How Does Product Recovery Affect Quality Choice? Gil Souza, Indiana University (with Atalay Atasu, Georgia Tech) 1 Quality Choice and Product Recovery.

Quality Choice and Product Recovery - Atasu & Souza

1

How Does Product Recovery Affect Quality Choice?

Gil Souza, Indiana University(with Atalay Atasu, Georgia Tech)

Presented at:


2

Some Self Promotion

• Released in April 2010• Self contained chapters• Written with managers

in mind• Can be used in MBA

classroom


3

Agenda1. Strategic vs. tactical decisions on CLSCs

1. Example of a Strategic Problem: Network Design

2. Motivation: examples of firms that practice product recovery and how that impacts their product quality design

3. Research question4. Related literature5. Model structure and basic assumptions6. Structural results7. Extension 1: Competition8. Extension 2: Differentiated remanufacturing9. Conclusion


4

Strategic vs. Tactical Decisions in CLSCsLevel Examples

Strategic

•Network design: location and size of collection centers, aggregation centers, reman. facilities, etc.•Should the OEM engage in remanufacturing?•Pricing of new and remanufactured products•Impact of recovery on product design

Tactical

•Acquisition of returns: when, how many, at which price, at which quality?•Disposition decision: what should a firm do with a product return: remanufacture, dismantle for parts, recycle, etc.?

Operational•Disassembly planning: sequence and depth•Job shop control after disassembly

October 2004 Gil C. Souza 5

Reason for return % of Returns

Product defective 20%

Could not install 28%

Not compatible with user needs

40%

Convenience (better price elsewhere, remorse, etc.)

12%

Example of Network Design Problem: Consumer Returns at HP

Why Do Consumers Return Products?


What Happens to Commercial Returns?

Distributor Retailer SalesFactory

Evaluation of returns

Returns

RemanufacturingSecondary market

Returns

Sales

New returns

Disposal

•Large delays in reverse chain (e.g., 60-90 days for remanuf. of printers)

•This is critical for time-sensitive returns (e.g. printers)


Queuing Network Model of Commercial Product Returns Supply Chain

Distributor Retailer

Sales

Factory

Evaluation of returns

eReturns

pr

Remanufacturing

r

m

r

(1 ) rp

(1 ) rp r d sf

c Customer

Consumption

r

r

Note: Nodes are modeled as queuesp = proportion of “new” returns = net demand rate (demand per day)r = returns ratei = processing rate at node i

Sales to secondary market2

(1 ) rp

Reverse chain: M/M/1 queuesForward chain: M/G/ queues

Gil C. Souza 8

HP Deskjet Printers: Value of One Day

Portion of Network Value of One Day (US$)Between the customer and evaluating facility

$93,797

Between the evaluating facility and remanufacturing

$72,475

Between remanufacturing and the secondary market

$79,489

•Large opportunities in the reverse network because of existing delays a reduction of current 40-day delays to 20-day delays implies $1.5 million savings over life cycle of printer!

Bosch Power Tools: Value of One Day

Portion of Network Value of One Day ($)Between the customer and evaluating facility

$5,624

Between the evaluating facility and remanufacturing

$11,623

Between remanufacturing and the secondary market

$12,748

•Number are much lower than HP’s (particularly considering very long life cycles for power tools)

•It is clear that Bosch needs an efficient reverse supply chain, whereas HP needs a responsive reverse supply chain


10

Quality Choice and Product Recovery: The Interface Between Product Recovery and Product Design


11

Motivation 1: Undifferentiated Remanufacturing - Xerox

• Newly manufactured copiers contain a significant amount of remanufactured modules– We denote this case here a “remanufactured” copier

• A higher quality module (e.g., faster paper feeding) increases a consumer’s WTP for the product (but more expensive to produce)

• If Xerox can remanufacture that module (when the product comes back), it can decrease its average production cost for the product

Paper InputModule

Finishing Module

Imaging Module

Xerox iGEN 3


12

Motivation 2: Differentiated Remanufacturing (Cummins)

• Remanufactured engines (or parts) are sold at an average discount of 35% relative to new - imperfect substitutes for new engines (or parts)

• A more powerful turbocharger, for example, increases consumers’ WTP for the engine but that turbocharger is more expensive to produce

• If Cummins can remanufacture the turbocharger, it can reduce its average remanufacturing cost for the engine

• Cummins said that remanufacturability impacts product design

Before Remanufacturing

After Remanufacturing


13

Motivation 3: Recycling • Costly recycling: European Recycling Platform (ERP, a

consortium founded by HP, Sony, Braun and Electrolux) pays 7.5¢/kg to recycle small appliances in Europe – Recycling here mandated by take-back legislation– The link between design quality and recycling is the sales

volume of new appliances • Example: increasing quality increases WTP; you may then increase

prices, reducing sales volume (and consequently future recycling volume & associated costs)

• Profitable recycling– Cell phones recycling can be profitable (Geyer and Doctori

Blass 2009)


14

Research Question• Take-back and reprocessing (recycling or

remanufacturing) of used products by an OEM is driven by two factors:1. Take-back legislation (mandated)2. Value recovery (mostly economic reasons)

• How does the possibility of product recovery impacts the design quality of a product?

– Quality here is interpreted as an observable performance measure that increases consumers’ valuation for the product


15

Related Literature• Closed-loop supply chains: why engage in

remanufacturing?– Atasu et al. (MS 2008), Debo et al. (MS 2005), Ferguson and

Toktay (POM 2006), Guide & Van Wassenhove (OR 2009), Hauser & Lund (2003)

• Closed-loop supply chains: economic impact of take-back legislation on product design– Atasu & Subramanian (2009), Plambeck & Wang (MS 2009)– Our paper differs here in that our design variable is quality—it

impacts consumer valuation for new product• Marketing: quality choice models

– Mussa & Rosen (1978), Moorthy (1988), Desai (2001)– Difference is that we consider product recovery

16

Model: Basic Assumptions• Monopolist, single period, single product

– We consider duopoly extension later– Single period: slice of a “steady-state” infinite horizon

• Consumer’s valuation for quality ~ U[0,1]


u s p

Quality Price

Consumer type

Net utility

u 0Market size = 1

0 1ps

1p

qs

Consumers with positive utility


17

Three Recovery Types• Quality recovery: quality inducing

components are recovered and (re)sold again. Profitable. Typical example: remanufacturing

• Profitable material recovery: recovery of quality inducing components not possible, but recovery overall is profitable. Typical example: profitable recycling [of cell phones]

• Costly recovery: recovery overall is not profitable; mostly done due to legislation. Typical example: costly recycling [WEEE]


18

Basic Assumptions II: Recovery• A fraction (exogenous or decision variable) of total

products sold (in previous period) can be recovered• If recovery is in the form of quality recovery:

– is the fraction of demand q met with recovered products (1 - is fraction of demand met with new)

– This is the undifferentiated remanufacturing case– We have extended the model to the differentiated

remanufacturing case; later• If recovery is in the form of recycling (costly or profitable)

– All demand q is met with new products– is proportion of end-of-use products q recycled


19

Basic Assumptions III: Costs (Quality Recovery case)

• Unit variable cost of producing a brand new product with quality s is (Moorthy 1988):

• Average variable cost of production to meet total demand q with recovery level t is – Where is average cost of quality with

recovery – r is cost savings from remanufacturing of quality-

inducing components– n is remanufacturing cost that is independent of the

quality choice

2nc c s

2mc s n

mc c r


20

Basic Assumptions IV: Costs (Recycling case)

• Recycling can be thought of a particular case of remanufacturing where r = 0 (no quality recovery)

• That means and thus avg cost is – If recycling is profitable, n < 0– If recycling is costly, n > 0

mc c 2cs n


21

Basic Assumptions V: Additional Convex Increasing Collection Cost

• Literature suggests diseconomies of scale in recovery costs as recovery volume increases (Ferguson and Toktay 2006)

• We also have empirical evidence from two data sets

North American Chemical Company European Consumer Electronics Firm


22

Our General Model Is Then:

2 2

, , ( ) ( )max m

p sq p c s n q

Avg cost of productionConvex collection (recovery) cost

mc c r Notation reminder

Non-negative sales quantitys.t. 1 0 p

qs

Non-negative price 0p

Logical constraint 0 1

Firm choosesPrice p Quality sCollection rate


23

Benchmark Result: No Recovery (Moorthy 1988)

2

, 1max

p s

pp cs

s

* 13NRs

c

* 29NRp

c

* 127NR c


24

Result 1.1 (Legislation Case: Exogenous)

• The optimal quality level with quality recovery or costly recycling (n > 0) is higher than the non-recovery case:

• Due to take-back costs, take-back legislation provides an incentive for firms to increase the quality of their products: they are able to charge higher prices, decreasing consumption (and thus environmental impact)

* *NRs s


25

Result 1.2 (Legislation Case: Exogenous)• The optimal quality level with profitable recycling (n

< 0) is: – Higher than the non-recovery case ( ) if– Lower than the non-recovery case ( ) if

• With profitable recycling and low enough , collection costs are low enough so that recycling overall is profitable– Firm has an incentive to choose a lower quality level,

charge a lower price, and sell more• With high enough , legislation accomplishes its

goals of reducing environmental impact through higher quality products, lower consumption

* *NRs s

* *NRs s


26

Result 2.1 (No Legislation Case: is a Decision)

• The optimal quality level with quality recovery is no lower than the non-recovery case:

• With costly recycling, • Only difference from the case with take-back

legislation (exogenous ) is that for high enough recovery cost n, the optimal collection rate is zero ( ), implying

* *NRs s

* 0 * *NRs s

* *NRs s* 0


27

Result 2.2 (No Legislation Case: is a Decision)

• The optimal quality level with profitable recycling (n < 0) is: – Same as non-recovery case ( ) if– Lower than the non-recovery case ( ) if

• Bottom line: for any type of product recovery, regulated markets for product recovery (setting a minimum recovery rate ) perform better—higher product quality, lower consumption—than in unregulated markets

* *NRs s n n

* *NRs s n n


28

Environmental Impact of During Life Cycle with Recovery

• Denote ex as environmental impact per unit during product life cycle stage x. – Production: new + rem.– Use: – End of life: fraction going to landfill =

• Total environmental impact– EI = production + use + end-of-life–

(1 ) ( )p r p re q e q e eue q

(1 )eole q

(1 ) (1 )

( )p r u eolEI e q e q e q e q

q E e Increasing in q and decreasing in


29

Environmental Benefits of Recovery Compared to No-Recovery Case

Recovery Scenario Environmental Benefits

Quality recovery +Profitable Recycling - or +

Costly Recycling +

Remember that consumption (q) is always lower under quality recovery and costly recycling, but higher under profitable recycling


30

Impact on Consumer Surplus ( = 0)

Recovery Scenario

Impact on Manufacturer Profits and Consumer Surplus

Quality recovery +Profitable Recycling

+

Costly Recycling -

*

1/

Consumer Surplus2p s

s p d

0 1/p s

Consumer surplus and manufacturer’s profits are aligned (if collection costs are linear in recovery rate)

Consumers who buy


31

Extension 1: Duopoly• Consider two identical firms with the same

cost structure• One possible equilibrium is symmetric: firms

price at marginal cost, and obtain zero profit• We focus on asymmetric case, where one firm

chooses a high quality (sh) at price ph, and the other a low quality (sl) at price pl

• Thus, consumer of type gets utility from firm i

i i iu s p


32

Quantities Sold By Each Firm

l

l

ps

h l

h l

p ps s

0 1

( )

l

l h

qu u

( )

h

h l

qu uNothing

( 0)lu

Consumer type θ


33

Problem Faced by Each Firm i

2

,max ( )

s.t. , 0i i

i i i m ip s

i i

q p c s n

p q

Again, solve sequentially: for given si and sj, each firm optimizes its prices. Then, substitute pi(si,sj) into above, and find optimal si, sj


34

Results• Can find closed form expressions for prices as a

function of qualities• However, solution of optimal quality levels can only

be found numerically• Results confirm monopoly results

– With quality recovery or costly recycling, both firms chose a quality level that is higher than under no recovery: s* > sNR

– Quality level can be lower than no-recovery case if recovery is in the form of profitable recycling

– Firm’s profits (and consumer surplus) are higher than no-recovery case with profitable recovery

35

Impact of Recovery on Firm Quality under Competition


Quality recovery

Costly recovery

Profitable material recovery

Quality of low-quality firm

Quality of high-quality firm

Quality gap between firms increases with recovery

Quality gap between firms decreases with recovery


36

Extension 2: Differentiated Remanufacturing (Quality Recovery)

( )r ru s p

n nu s p Consumer’s utility for a new product of quality s and price pn

Consumer’s utility for a product of quality s that is remanufactured and sold at price pr

Where is the consumer’s perceived “quality gap” between a new and remanufactured product. This results in:

1 n rn

p pq

n r n rr

p p p pq

s


37

Optimization Problem: Differentiated Remanufacturing

Can show that, just like in the undifferentiated remanufacturing case:

* *NRs s

2 2 2

, , ( ) ( )max

n r

n n r r np p s

q p cs n q p c r s q

s.t r nq q

, , , 0n r n rp p q q


38

Conclusions: Impact of Product Recovery on Quality Choice

• Firms typically design products with higher quality when there is product recovery (compared to the case of no recovery)

• Exception occurs if there is profitable material recovery at relatively lower recovery rates mandated by legislation

• Results are robust

How Does Product Recovery Affect Quality Choice? Gil Souza, Indiana University (with Atalay Atasu, Georgia Tech) 1 Quality Choice and Product Recovery.

Documents

commercial returns

returns ratei

consumer returns

gil souza

proportion of new returns

day delays

evaluating facility

timesensitive returns