E Ford Merkle Jun 8

Report
Supply Chain Monitoring
Ford Parts Supply & Logistics
Case Study
Roger Merkle
Former Manager, North American Inventory Planning Department
Ford Parts Supply & Logistics (PS&L)
Business Overview
• PS&L operates a network consisting of:
>
>
>
>
198K unique part numbers, over 1M SKUs (part/location)
Wide mix of velocity, size, and value
Vehicle base - 50 million vehicles on the road, 35 model years
Logistics network - 2,000 suppliers and 5,900 authorized
dealerships
> 18 HVCs, 3 HCCs, 1 LV/LC, 1 PRC, and 1 NPD
> Service Parts - US, Canada, Mexico and direct global export
• Complex logistics hubs, many containers, railcars, suppliers,
packagers, sources/destinations, and paths
• High degree of magnitude and complexity
SUPPLIERS
PACKAGERS
REDISTRIB.
CENTER
REGIONAL
DISTRIBUTION
DEALERS
Ford PS&L … Gearing Up for Change
• Acquired and centralized
relevant data sources
• New systems for forecasting, inventory planning,
DRP, electronic supplier communication and
management
• Implemented Supplier
Performance Monitoring
• Reduced inventory by
two thirds
• Record customer service
levels
• Record turn rates
Ford - Lack of Integrated Data
DISTRIBUTION
PURCHASING
SUPPLY
Low purchase price
(High
Inventory)
Low Inventory
Reliable Suppliers
(Stable
Flexibleschedules)
schedules
(Long
Short lead
lead times)
times
Low inventories
(High
inventories)
Short lead
lead times)
times
(Long
Flexible
Transportation
(Process-focused)
High customer
service
Stable Fixed Costs
Stable Fixed Costs
Low inventory
Low inventory
Stable Part Mix
Stable Part Mix
Stable Schedules
Stable Schedules
Low Transportation
Cost
Low Transportation
Cost
CUSTOMER
FULFILLMENT
Voice of
of the
the customer
Voice
customer
(HighService
inventories)
High
Levels
(High safety stock)
Managed inventories
(High Costs)
Managed safety
expediting, overtime
stock
Managed Costs
expediting, overtime
SELL
Returns
SOURCE
(MAKE)
DELIVER
RETURN
Ford’s Business Challenges
Ford PS&L Supply Chain Overview
PLAN
SOURCE
SUPPLIERSData
Analytical
updated
weekly, at
S
S
S bestS
Count - 2,000
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
Carrier
S
S
S
S
Highly
Reactive
S
S
S
S
focus
on
S
S
S
S
backorders
and
S
S
S
blameS assessment
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
S
DELIVER
REGIONAL
DISTRIB.
CENTERS
Different cultures,
processes and
practices at each
PDC
node
Reclass
Emergency
Increasing
complex supply
Bypass /
chain including
Redirect
external partners
Emergency
and sources as
Bypass /
well as non
Redirect
traditional
channels
DEALERS
Material expedited
by teams of
D people
D
D
D
at headquarters
D
Count - 5,900
Count - 8
PDC
Carrier
fro
m
D
DC
D
x-fer
PDC
S
S
D
D
D
D
D
Emergency
PDC
Metrics not aligned
- Data not common
PDC
Stocking PDC Consumption
D
D
D
D
D
PDC
D
D
D
D
Stocking PDC Consumption
Forecast
accuracy
PDC
and safety stock
management
RETURN
D
D
D
D
D
D
D
* Voluminous
reports - both
paper and
electronic
* Labor-intensive
to collect
RETURN
D
D
Little ability
toD
D
Stocking
D
D
PDC Consumption
D
prioritize which
D
D
D
Consumption
D
actions are
critical
D
D
D
to the business
D
D
D
D
D
requirements
D
PDC
S
S
D
D
D
S
S
D
D
an
yP
DC
ny P
to a
CENTER
S
S
S
High levels of
processing time
variability NATIONAL
PARTS
+
DEPOT
Forecast error over
supply chain process
Reclass
Bypass
=
High safety stock
PACKAGERS
PARTS
levels REDISTRIBUTION
Count -7
Carrier
(MAKE) / DELIVER
D
D
D
RETURN
D
D
D
D
D
D
D
Ford’s Requirements
•
Culture Change
•
Model-based analytics
•
Closed-loop issue management
•
Comprehensive visibility
•
Support segmentation
> Enable lean performance of existing
systems
> No disruption to existing operations
•
Prediction
> Support intelligent, proactive
analysis vs. reactive
> Predict impact of current plan within
lead time for resolution
•
Prioritization
> Drive data to lowest actionable level
in organization
> Identify high-impact opportunities
> Combine forecasted and actual
demand levels
> Manage material velocity based
upon any desired variable of
prioritization
> Manage escalation
> Adapt to any supply chain, any level
of data availability
> Calculate metrics across “white
spaces” where data availability is
poor
> Combine varying sources of data
> Support analysis of current
operating business systems
> Manage variability in real time w/
feedback to analysts / source
systems
> Manage approval process for
recommended changes
> Identify segments and processes
with biggest problems
> Locate specific material throughout
the supply chain
> Assess historical performance
Build vs. Buy Decision
• Why not custom?
> Integrated solution (data acquisition, data model, analysis,
prediction, ad-hoc OLAP capability, security, alerting,
administration)
> Investment in complex algorithm development
> Speed of implementation (rapid ROI)
> Proven business value
> Teradata Supply Chain Intelligence (SCI) provides Standard KPIs,
Reports & Alerts
> SCI Based on Industry Standards & Best Practices
• Technology Benefits
>
>
>
>
>
>
Scalable database architecture
Operational use of analytics
Expandable and configurable data model and analytics
Reduced Support Costs
Multiple database repository support
Developed Exception Management System
(alerting, escalation, message broadcasting)
Ford purchased Supply Chain Intelligence (SCI) from
Teradata, a division of NCR
SCI Creates Actionable Information
• Process the analytic results into actionable information in the
format and level appropriate for the operation
• Provide analytic results for 4 distinct audiences:
> Management - personalized for responsibility
• Performance metrics and trends for product, processes (including alerts
themselves) and lines of business.
> Analysts
• Performance metrics identifying exceptions and outliers.
• Predictive performance and opportunities based upon statistics.
• Specific reports that address points of interest
– Recalls, missing, new product, new processes, new facilities etc.
> Operations
• Reactive alerts (standards) – events that exceed standard
• Proactive alerts (critical) – product to be re-prioritized to prevent an issue
> Partners
• Late shipment reports, trend analyses
Solution Example
Ford’s Three-Pronged Solution
SCI Modeling & Segmentation
Supply Chain Modeling
• Utilize daily product position
and business requirements
snapshot
• Re-calculate projected
quantities and time via models.
Projected customer service
levels via variability analysis
• Comparison to expected
aggregated Demand by family,
SKU, path
FORD SUPPLY CHAIN EXAMPLE: PS&L Operation
Demand
SUPPLIERS
SUPPLIERS
(Count: 2,000)
CUSTOMERS
PACKAGERS
(Count: 7)
REDISTRIB.
CENTER
(Count: 1)
REGIONAL
DISTRIBUTION
(Count: 10)
DEALERS
(Count: 5,900)
• Detailed analysis of segment or
aggregate performance by
time, yield, capacity,
constraints, …….
• Model management at segment
levels include:
o Segment lead time
o Yield
o Split/merge/path selection
• Long term highly accurate
forecasts not required
• Track and tune standards over
time
REDISTRIBUTION
CENTER
Allocation
Segment
INTERVALS
Ship to In-Yard
REGIONAL
DISTRIBUTION
In-Yard to
Receipt
Receipt to
Stock-Keep
Critical Alerts – In Yard
Critical Alerts – In Yard
Over Standard Alerts
Inventory Visibility
SCI – From Two To Eight Opportunities
To Avoid Back Orders
Change
Ship
Date
None
SUPPLIERS
PACKAGERS
Change
Ship
Date
1. Receive
2.Process
1.Normal/
Critical
REDISTRIB.
CENTER
1.Receive
2.Stockkeep
3.Normal/
Critical
None
REGIONAL
DISTRIBUTION
1.Receive
2.Stockkeep
DEALERS
Standards Management
Parameter Management scores
actual cycle time vs. the current
model parameter to detect
segments with a poor fit.
The actual cycle time for this
O/D pair averages over 6
days, so the assumptions
for segment cycle time are
not modeled properly.
For the shipments through this
segment, the average cycle time is
compared to the standard.
The current model does take average
segment variability into account. Now
actual variability can be incorporated
into the model to drive accurate safety
stock
Drilling into the detailed data for
each segment supports outlier
identification and users can evaluate
how well the model fits each
segment within the supply chain.
New model parameter evaluated
for fit with actual shipment
transaction cycle times
Histogram shows the
distribution of shipment
cycle times as a
percentage of the total
shipments
All detailed data supporting the
evaluation is available in the drill.
0
2 NEW PDCs (location unknown)
3 NEW PDCs (location unknown)
ADDITIONAL USERS
2 NEW PDCs (location unknown)
ADDITIONAL USERS
NEW PDC (location unknown)
ADDITIONAL USERS
ADDITIONAL USERS
CHICAGO
SAN FRANCISCO
MEMPHIS
ATLANTA, DALLAS
DETROIT, LA, KANSAS CITY
PLANNERS
ANALYSTS
Ford: Growth in User Base
FORD USER COUNTS - ACTUALS THROUGH PROJECTIONS
350
300
250
200
Running
Web User Count
150
Running
Alert User Count
100
50
Jul-01 Aug-01 Sep-01 Oct-01 Nov-01 Dec-01 Jan-02 Feb-02 Apr-02 Jun-02 Oct-02 Apr-03 Aug-03 Feb-04 Sep-04
Ford – ROI
Benefits:

• The SCI Solution gives Ford the ability to
predict and prevent potential back orders not just react
• Powerful Analysis of “Every SKU, Every
Day” on hundreds of millions of dollars of
inventory
• Daily metrics and historical trending that
allow reality-based planning to be linked
with execution management




Improved Customer
Service levels
Less overtime,
expediting, and
special handling
Higher margins
No new data sources
required
Results:


• All members of the network now perform as
a synchronized team!
Reduced Inventories


10% One-Time and
Recurring Reductions
in Inventory
20% Back Order
Reduction
25-30% Reduced
“referrals”
30% Cycle Time
Reduction
Savings in first 6 months alone was five times the cost of the system.
Thank You for your Attention
Email:
[email protected]

similar documents