MSC- Who We Are and What We Do
107Countries Served
>390Number of Suppliers(includes component suppliers)
~$8B Total Amount of Annual Spend
30KShip To Locations
1TBTotal Daily SC Data Generation
1.7M+ Est. FY17 Number of Retail and Commercial Sales Orders
33Number of Mfg & Distribution Operations
>42,000Number of Active SKU’s
2.0M+Est. FY17 SAP Deliveries
>77M Units Manufactured & Shipped
Strategic
SourcingCare
Manufacturing
Supply Chain Mgmt./
New Product Introduction
Global
Supply
Chain
Transformation,
Technology
and Analytics
Integrated
Business
Planning
Safety,
Compliance
& Sustainability
Our Products
PC Hardware
Surface Hub HoloLens
Software
Xbox
Surface
Surface Studio
MSC Global Presence
Argentina
Colombia
Canada
Australia
Austria
France
Japan
Singapore
Malaysia
UK
UAE
Germany
Russia
Brazil
Hong Kong SAR
Czech Republic
Netherlands
TaiwanMexico India
Finland
Microsoft Supply Chain Challenges
Pricing and Margin Pressure
Increasing Customer Service Expectations
1 Digitized Supply Chain
MSC’s 1DSC Transformation Journey
The Azure Platform Enables our Connected Digital Supply Chain Foundation
Supply Chain Partner Systems
Warehouse
ManagementShop Floor Control System Transportation
ManagementBOM Management
Corporate systems
CommerceDevice Telemetry CRM
SubscriptionsProduct
Catalog
Sales
Point of Sales Freight
Management Enterprise
Services
Supplier ERP
Data & Advanced Analytics Platform
HD Insight Event
HubsAzure Data LakeAzure Data
Factory
Supply Chain Cloud
Supply Chain Services
Azure API
Management
Microservices Azure Application
Insights
Microsoft Supply Chain Core Transactional Systems
Make
(MDS)Deliver Care
Dev/NPIFinance
Source Plan
Drive world class Customer Care, Product Quality & Compliance
Enable Scale and Complexity
Effectively Deliver Products & Services to Market
Aggressively Manage Costs
Connected: MSC Purpose is reflected in what we measure via Power BI
Microsoft Manufacturing Digital Transformation: Leading
the Industry with Microsoft Tools/Technologies
Factory Video
Speed
Scale Capture - 1B data points/day, Analyzed < 1%
Accessibility to data measured in hours/days
Predictive Factory’s with MDS, a Azure-based system for real-time factory floor & product lifecycle analytics
Before: Static, siloed data (TCS, Shop Floor…) After: Enables proactive alerting, predictive analytics
MTE TCS
T2T1
JDM1
Shop Floor
Any T1 or Critical T2
Returns
JDM1 Shop
Floor
TCS TCS
Share
Point
SQL
EIP
Other T1Critical T2T2
JDM1
Reduced Returns, Higher Yields, Better Productivity
Returns
System enables Power BI, machine learning, advanced analytics & decision making capability
E2E Visibility from Incoming to Customer Proactive Alerting, Real time Insights
Incoming Part
Quality & Availability
CTF/CTQ
IoT Process Monitoring
Product
Yield
Outgoing
Quality
Field
Quality
Returns
Repair &
Refurb
T2
Quality
Output & Shipment
Progress
Incoming Manufacturing Customer
Proactive
Alerting
Predictive
AnalyticsBig Data
Mining
Real time
Insights
• Creating Clarity via Personalized Dashboards
• Increasing Collaboration, Connected Data Streams
enable Teamwork
• Improving Factory Productivity/Optimization
• Big Data insight and Machine Learning
• Proactive Alerting
• Predictive Analytics
Implementation is changing the way we (Microsoft & partners) manage the day to day factory operations
Personalized dashboards align to Factory Manager’s purpose
Example: Factory Manager Product Dashboard Example: Proactive Alerting
Surface
Pro 4
Alerts identify one tester drifting out of control.
Tester taken off line, and issue corrected before yield impacted
Alerts identify significantly different battery placement measurements on one line vs others
Fixture alignment issue corrected before quality impact
Proactive alerting monitoring critical IoT enabled measurement data
Analytics engine evaluated multiple attribute
variables to predict top patterns across
Wafers, Chips
Model predicted optimal patterns and
optimized manufacturing parameters for
improved yield
Results
Predictive
insights
Factory
Data
Azure ML
Targeted patterns/combinations
demonstrated ~30% yield improvements
and $2M in scrap reduction
Assembly component, Genealogy, MES,
Environmental
Training Machine Learning models to
identify drivers/patterns of SP4 returns
attributed to critical HW genealogy,
software telemetry & CSS data
Current model predicts ~80% recall rate for
a returned device.
Improving precision with additional Factory
and Customer/Retailer data
Expected
Results
Current
insights
Azure ML
Targeting to achieve:• Improved CSS/Customer interaction
• 1st Returns reductions, better quality
• Improved returns forecasting
Product
Big DataProduct Experience• SW release/updates
• App/accy Incompatibility
• Windows telemetry
Product Design & Mfg• Mfg/Test process
• Components/supplier
• Hardware/device design
As a result of the foundation, we are now starting to use predictive analytics & machine learning in the factory
Example: Predictive Yield Improvement Example: Predictive Return Insights
Digital Transformation increases Pace
Data is the new Currency