TAFTA Rate as %

Volume
0%
15,160,768

provides business case cost/benefit). Excess v. Best Pref

Page 22 Global trade analytics

Enterprise data
Examples of outcomes
► Energy (Electrical) Product Manufacturing Company
► Global duty model (based on aggregated
classification) to support compliance and cost
reduction
► Country trade profiles by BU that identify the key
areas for cost reduction
► Benefits estimation $10m in Year 1
► Data aggregation and cleansing for SAP GTS

► Tobacco manufacturer global sourcing, manufacturing and

distribution
► Key raw materials (leaf) to primary tobacco plants
► Intermediates from primary to product finishing plants
► Finished goods to end markets
► Duty optimization tool now in use by procurement/
planning teams for more than five years
► Benefits not disclosed, but understood to be very
significant

Page 23 Global trade analytics

Some final comments

► Duty modeling can provide country, regional and global

insights to customs duty costs, risk and opportunities
relevant to one or many stakeholders
► The model allows targeting and drill-down to direct effort
and resources to where they provide the best return
► Predictive modeling can support business change
strategies and point to key trade impacts, e.g., supply
chain or transfer price model changes and TP
adjustments.
► Modeling and broker/customs analytics are essentially
complementary – which one to use, and when, depends
on the direction and role of the trade function within the
business.

Page 24 Global trade analytics

Risk modeling – predicting risky
transactions

Page 25 Global trade analytics

Defining the problem

► Facing significant pressures to limit new Full Time

Employees (FTEs) in the trade compliance departments,
while still being required to help grow business and
regulatory risks
► Spending too much time on tactical activities detracts from
addressing strategic issues
► Expansive “import/export” footprint is difficult to monitor
► Fees and fines resulting from lack of proper controls
► Growing threats to an organization and its risk and
compliance framework from unconventional elements
► Competitive business pressures and regulatory needs
requiring sectorized compliance perspective
Page 26 Global trade analytics
Unifying insights with advanced analytics

► Involves building Individual global trade

economic/statistical models in analytic insights can be
order to understand key behavior powerful, but may not be a
drivers useful guide to action
when multiple insights
► Greater precision to understand
compete for attention.
why something has happened or
what will happen by relating data
inputs to data outcomes
► Example: What is the probability this
shipment has a misclassification?

Page 27 Global trade analytics

The benefits of advanced analytics

Quantitative, risk-based approach to Free up current resources to work on

compliance monitoring business support or other more
value-added activities
Enables periodic monitoring to detect
errors early on before they become
systemic “Multivariate” approach uncovers
main behavior drivers and anticipates
riskier transactions
Additional capability that enhances
compliance management activities
Tap into and access an experienced
knowledge pool to provide insights
Internal control testing – evaluate the and assist with process if possible
sufficiency of process controls and Reasons for using
identify additional (more specialized)
process controls
advanced analytics Improve the predictability of duty
spend

Evaluate the effectiveness of Improve compliance, operational

compliance audits efficiency

Advanced analytics helps optimize your risk and compliance activities to help you govern your
business and achieve strategic objectives with predictable cost.

Page 28 Global trade analytics

Case study in predicting tariff classification
errors
A commonly used
The number of unique shipper reduces
items – each requiring a uncertainty and lowers
separate classification – the chance of a data
increases complexity entry error or miss
classification

Product Daily line Daily unique HTS Code Shipper

Import date Export date CO POE Mode HTS Currency Shipper Daily unique POE
description total codes frequency frequency frequency
4/5/2014 3/31/2014 SG 3901 40 9018903000 USD ACME, SG SWITCHBOARD 122 77 9 1.0% 0.1% 21.3%
4/5/2014 3/23/2014 CR 4909 11 9018908000 USD ACME, CR TRANSFER SET 122 77 9 41.6% 0.9% 30.2%
4/5/2014 3/23/2014 CR 4909 11 9018908000 USD ACME, CR MEDICAL DEVICE 122 77 9 41.6% 0.9% 30.2%
4/5/2014 3/31/2014 US 5271 40 9026104000 USD ACME, CR METER 122 77 9 0.05% 0.01% 30.2%
4/5/2014 3/23/2014 CR 4909 11 9018908000 USD ACME, CR HOUSING 122 77 9 41.6% 0.9% 30.2%
4/5/2014 3/31/2014 US 5271 40 9026104000 USD ACME, CR METER 122 77 9 0.05% 0.02% 30.2%
4/5/2014 3/23/2014 CR 4909 11 9018908000 USD ACME, CR PUMP 122 77 9 41.6% 0.9% 30.2%
4/5/2014 3/31/2014 US 5271 40 9032896040 USD ACME, CR CALIBRATOR 122 77 9 0.1% 0.01% 30.2%
4/5/2014 3/31/2014 AT 2720 40 3002100190 EUR ACME, AT SOLUTION 122 77 9 12.2% 0.04% 7.1%
4/5/2014 3/23/2014 CR 4909 11 9018908000 USD ACME, CR SECUREMENT 122 77 9 41.6% 0.1% 30.2%
4/5/2014 3/23/2014 CR 4909 11 9018908000 USD ACME, CR SYRINGE 122 77 9 41.6% 0.9% 30.2%
4/5/2014 3/23/2014 CR 4909 11 9018908000 USD ACME, CR BAG 122 77 9 41.6% 0.3% 30.2%
4/5/2014 3/31/2014 CH 4110 40 3002100190 EUR ACME, BE ADAPTER 122 77 9 12.2% 0.1% 10.4%

HTS codes outside of the

Certain HTS codes(e.g., client’s regular import mix
parts of, basket provisions, increase the probability of
etc.) are inherently more an incorrect classification
challenging to determine

Page 29 Global trade analytics

Insights on key outcome drivers

15
► Days with many unique product
codes had higher errors

Restatements
10
► Low volume shippers had error
rates close to 100%

5
► Infrequently used HTS codes
accounted for nearly 80% of

0
0 20 40 60 80
Unique product codes per day
errors.
.8

1
.8
.6
Percent restatements

Restatement rate
.6
.4

.4
.2

.2
0

0 10 20 30 40 0 10 20 30
Frequency by HTS code Total shipments for each shipper
Page 30 Global trade analytics
Unifying insights with advanced predictive
analytics

Unique
products

Statistical
Shipper
volume

Predicted
model

HTS
frequency
probability
Country
of origin

HTS
complexity

Page 31 Global trade analytics

Predicting tariff classification errors

Effective risk management is about identifying potential issues before they turn into systemic errors.
Advanced analytics can help identify which shipments represent the greatest risk of leading to a
compliance error.

5 April 2014
Exporting Predicted
country HTS number Shipper Product description probability
SG 9018903000 ACME, SG SWITCHBOARD 81.50%
CR 9018908000 ACME, CR TRANSFER SET 1.79%
CR 9018908000 ACME, CR MEDICAL DEVICE 1.79%
CR 9026104000 ACME, CR METER 92.49%
CR 9018908000 ACME, CR HOUSING 1.79%
CR 9026104000 ACME, CR METER 92.49%
CR 9018908000 ACME, CR PUMP 1.79%
CR 9032896040 ACME, CR CALIBRATOR 92.40%
AT 3002100190 ACME, AT SOLUTION 2.23% High-risk
CR 9018908000 ACME, CR SECUREMENT 1.79% shipments are
flagged
CR 9018908000 ACME, CR SYRINGE 1.79%
CR 9018908000 ACME, CR BAG 1.79%
BE 3002100190 ACME, BE ADAPTER 12.01%

Page 32 Global trade analytics

Predicting tariff classification errors

“Back-testing” against actual tariff classification errors was found to have a high predictive validity.

5 April 2014
Exporting Predicted Error
country HTS number Shipper Product description Probability Y/N
SG 9018903000 ACME, SG SWITCHBOARD 81.50% Y
CR 9018908000 ACME, CR TRANSFER SET 1.79% N
CR 9018908000 ACME, CR MEDICAL DEVICE 1.79% N
Both errors
CR 9026104000 ACME, CR METER 92.49% N were correctly
CR 9018908000 ACME, CR HOUSING 1.79% N identified
CR 9026104000 ACME, CR METER 92.49% N
CR 9018908000 ACME, CR PUMP 1.79% N
CR 9032896040 ACME, CR CALIBRATOR 92.40% Y
AT 3002100190 ACME, AT SOLUTION 2.23% N
CR 9018908000 ACME, CR SECUREMENT 1.79% N
CR 9018908000 ACME, CR SYRINGE 1.79% N
CR 9018908000 ACME, CR BAG 1.79% N
BE 3002100190 ACME, BE ADAPTER 12.01% N

85% of transactions correctly classified 100% of errors correctly identified

Page 33 Global trade analytics

Case study summary

► Total invoice lines: 1115 100%

Share of total reclassification

90%
80%

► Error rate: 32.7% 70%

60%
50%

value
Total invoices lines > 80%
40%
► 30%
20%
predictive probability: 146 10%
0%

33%
62%
73%
78%
82%
85%
87%
89%
91%
93%
94%
95%
96%
97%
97%
98%
98%
99%
99%
99%
100%
► Total invoices lines > 90% Share of total transactions
predictive probability: 77
51.6%

► Total invoices lines > 95% Transactions Restatements

predictive probability: 23 32.6% 32.6%

23.6%

15.9%
13.6% 12.9%
7.9%
6.0%
3.3%

0%-20% 20%-40% 40%-60% 60%-80% 80%-100%

Page 34 Global trade analytics

Key success factors

► Identify and align on the most critical business problems to solve and build the
solution around it – identify the key questions or themes that the global trade function
wants to get answers to, but has not had the capability to do historically

► Build the business case – articulate the value, what constitutes success and the high
level roadmap to guide the journey

► Obtain executive commitment – and maintain it throughout the journey for success to
be realized

► Collaboration and compliance – seek input from a broad range of functional areas
and supply chain partners and obtain buy-in as far as information relay, update cadence,
and accuracy/quality of data are concerned

► Conscious focus on data management – define and enable robust data management
capabilities that support high data quality

Page 35 Global trade analytics

Circular 230 disclaimer

This presentation is provided solely for the purpose of

enhancing knowledge on tax matters. It does not provide tax
advice to any taxpayer because it does not take into account
any specific taxpayer’s facts and circumstances.

These slides are for educational purposes only and are not
intended, and should not be relied upon, as accounting
advice.

Page 36 Global trade analytics

Disclaimer

► EY refers to the global organization, and may refer to one or more, of the member firms
of Ernst & Young Global Limited, each of which is a separate legal entity. Ernst &
Young LLP is a client-serving member firm of Ernst & Young Global Limited operating in
the U.S.

► This presentation is © 2014 Ernst & Young LLP. All rights reserved. No part of this
document may be reproduced, transmitted or otherwise distributed in any form or by
any means, electronic or mechanical, including by photocopying, facsimile
transmission, recording, rekeying, or using any information storage and retrieval
system, without written permission from Ernst & Young LLP. Any reproduction,
transmission or distribution of this form or any of the material herein is prohibited and is
in violation of U.S. and international law. Ernst & Young LLP expressly disclaims any
liability in connection with use of this presentation or its contents by any third party.

► Views expressed in this presentation are those of the speakers and are not necessarily
those of Ernst & Young LLP.

Page 37 Global trade analytics