Online Mass Customization and the Customer Experience
Arnold Kamis, Bentley CollegeMarios Koufaris, Baruch CollegeTziporah Stern, Baruch College
NJIT - November 22, 2004
Mass Customization
The ability to manufacture individually customized products and services on a mass production scale without a significant cost increase.Online examples: NikeID Reflect.com Dell
Online interfaces
Alternative based: Select from all possible customized product
options
Attribute based: Select from all customizable product
attributes to customize product
Question based: Answer relevant questions and have
product customized for you
Research questions
What is the best interface? Simple vs. Powerful
How much variety to offer? Is more choice always good?
Who are the customers/users? Are all users having the same
experience?
Theoretical background
Behavioral decision-making: Discovered Preferences (Payne et al. 1988;
Coupey 1994; Fennema et al. 1995) Rational model – Optimal choice exists Less realistic in customer decision-making
Constructed Preferences (Payne et al. 1992; Rosenshein 2000) No objective optimal choice Preferences dynamically constructed What is best decision support tool for self-
customization?
Theoretical background
Task-technology fit model (Goodhue, 1995): Appropriate technology for specific user
tasks increases performance and satisfaction
Also important are user characteristics What is appropriate technology
(interface) for online mass customization task?
Study variables
Measure success of system through customer experience:
Perceived ease of use Perceived usefulness Perceived control Enjoyment
Effect of the interface: alternative vs. attributeEffect of variety: Number of possible product versionsUser characteristics: Computer anxiety and computer playfulness
The Interface
Alternative-based: Customers must deconstruct product into attributes and
values May inflate perceived variety (3 attributes x 5 values = 125
alternatives) Not very interactive, just pick one from the list
Attribute-based: Product already deconstructed Actual variety obvious (3 attributes x 5 values = 15 choices) Interactive, experimental, game-like
Hypothesis: Overall, online customers will experience higher (perceived control, perceived usefulness, enjoyment) with an attribute-based MC interface than with an alternative-based one
Variety: Is more choice always better?
Hypothesis: Overall, enjoyment of online MC customers will follow an
inverted U-shaped curve as variety increases.
Satisfaction Regret
Positiveness of consumption experience
Variety
Source: Desmeules, 2002
Variety and Interface
Attribute-based interface reduces task complexity and cognitive effortHypotheses:Enjoyment will start decreasing at a higher variety for users who use the attribute-based MC interface than for those who use the alternative-based one.(Perceived ease of use, perceived control) will decrease in the alternative-based interface but will remain constant in the attribute-based interface as variety increases.
The User
Computer anxiety and computer playfulness will moderate the effect of interface on customer experienceHypotheses:
Overall, the difference in (perceived control, perceived usefulness, perceived ease of use, enjoyment) between using the attribute-based and using the alternative-based interfaces will be larger for users with low computer anxiety than for those with high computer anxiety.
Overall, the difference in (perceived control, perceived usefulness, perceived ease of use, enjoyment) between using the attribute-based and using the alternative-based interfaces will be larger for users with high computer playfulness than for those with low computer playfulness.
Methodology
2x2x3 factorial design Interface: Alternative – Attribute Product: Watch – Backpack Variety: 8 – 54 – 150 possible product
versions
Web pages created by researchers from NikeID.com and Factory121.com with no brand identifiers
Methodology
Baseline experience: Subjects used all three interfaces (including
a question-based one) at 54-scale to customize a candle
Then randomly assigned to one of 12 conditionsTold that they may win customized watch or backpack (incentive)At the end, filled out survey on their experience and themselves
Data Analysis
329 subjects recruited online through Zoomerang’s MarketToolsNationally representative sampleRan CFA with AMOS 4.0 to test validity and reliability of survey instrumentRan ANCOVAs to test hypotheses Covariates: Product Involvement and
Web Skills
ResultsInterface Perceived
Control Perceived Usefulness
Enjoyment
Alternative 5.226 (0.092) 3.833 (0.115) 4.179 (0.107) Attribute 5.913 (0.090) 4.725 (0.113) 5.112 (0.105) Mean differences are significant at p<0.001
Hypotheses supported: Overall, attribute-based interface provides a better customer experience than alternative-based interface.
ResultsEnjoyment
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
5
5.1
8 54 150
Variety
Est
imat
ed M
arg
inal
Mea
ns
Main effect of variety significant at p<0.01
Hypothesis supported: Enjoyment follows inverted U-shaped curve as variety increases
Interaction effect of interface and variety not significant
ResultsPerceived Control
4.5
4.74.9
5.1
5.35.5
5.7
5.96.1
6.3
8 54 150
Variety
Est
imat
ed M
arg
inal
Mea
ns
Alternative
Attribute
Interaction effect of interface and variety significant at p<0.01
Hypothesis supported: Perceived control decreases as variety increases in alternative-based interface but remains constant in attribute-based interface
ResultsPerceived Ease of Use
4.8
5
5.2
5.4
5.6
5.8
6
6.2
6.4
8 54 150
Variety
Est
imat
ed M
arg
inal
Mea
ns
Alternative
Attribute
Interaction effect of interface and variety significant at p<0.05
Hypothesis supported: Perceived ease of use decreases as variety increases in alternative-based interface but remains constant in attribute-based interface
ResultsEnjoyment
3
3.5
4
4.5
5
5.5
6
Low High
Computer Anxiety
Es
tim
ate
d M
arg
ina
l M
ea
ns
Alternative
Attribute
Perceived Usefulness
33.23.43.63.84
4.24.44.64.85
Low High
Computer Anxiety
Es
tim
ate
d M
arg
ina
l M
ea
ns
Alternative
Attribute
Categorized computer anxiety into low-medium-high and excluded medium
Interaction effect of interface and computer anxiety significant at at least p<0.05
Hypotheses partially supported: Computer anxiety moderates effect of interface on (a) enjoyment and (b) perceived usefulness (but not perceived control or perceived ease of use)
ResultsPerceived Control
5.2
5.4
5.6
5.8
6
6.2
6.4
6.6
Low High
Computer Playfulness
Alternative
Attribute
Perceived Ease of Use
4
4.5
5
5.5
6
6.5
7
Low High
Computer Playfulness
Alternative
Attribute
Categorized computer playfulness into low-medium-high and excluded medium
Interaction effect of interface and computer playfulness significant at p<0.05 for p. control and significant only at variety=150 for p. ease of use
Hypotheses partially supported: Computer playfulness moderates effect of interface on (a) perceived control and (b) perceived ease of use at high variety levels (but not perceived usefulness or enjoyment)
Contributions
Theory/Research: Attribute-based decision-making tool best for
constructed preferences Appropriateness of attribute-based tool increases
with task complexity Verified core ideas of TTF in context of online mass
customization, including often ignored user characteristics
Reduction in cognitive effort in online user tasks may outweigh reduction in physical effort
Computer anxiety affects perceptions on outcome while computer playfulness affects perceptions on process
Contributions
Practice/Industry: Increase in variety of products without
appropriate tools to make decisions can have negative effect
Not all customers are the same – More powerful interface is often also more complex, overwhelming some customers
Since computer anxiety affects perceptions regarding outcome, always provide ways to backtrack or undo actions
Future work
Part of a larger study Look at question-based interface Impact on behavioral intention Other user characteristics: personal
innovativeness in IT
Suggestions for further research Study other products, especially experiential
ones with non-visual attributes Examine variety at a more detailed level
Questions?Comments?
Marios KoufarisDepartment of Statistics and Computer Information
SystemsZicklin School of Business
Baruch College, [email protected]