Transcript Quick Response in Manufacturer

Quick Response
in Manufacturer-Retailer
Channels
Ananth V. Iyer • Mark E. Bergen
Presented By
Evren Körpeoğlu
Quick Response
 Quick Response is a strategy focusing
on providing shorter lead times.
 It enables orders to be placed closer
to the start of the selling season.
Quick Response & Apparel Industry
 Used mainly in Apparel
Industry, because of;
 Long lead times between order
placement and delivery of
product.
Quick Response
 To Reduce the lead times;
Information Sharing using EDI
Logistics improvements
Increased use of Air Freight
Point of sale scanners and
barcoding
 Improved manufacturing
methods such as laser fabric
cutting or modular sewing cells




Is it Pareto Improving?
 Do both the manufacturer and the
retailer benefit from QR?
 This paper investigates the effect of
QR and how to make it become
Pareto improving.
Model Structure
 Old System:
 L = Lead Time
(5 to 8 months)
 QR System:
 L2 = Lead Time (2 to 5 months)
 L2≤L
 L2 = L – L1
Demand Structure
 We have two levels of demand
uncertainity:
 Uncertainity of the number of people
who come to the store and their
propensity to buy
 Uncertainity about mean demand
Demand Structure before QR
 N(θ,σ2): distribution of demand during season
assuming that we know meand demand
N(μ,τ2): mean demand estimation
 m(x)~N(μ,τ2+σ2) → demand distribution
time 0
at
Demand Structure after QR
 Data collected during period L1 is used to
estimate the demand of the season.
 If the demand in period L1 is d1, then:
m(x|d1)~N(μ(d1), σ2 + 1/Ρ)
 
 d1
 (d1 )  2 2  2 2
   
2
2

1

2

1

2
QR and the Channel
Model Parameters:
c: Cost per unit to buy product from manufacturer
∏: Goodwill cost per unit
h: holding cost per unit of product left over at the
the end of the season
w: production cost per unit
r: retailer’s revenue per unit
The Old System
Expected
Revenues
Q


Q
EP(Q)   rxm( x)dx   rQm ( x)dx
Q


Q
 h  (Q  x)m( x)dx    ( x  Q)m( x)dx  cQ
Holding Costs
Goodwill Costs
The Old System
Optimum initial inventory:
I old    Z ( s )   
2
Optimal service level (s):
r  c
s
r  h
2
The Old System
Maximum Expected Profit:
EPold  (r  c)   {( c  h) Z ( s )  (r  h   )br ( Z ( s ))}  2   2
The Old System
Expected Quantity Sold:
I old  sold 
I old


I old
 xm( x)dx 
2
2
I
m
(
x
)
dx



b
(
Z
(
s
))



r
 old
Expected Quantity Left over:
I old left 
I o ld
2
2
(
I

x
)
m
(
x
)
dx

(
Z
(
s
)

b
(
Z
(
s
)))



r
 old

QR System
 Retailer’s inventory choice involves two steps:
1. Observe the demand during L1 and use it to estimate
seasonal demand
2. Choose optimal inventory to maximize expected
profits with estimated demand. The demand before
season,d1, directly affects this choice.
QR System
Optimum initial inventory:
EI QR    Z ( s )  
2
1

Maximum Expected Profit:
EPQR  (r  c)   {( c  h) Z ( s )  (r  h   )br ( Z ( s ))}  
2
1

QR System
Expected Quantity Sold:
EI QR sold    br ( Z ( s ))  
2
1

Expected Quantity Left over:
EI QRleft  ( Z ( s )  br ( Z ( s )))  
2
1

Old System vs QR System
Since:
    
2
2
 EIQR-sold ≥ EIold-sold
 EIQR-left ≤ EIold-left
2
1

Old System vs QR System
Expected Manufacturer Profit in old system:
EPold mfr  (c  w){  Z (s)    }
2
2
Expected Manufacturer Profit in QR system:
EPQR mfr  (c  w){  Z ( s )  
2
1

}
When is QR Pareto Improving?
Lemma 1:
 If s≤0.5 then under QR we have an increase
in manufacturer and retailer expected
profits because Z(s) < 0
 Thus, under low service levels, QR is pareto
improving.
When is QR Pareto Improving?
In the case that manufacturer pays a salvage
credit per unit, if this credit is large enough,
then QR will be pareto improving even if
s≥0.5
EPQR mfr  (c  w)   ( wZ ( s )  cbr Z ( s ))  
2
1

}
When is QR Pareto Improving?
Lemma 2:
 If s≥0.5 then under QR we have an increase
in expected retailer profits but a decrease in
expected manufacturer profits
Other Methods...
 Commitments using service level
 Commitments regarding the
wholesale price
 Volume commitments across products
Commitments using service level
 After QR, retailer may use a higher
service level but it is not beneficial for
the manufacturer
 Since the increase in service level can
not be shown in the model, the
goodwill costs are changed.
Commitments using service level
Theorem 1. If
1  (1 /(1  d 2 ))
s≥0.5, d=σ/τ and y 
1  (1 / d 2 )
then increasing ∏ to ∏’=(s’(r+h)+c-r)/(1-s’)
and s’ such that Z(s’)=Z(s)/y makes QR
Pareto improving.
Commitments using service level
Lemma 3. If
1  (1 /(1  d 2 ))
s≥0.5, d=σ/τ and y 
1  (1 / d 2 )
then increasing ∏ to ∏’=(s’(r+h)+c-r)/(1-s’)
and s’ such that Z(s’)=Z(s)/y, results in
EIQR-sold(s’) ≥ Iold-sold and EIQR-left (s’)≤ Iold-left
Commitments using service level
Other ways in using service level;
 Contractual commitment between
manufacturer and retailer to provide high
service level
 Use of cooperative advertising that uses
service level guarantee
 Increasing the goodwill cost
Wholesale Price
Quantity Discounts:
 If manufacturer forces retailer pay c1>c
if the quantity purchased is smaller than
the Iold, and c if it is larger than Iold,
manufacturer may benefit from QR.
Lemma 4:
There exists a quantity discount scheme
that makes QR Pareto improving
Wholesale Price
Lemma 5:
There exists a wholesale price c* that
depends on system parameters such that if
c≥c*, then there exists no flat wholeslae
price commitments that can make QR
Pareto improving.
Volume Commitments across
Products
 Assuming we have M products, the
buyer should have a commitment that
the total amount of purchased
products will be as much as before
the QR.
Volume Commitments across
Products
Lemma 6:
The impact of a volume commitment of I(M),
which is initial inventory across M products, on
expected profit for the retailer is given by
EPQR  (r  c) M  (c  h) I ( M )  (r  h   ) M  2 
1


 
z
( z )br ( z )dz
Thus a volume commitment of
M(μ+Zs√σ2+ τ2) across M products makes
QR Pareto improving.
Volume Commitments across
Products
Volume commitment maintains
manufacturer’s profit but genrate an
expected service level which is below
that is required for service level
commitments
Volume Commitments across
Products
Lemma 7:
The expected service level s(M) under a
volume commitment of MIold across M
products, s(M), is s≤S(M)≤s’, where is s’ is
defined as in Theorem 1
Conclusion
 Impact of a channel view on Quick
Response for fashion apparel industry
is examined.
 Classical inventory model are used to
show these effects.
 Three ways for Pareto improving are
consider: service level, price and
volume commitments.
THANK YOU FOR LISTENING!...
Comments & Questions