Optimizing Supply Chains Through Machine Learning 2007-2016 ... Relationships from Sales Databases...
Transcript of Optimizing Supply Chains Through Machine Learning 2007-2016 ... Relationships from Sales Databases...
Optimizing Supply Chains Through Machine Learning
Lalit Wadhwa VP, Global Supply Chain Operations
AVNET, Inc.
@wadhwal https://www.linkedin.com/in/lalitwadhwa
Topics
• About Avnet
• Digital Supply Chains
• Data in Digital Supply Chains (DSC)
• Convergence of DSC & Machine Learning
• Case Studies and Applications
• Drivers for Machine Learning Adoption
• Wrap Up
Who We Are And What We Do
Who
We Are We are one of the world’s largest
global distributors of electronic
components, computer products
and embedded technology
serving customers in
more than 125 countries.
What
We Do
Financial
Scope For the fiscal year ending
July 2, 2016 we generated
revenue of $26.2 billion.
Who
We Are
What
We Do
Financial
Scope We are one of the world’s largest
global distributors of electronic
components, computer products
and embedded technology
serving customers in
more than 125 countries.
We connect the world's leading
technology companies with more
than 100,000 customers by
providing cost-effective, value-
added services and solutions.
For the fiscal year ending
July 2, 2016 we generated
revenue of $26.2 billion.
Company Snapshot
Americas
$10,424B
(40%)
Asia
$7,985B
(30%)
EMEA
$7,811B
(30%)
• Named to the FORTUNE Most
Admired list for technology
distribution, 2007-2016
• Top Business Technology
Innovator on the 2016
InformationWeek Elite 100
• No. 102 on the 2016
FORTUNE 500 (U.S.)
• No. 380 on the 2016
FORTUNE Global 500
• Named a World's Most Ethical
Company by Ethisphere
Institute 2014, 2015 and 2016
An Industry Leader
FY16 Annual Revenue
$26.2B
Avnet, Inc.
$9.7B
Avnet
Technology
Solutions
$16.6B
Avnet
Electronics
Marketing
• Headquartered in
Phoenix, AZ
• Founded in 1921
• AVT listed on
the NYSE in 1960
• 800+ suppliers
• 100,000+ customers
• 100 acquisitions
announced or closed
since FY91
• 17,700 employees
worldwide
Fast Facts
37%
63%
Disruption of the Supply Chain
Traditional Supply Chains
Advanced Analytics &
Machine Learning
Platforms
IoT and M2M Additive
Manufacturing
Advanced Robotics
Drones
Characteristics of Digital Supply Chains
Customer-Centric
Connectivity
Visibility
Traceability
Analytics & Intelligence
Agility
Scalability
Data Types in Digital Supply Chains
Source: Big Data Analytics in SCM, Ivan Varela Rozados & Benny Tjahjono
What is Machine-Learning?
Question Data
Collection & Preprocessing
Learning Algorithm
Model
New Data Model Predicted
Output
Types of Machine Learning
Supervised – Evaluate New Data Based on Prior Data
• Regression
• Classification
Unsupervised – Discover Patterns
• Segmentation
• Clustering
• Association
Reinforcement – Autonomous Agents
• Cumulative Reward Maximization
Structured Data Type - Examples
Process - Source
• Supplier
• Product
• Pricing
• Quantity
• Lead Time
• Inventory
• Shipment
• Quantity Multiples
• Financial Data
• Capacity
• Yield
Process - Deliver
• Customer
• Forecast
• Orders
• Product
• Lead Time
• Scheduling
• Pricing
• Promotions
• Discounts
• Shipping
• Return
• Financial Data
Process - Return
• Customer
• Product
• Return Reason
• Quantity
• Condition
• Product Reviews
• Financial Data
Business Query
• Customer Segmentation Based on 7 Attributes
Data Sources
• ERP, CRM, External Databases
ML Algorithm(s)
• K-means, Hierarchical Cluster Analysis (HCA)
Benefits
• Customized Product, Pricing & Marketing Strategy for Each Segment
Customer Segmentation
Business Query
• Discover Product Association Relationships from Sales Databases
Data Sources
• ERP, CRM
ML Algorithm(s)
• Apriori, Max-Miner
Benefits
• Targeted Promotion, Sales Uplift
Product Association
Business Query
• Optimize Product Pricing to Achieve Predetermined Objectives
Data Sources
• ERP, RFQ, External Databases
ML Algorithm(s)
• Decision Tree, Random Forest
Benefits
• Improve Quote Win-Rate and Profitability Simultaneously
Pricing Optimization
Business Query
• Demand Estimation for New Product Introduction
Data Sources
• CRM, Sales Data, External Databases
ML Algorithm(s)
• Lasso Regression, SVM, Random Forest
Benefits
• Simultaneous Improvement in Inventory Profile & Margin
Demand Estimation
More Examples
Sales Pipeline Win Propensity Prediction
Supply Chain Security & Fraud Detection
Forecasting Obsolescence Risk
Optimizing Warehouse Operations
Manufacturing Process Optimization Through ML
At the Intersection of Supply Chain and Machine Learning
Drivers for Machine Learning Adoption
• Alignment with Business Strategy
• Pilot Projects
• Change Management
– Data Driven Decision-Making Culture
• The Role of Human Insight
• Talent
– Convergence of Business, IT and Data Science Experts
Challenges To Be Aware Of
• Data Relevancy
• Data Quality
• Data Quantity
• Carefully Separate Correlation from Causation
• Static Models in Constantly Changing Business Environment
Wrap Up
• Digitization of the Supply Chains
• Transformational Impact of Machine Learning
• Machine Learning in B2B Supply Chains
• Succeeding with Machine Learning Adoption