Transcript presentation source

American Lighting
Products
 What drives inventory in this system?
 Safety stock at the MDC’s
 Shipping quantities to the MDC’s
(Truckload)
 Production lot sizes
 Number of products
 ...
Inventory for a product
 What is average system inventory for a
product?
Net System Inventory
Objective
 “Set for each product item at each MDC
based on historical sales levels”
 “The NSO is the reorder point for the
factory to produce another lot of product”
 “the Average system inventory for a
product is the NSO + 1/2 a lot size”
Is this Consistent?
 “The NSO is the reorder point for the
factory to produce another lot of product”
 “the Average system inventory for a
product is the NSO + 1/2 a lot size”
Inventory
System Inventory
NSO
Production
Time
Inventory
At an MDC
Base
Stock
Level
Truckload
Production begins
Time
What is Base stock level at
MDC?
 Expected “Lead Time” Demand
 Demand from time MDC hits base stock level
to time first truck load arrives
 Production times
 delivery lead times
 time till system inventory hits NSO
 Safety stock to cover for variability
What can distribution do?
 Lot sizes?
 Expected lead time demand?
 Can consolidation affect this? How?
 Safety stock?
Safety Stock
 Customer Service level
 95% to 98%
 1.64 to 2.05
 Variance in lead time demand
 Estimating effect of consolidation
 Square root law
 Poisson Demand
 Forecast Error
Square Root Law
(For Safety Stock!)
 Two independent random variables
 X with mean  and variance 2
 Y with mean  and variance 2
 What about X+Y?
 2
 2 2
 Standard Deviation of “Consolidated MDC”
 2 
Poisson Demand




If we don’t know variance…
Assume Poisson demand
variance = mean
Std Deviation of Lead Time Demand =
square root of Average Lead Time
Demand
Forecast Error
 Forecast errors due to variability of
demand
 Forecast Accuracy




Product at MDC <50%
Product 50-70%
Product Family 70-90%
Market 90-100%
What is the Safety Stock?
 Poisson Model
 At the very worst all 700 products have the
same volume
 Safety stock < 2*700*sqrt(Volume/700)
 That gives total safety stock of <2 million
 That’s less than 8% of inventory!
Production Lead Time
(Again)
To MDC 1
To MDC 3
Truck load
Time
Reducing Average Lead
Time Demand
 Base Stock at MDC depends on how many
MDCs carry the same product
 With 8 MDCs
 Expect to wait till the fourth truck
 With 4 MDCs
 Expect to wait till the 2nd truck
What is Base Stock Level?
 Safety Stock
 Expected Lead Time Demand
 Inbound lead time < 7 days
 Guess at time to load a truck





Total Volume = nearly 5,000 truckloads
2,500 truckload per plant
50 truckloads per week per plant
7 truckloads per day
Difference between 2nd truck and 4th is a matter
of hours.
Conclusions
 Safety Stock < 8% of inventory
 Expected lead time demand insensitive to
consolidation
 Distribution cannot get close to 20%
inventory reduction without affecting
customer service levels
How to Reduce Inventory
 Smaller lot sizes at the factory
 Spread out 3 weeks maintenance and
vacations
 Trim product line
Critique
 What did we ignore?
 Where else can Distribution look for
inventory reductions?
NSO vs Base Stock
 Time till System reaches NSO after MDC
reaches base stock
 Interested in the
 Mean?
 std. Deviation?
Guesses
 Mean is zero
 Std. Dev. grows with the number of MDCs
(simple simulation supports this)
NSO effects
 How big is the effect?
 Can we change replenishment policy to
eliminate effect?
 Can we reduce effect with consolidation?
What is NSO?
 Estimate NSO due to this delay
 Probably significant because we have such
poor service levels (as low as 78%) with
such high inventory levels.
 How to get better numbers?
 Ask (not an option here)
 Analysis
 Simulation