Class 8: Norton Auto Supply Role of RDC?

• Application of multi-period inventory models • Geographic location

• Location and lead time pooling • Fast customer service
• Distribution Strategies • Buffer
• Demand forecasting
Mfg. Plant External • Smooth demand for CDC
Suppliers
• Lowers transportation costs (than direct shipments to
CDC dealers)
• Hold safety stock for dealers
20 RDC
RDC’s – Lowers capital burden on dealers
– Lower safety stock: “Risk pooling” (location pooling)
Dealers

Fall 2009 Svenja Sommer 1 Fall 2009 Svenja Sommer 2

RDC: “Risk Pooling – through Locations” Why does location pooling work?

safetystock at warehouse 1
• Location pooling reduces =
N individual dealers safetystock at dealers N
demand uncertainty as 1
Periodic review – P Std. dev. – σ measured with the coefficient
Safety stock at each dealer =Z xσ x P+L of variation
• Reduced demand uncertainty
Total safety stock = N × Z × σ x P+L reduces the inventory needed 0.7
to achieve a target service
level 0.5
Replace N dealers by 1 single RDC :
• But there are declining
Periodic review – P marginal returns to risk 0.33
Std. dev. of total demand = σ⋅ N pooling!
Î Most of the benefit can
Safety stock at warehouse =Z x σ x N x P+L be captured by pooling
only a few territories. 1 2 4 9 N
Fall 2009 Svenja Sommer 3 Fall 2009 Svenja Sommer 4

Location pooling pros, cons and alternatives Why CDC?

• Pros:
– Reduces demand uncertainty which allows a firm to reduce inventory,
increase service, expand the product line, or a combination of all three. • Single point of contact for suppliers
• Cons: • Central supplier management
– Location pooling moves inventory away from customers:
• this creates an inconvenience for the sales reps. • Economies of scale in manufacturing & purchasing (quantity
• May create costs to ship product to customers, but may reduce inbound discounts?)
transportation because of consolidation.
• Alternatives: • Managing service to RDC’s
– Virtual pooling:
• Information control
• Each rep keeps her own inventory, but shares inventory with nearby
reps if needed. • Lower transportation costs (than direct shipments to RDC)
– Drop shipping:
• If a firm doesn’t have enough demand at each location to justify • Lower safety stock: “Risk pooling” (lead time pooling)
holding inventory, the firm can location pool with other firms via a
drop shipping firm, e.g., Alliance Entertainment holds inventory and
performs fulfillment for Circuit City’s online DVD store.
Fall 2009 Svenja Sommer 5 Fall 2009 Svenja Sommer 6

1
 CDC: Lead time pooling – consolidated distribution Consolidated distribution

Consolidated distribution reduces inventory via lead time risk pooling

Current system: direct from supplier
– Only need to decide the total quantity to ship from the supplier, not a total
8 week lead time quantity and its allocation across locations. Hence, some uncertainty is avoided.
RDC 1
.. – Most effective if demands are negatively correlated across locations.
Supplier .
– Most effective if the supplier lead time is long and the DC to store lead time is
RDC 50 short.
– But consolidated distribution increases total distance traveled and total lead
time from supplier to stores.
Proposed system: centralized inventory in a distribution center
1 week Other benefits of consolidated distribution:
lead time
RDC 1 – Easier to obtain quantity discounts in purchasing.
8 week lead time ..
Supplier CDC . – Easier to obtain economies of scale in transportation:
• DC may be able to order a full truck load every day, whereas an individual
RDC 50
store might requires much more time to order a full truckload)
Fall 2009 Svenja Sommer 7 Fall 2009 Svenja Sommer 8

CDC: Inventory Policy ? Exhibit 1: ABC Classification at Norton

Continuous Review Policy cumulative

Investment
Avg. length of a cycle
100%
= 1 month for “A” parts 90%
Q
=
R
Avg. weekly demand at CDC 55%
C
= 20 x avg. weekly demand at RDC
B

Std. dev. of weekly demand at CDC A

Number
= √20 x std. dev. of weekly demand at RDC of Part #
3.7% 16.4 %

Fall 2009 Svenja Sommer 9 Fall 2009 Svenja Sommer 10

SK 3809 – Current Policy: SK 3809 – Recommended Policy:

Q = one month’s demand at CDC 99% cycle service level ⇒ Z = 2.33

= 4 x 20 x 1300 = 104,000 units
ROP = avg. lead time demand + Z x std. dev. of LT demand

= 20 x 1300 x 1 + 2.33 x 100 x 20 x 1

ROP = Avg. lead-time demand + 4 weeks avg. demand
LT LT
= 1 x 20 x 1300 + 4 x 20 x 1300
= 26,000 + 1043
= 130,000 units = 27,043 Safety stock
Safety Stock: 104,000

Why is safety stock calculation NOT a good idea?

Fall 2009 Svenja Sommer 11 Fall 2009 Svenja Sommer 12

2
 SK 3809 – Recommended Policy: Costs

2 R S Annual ordering costs + Annual Inv. holding cost

Q= =19,447 ≈ 20,000
H Cycle Safety
SR Inv. Stock
H ⋅Q H x safety stock
where: Q
2
H = i x C = 0.26 x 2.75 SK 3809: Across all Parts:
Current policy: $112,840/yr $3.43 Million/yr
= $0.715/unit/year
Recommend: $14,650/yr $0.23 Million/yr
R = 20 x 1300 x 52 (annual demand)
• Difference = Savings by reducing Q & by reducing SL to 99%
S = $100 • Why ok? Stockout at CDC ≠ stockout at dealer; RDC holds safety stock!
Fall 2009 Svenja Sommer 13 Fall 2009 Svenja Sommer 14

RDC – Current Policy ? RDC Recommended Policy - SK 3809

Periodic review policy with P = 1 week, L = 1 week 98% - “Fillrate” Service Target
2% of demand is short
OUL = Avg. demand over (P+L) periods Avg. shortages/cycle
+ 2 weeks of demand = 0.02 x 1300 = 26 units

SK 3809 Safety Stock Avg. shortages/cycle = L(z) x σx P+L

OUL = 2 x 1300 + 2 x 1300 = 5200
26 = L(z) x 100 x √2
Safety Stock
26
Again: Why is safety stock calculation NOT a good idea? L(z) = = 0.184 ⇒ Z = 0.55
100 x 2
Fall 2009 Svenja Sommer 15 Fall 2009 Svenja Sommer 16

RDC Recommended Policy - SK 3809

OUL = Avg. demand over (P+L) + Z x σ x P+L

Inventory Cost of SK 3809 Across all Parts:
= 2 x 1300 + 0.55 x 100 x 2 at 1 RDC:
Current: $3253/year $894,842 / yr
= 2600 + 78 = 2678
Proposed: $1450/year $692,175 / yr
Safety Stock

Savings at 1 RDC: $1803

Average Inv = OUL – (R*P) / 2 = 2678-1300/2 =2028
Savings across 20 RDC ≅ $36,000 $202,667 / yr
(Norton owns its pipeline inventory!)

Fall 2009 Svenja Sommer 17 Fall 2009 Svenja Sommer 18

3
 “FedEx” Proposal

96% service off shelf Safety Stock = 11 units

3% by FedEx (24 hour) Reduction in safety stock

= 78 – 11 = 67 units
SK-3809
Inv. Cost savings = 0.715 x 67
96% Fillrate = $48/yr
4% of demand short
No. of Fed-Ex shipments
Avg. # of shortages/cycle = 0.04 x 1300 =52 / cycle = 0.03 x 52 x 1300 = 2028 units/yr
L(Z) = 0.369 ⇒ Z = 0.08
⇒ Probably not worth it for Item SK 3809 …

Fall 2009 Svenja Sommer 19 Fall 2009 Svenja Sommer 20

For what items would you use FedEx ? Norton Auto Supply

• light, small ⇒ low FedEx cost RDC CDC

Service Measure: Service Measure:
• high unit cost ⇒ high holding cost
Fillrate Cycle Service Level
• long lead-time Ordering Policy: Ordering Policy:
large safety stock
Periodic R.P. Continuous R.P.
• high uncertainty
• Lower demand variation at CDC • Lower safety stock
• Easier to coordinate • Economies of scale in
Note: could use Newsvendor fractile to determine optimal service level • Joined ordering of multiple production / transportation
before FedEx shipment: products from CDC • Take advantage of quantity
• excess cost = holding cost per cycle • Combine shipments to different discounts
• shortage cost = shipping cost per FedEx RDCs
Fall 2009 Svenja Sommer 21 Fall 2009 Svenja Sommer 22

Distribution Strategies Cross-docking: Moving products directly from inbound to

outbound vehicle without storage in between.
Shipping
Sources Sources Sources
Store T Store V Store W Store X Store Y Store Z

Mixed
pallets
Warehouse Dock

Roll-
Pallet Breaking & Assembly over
stock

Full pallets -
CLIENTS CLIENTS CLIENTS single
product
Warehousing Direct Cross-docking
(one or several) delivery (no inventory!)

Plant A Plant B Plant C Plant D Plant E

Receiving
Fall 2009 Svenja Sommer 23 Fall 2009 Svenja Sommer 24

4
 Cross-dock Warehouse Cross-docking Direct Ship

Cost Factors
• inventory High Low Low
• transportation Low (FTL) Low (FTL) High (LTL)
• facilities and High Medium Low
handling
• information Low High Medium
systems

Service Factors
• Response time to Fast Medium Long
end customers
• Product variety Expensive Easier Easier
• Total lead time Long Medium Low

Type of Product Low cost, low High demand, low Bulky, easily
obsolescence variability damaged items
Fall 2009 Svenja Sommer 25 Fall 2009 Svenja Sommer 26

Wal-Mart

• Staple stock - items that customers expect to find in the same

place in every Wal-Mart (toothpaste, shampoo, etc.).
– These items are warehoused so there are no stockouts.
– Pull based distribution
• Direct ship - items that Wal-Mart buyers have gotten a great
deal on and are pushing out to stores.
– These items are cross-docked; stores don't know what they're
getting until the truck arrives!
– Push-based distribution

Fall 2009 Svenja Sommer 27