© Petersen and Kumar

Perceived Risk, Product Returns, and Optimal Resource Allocation:

Evidence from a Field Experiment

J. Andrew PetersenAssistant Professor of Marketing and Assistant Director of the Center for Integrated Marketing and Sales (CIMS)

– Kenan-Flagler Business School, University of North Carolina at Chapel Hill

V. Kumar Regents’ Professor, Chang Jiang Scholar, Richard and Susan Lenny Distinguished Chair & Professor of

Marketing, Executive Director, Center for Excellence in Brand & Customer Management (CEBCM), and Director of the Ph.D. Program in Marketing – J. Mack Robinson College of Business, Georgia State University, Atlanta GA

Applying Field Experimentation to Behavior ResearchMarch 14 – 15, 2014

Field Experiments Marketing Conference

© Petersen and Kumar

Economics of Product Returns

2

Current Cost

$100 Billion in Reverse Logistics

Reduces profit by around 3.8% per retailer

About 6% of all products are returned(Some categories with rates > 25%)

© Petersen and Kumar

Why Return Products?

3

*Source: Lawton (2008)

1• “No Trouble Found” (68%)

2• “Buyer’s Remorse” (27%)

3• “Defective” (5%)

A study found that 95% of all product returns occurred for two main reasons:

© Petersen and Kumar

Who Returns Products?

Firm% of Customers Who Returned a Product

or Filed a Complaint

High-tech B2B Firm 64%

Catalog Apparel Retailer 70%

General Merchandise Retailer 75%

Financial Services Firm 77%

4

© Petersen and Kumar

What are Firms Doing?

5

Product Return Policy Decisions

Firms often make product return policies more strict through changes in return timing (e.g. 30 day limit), restocking fees (e.g. 15%), etc.

Firms are spending fewer marketing resources on customers who return products (Petersen and Kumar 2009) or not even considering product returns at all

Resource Allocation Decisions

© Petersen and Kumar

What are Firms Doing? (cont.)

6

Resource Allocation Strategy

No Formal

Allocation

Model

Based on

Rank-

order

Use an

ORA

Algorithm

Customer

Value

Measure

No Formal Measure 2 0 0 2 (3.5%)

RFM/PCV 11 12 0 23 (41%)

CLV w/o Product Returns 5 5 1 11 (20%)

CLV w/ Net Buying 10 7 1 18 (32%)

CLV w/ Product Returns 0 0 2 2 (3.5%)

28 (50%) 24 (43%) 4 (7%) 56

Results of a Survey of 56 Managers of Retail Firms

© Petersen and Kumar

How Do Customers React?

7

Do Product Return Policies and Product Return Behaviors Impact Customer Purchase Behaviors?

Used as a signal for purchase (Nasr-Bechwati and Siegal 2005)

Higher leniency = higher purchase rates (Wood 2001)

Free-based (versus fee-based) product returns generate more future purchases (Bower and Maxham 2012)

The more a customer returns (to a threshold), the more a customer purchases in the future (Petersen and Kumar 2009)

© Petersen and Kumar

Cost and Perceived Risk

8

The Role of Product Returns

Increases costs through reverse logistics and losses of revenues from purchases

Lowers perceived risks which positively influences future purchase behavior

© Petersen and Kumar

Key Research Questions

9

How do resource allocation strategies affect firm profit when firms:

1. Do not manage product return costs or perceived risk?

• No Formal Measure/RFM/CLV w/o Product Returns

2. Manage product return costs, but not perceived risks?

• CLV w/ Net Buying

3. Manage product return costs and perceived risks?

• CLV w/Product Returns

© Petersen and Kumar

Data

10

• B2C firm which sells footwear, apparel, and accessories through the Internet and mail order catalogs

Source

• A cohort of 935 customers who made their first purchase in Quarter 2 of 2003 is used for the model validation stage

Model Validation

• A random sample of 26,000 customers is used for the field experiment stage

Field Experiment

• Lenient return policy (100% money back with no time limit)Return Policy

© Petersen and Kumar

Measurement

11

1• No Formal Measure (Control Group)

2• Firm Strategy (RFM-based)

3• CLV w/o Product Returns (Benchmark Model 1)

4• CLV w/ Net Buying (Benchmark Model 2)

5• CLV w/ Product Returns (Proposed Model)

We use 5 different objective functions based on the types of strategies firm use for resource allocation:

We need to develop objective functionswhich can help firms optimally allocate resources

© Petersen and Kumar

Measurement (cont.)

12

T

tt

itittiti

r)(1

Marketing)sπ(PurchaseipRelationshPCLV

11,0, )(MODEL 3

MODEL 4

T

t

titi PCLV1

t

itit1,0,

r)(1

Marketing)asesπ(NetPurch)ipRelationsh(

MODEL 5

T

t

titi PCLV1

t

ititit1,0,

r)(1

Marketing)π(Returns)sπ(Purchase)ipRelationsh(

MODEL 1

No Formal Model

MODEL 2

Firm Strategy (RFM-based)

© Petersen and Kumar

CLV Model Comparison

13

MODELMarketing

(μ = 22.29)

NetPurchases

(μ = 142.55)

Purchases

(μ = 173.16)

Returns

(μ = 30.61)

No Cost/No Risk

(without Returns)

3.68

(16.5%)-

27.37

(15.8%)-

Cost/No Risk

(Net Purchases)

3.60

(16.2%)

46.28

(33.0%)- -

Cost/Risk (Proposed)

(with Returns)

3.57

(16.0%)-

27.21

(15.7%)

6.81

(22.2%)

MODELMarketing

(μ = 17.11)

NetPurchases

(μ = 133.70)

Purchases

(μ = 162.81)

Returns

(μ = 29.11)

No Cost/No Risk

(without Returns)

3.61

(21.1%)-

27.91

(17.1%)-

Cost/No Risk

(Net Purchases)

3.60

(21.0%)

45.82

(34.3%)- -

Cost/Risk (Proposed)

(with Returns)

3.54

(20.7%)-

27.35

(16.8%)

6.94

(23.8%)

Out-of-Sample Model Fit

In-sample Model Fit

© Petersen and Kumar

Experimental Design

14

Measurement 1(May 2009)

Treatment (May - July)

Observation (August - October)

Measurement 2(October 2009)

O1a O1b

O2a x2 O2b

O3a x3 O3b

O4a x4 O4b

O5a x5 O5b

Each group is made up of 5,200 customers (n = 26,000)

O1 = Control O2 = Firm Strategy O3 = No Cost/No Risk

O4 = Cost/No Risk O5 = Cost/Risk

© Petersen and Kumar

Field Experiment Results

15

Customer Group

Total Profit

from Purchases

in 6 Months

Total Profit Lost

from Product

Returns in 6

Months

Total Catalogs

and Emails Sent

in the First 3

Months

Average Profit

Per Customer

During the 6

Months

Average CLV

Per Customer

Post Study

Control $1.60M $403.20k 0 $235.20 $1,088.89

Firm’s Strategy $1.62M $373.20k 19,760 $240.55 $1,087.03

No Cost/No Risk

(without Returns)$1.71M $290.16k 17,199 $273.48 $1,172.17

Cost/No Risk

(Net Purchases)$1.80M $270.21k 14,664 $294.21 $1,223.16

Cost/Risk (Proposed)

(with Returns)$2.02M $201.16k 12,090 $352.77* $1,402.96*

* The values found for average profit per customer during the 6 months and average CLV per customer post study for the Proposed Model are statistically

significantly larger from each of the other segments

© Petersen and Kumar

Field Experiment Results (cont.)

16

Relative Gains for the Proposed Model

MODELProfit from

Purchases

Profit Lost from

Product Returns

Marketing

Costs

Average

Profit Per

Customer

Average

CLV Per

Customer

Control 26.3% -50.1% N/A 51.5% 28.8%

Firm’s Strategy 24.7% -46.1% -38.8% 46.6% 29.1%

No Cost/No Risk

(without Returns)18.1% -30.7% -29.7% 28.5% 19.7%

Cost/No Risk

(Net Purchases)12.2% -25.6% -17.6% 19.1% 14.7%

* The percentages in each cell represent the benefit the Proposed Model strategy provides the firm relative to the other strategy

© Petersen and Kumar

Correlations Across Models

17

What happens if we apply the proposed model to the other samples?

MODEL Firm’s Strategy Benchmark Model 1 Benchmark Model 2

Proposed Model 0.03 0.27 0.42* The correlations represent the pairwise correlation between the marketing resources spent on a customer using the strategy from the Proposed Model

(left column) and the strategy used on the field study group (top row).** All correlations are significant at p < 0.01

© Petersen and Kumar

Implications

18

Product returns play a significant role in lowering perceived risk and should not just be a managed cost

Managing costs and perceived risks can lead to both short- and long-term gains in firm profitability

Applying this framework as a field experiment provides significant value to managers

© Petersen and Kumar

Thank You!!!