Concept Testing Flashcards
Definition and Purpose
Research on concept may begin with qualitative approach
A quantitative research phase follows measuring consumers’ reaction to a
proposed product on multiple dimensions:
* Likelihood of purchase
* Perceived importance of the product
* Perceived quality of the product
Concept Test phase produces both
* a sales volume forecast and
* diagnostic information to guide the positioning in the
marketplace.
Executing a Concept Test - Overview
In addition to the usual sample selection issue, major executional
considerations are
1. The Concept Communication Method
a. Tone
b. Mode
2. The Consumer Response measured
a. Intended Purchase Measures
b. Overall Product Diagnostics
c. Special Attribute Diagnostics
d. Respondent Profiling Variables
Typically, data
to collect falls
in 4 classes
Tone: Factual vs. Persuasive Approach
1.a Tone: Factual vs. Persuasive Approach (I)
factual &
non-emotional ↔️ Emotional &
Persuasive
F:
“A low-calorie form of peanut
butter that can be used in most
diets.”
P:
“A marvelous new way to chase
the blahs from your diet has been
discovered by General Mills
scientists – a low-calorie version of
ever-popular peanut butter. As
tasty as ever and produced by a
natural process, our New Light
Peanut Butter will fit every weight
–control diet in use today virtually
without restrictions
F: Eliciting evaluation of concept –
core idea (+)
Unlike what consumer sees in
market place (-)
P:
Eliciting evaluation on concept
and communication strategy
➡️ Yields better behavioral
predictions
Mode: Words vs. Visuals
There are 2 modes and a combination in communication:
* Words only
* Visuals only
* Words & Visuals
There is no general rule as to which is better
* Alternative electives among MBA students words ➡️ words
* Designer clothing item ➡️ Visual
The Concept Communication Method
- None of the 6 cells dominates the others
- Important to recognize the impact of concept type on respondents’
reactions, in particular ‘purchase intent’ (PI) scores - What do you think increases PI scores?
Of course: Richer mode, richer scores - That is why you cannot compare scores generated
from different concept tests
Pfizer and Clairol tested identical concepts in different modes:
Mode of Concept Test
Words only
Pfizer A
20%
Pfizer B
35%
Clairol
33%
Mode of Concept Test
Words plus Visual
Pfizer A
46%
Pfizer B
52%
Clairol
50%
Difference 26% 17% 17%
Without recognition of mode, A would be better than B, when in fact B
dominates A in both modes.
The Consumer Response measured
Intended Purchase Measures cover
Intended Purchase Measures cover
* Purchase Intention (PI)
* Expected Frequency
PI is included in virtually every concept test
Purchase Intention (PI), typically queried:
“Based on this product description, how likely would you to be to buy
this product if it were available at a store in your area?”
Definitely
would buy
Probably
would buy
Might or might
not buy
Probably
would not buy
Definitely
would not buy
A 5-point-scale is most common, however 6-, 7-, and 11-point scales
are also often used.
Expected Frequency
* For nondurable goods, frequency is key to volume.
* PI is good indicator of trial, but volume forecast
requires knowledge of consumption habit or
special occasion item.
Expected Frequency, typically queried:
“Which statement best describes how often you think
you would buy this product if it were available to you?”
Once (or
more) per
Week
Once per
2 to 3
Weeks
Once per
month
Once per
2 to 3
months
Once per
4 to 6
months
Circa once
per year
Less than
once per
year
Never
For example, Coke measures in days per week
Following cases are usually addressed when appropriate:
* Different sizes
* Multiple units might be bought
Then, sales are calculated in the following way:
Sales volume
per household
in time period
=
% households in market who try
* Expected # purchases for triers
* Expected # units per purchase
Concept tests
usually try to
measure all
three
variables
The Consumer Response measured
Overall Product Diagnostics
Managers want to obtain data to understand
why purchase measures turn out the way they do
Standard battery of questions with respect to overall product
judgments:
- Uniqueness (differentiation from other products)
- Believability ➡️Most widely used, but may still generate low PI
- Importance in solving a customer’s problem
- Inherent interest
- Value for the money (here, price introduced with 5-point scale)
The Consumer Response measured
Specific Product Diagnostics
- When a concept has a number of attributes or benefits
offered, we need to know individual contribution to PI. - “You said that you [PI of respondent]. What is it specifically about the
product which makes you feel this way?”
It is often useful to collect
* perceptions of specific attributes
* their importance to the customer
S19 F11
Specific Product Diagnostics (III)
Schwartz (1989) suggests doing a quadrant analysis:
s20
Quadrant 1:
Consumer does not care
about
Quadrant 2:
Key communication
attributes important, and
product des well
Quadrant 3: Consumer does not care about
Quadrant 4:
Problem Quadrant:
improvement efforts
S21
S22
The Consumer Response measured
Respondent Profiling Variables
Type of consumers who respond in different ways
Demographics most obvious, but not necessarily diagnostic
1. Current purchase behavior
2. Perception of the category
3. Barriers to changing the brand
4. Influence in actual purchase occasion
For example, high satisfaction with current brand makes a switch to a
new brand less
Interpreting PI Scores (I)
- PI is always at heart of the concept test
- Suppose, low calorie peanut butter test yields:
1. Definitely would buy 15%
2. Probably would buy 45%
3. Might or might not buy 20%
4. Probably would not buy 10%
5. Definitely would not buy 9%
Is this a bad or good PI score?
What sales volume would you expect?
Interpreting PI Scores (II)
- Cannot be answered looking at these numbers.
- General rules of thumb on “good” scores exist:
- “… a concept test should receive 80% to 90% favorable answers (top 2
boxes) to encourage subsequent development work.” (Taylor et al.
1975)
Schwartz (1987) states following average scores across all product
categories:
* Definitely would buy 19%
* Probably would buy 64%
=
83%
Suppose, low calorie peanut butter test yields:
1. Definitely would buy 15%
2. Probably would buy 45%
= 60%
3. Might or might not buy 20%
4. Probably would not buy 10%
5. Definitely would not buy 9%
Our peanut butter fails both rules!
PI scores vary substantially across product categories:
S27
Note: It is only a rule of thumb, no more
Industry specific norms from:
* Published sources
* Firm’s own files
* Research firm files
Relation Between PI and Actual Purchase (I)
- Research shows strong correlation between PI and trials
- Top box good predictor of trial rate in consumer-packaged goods
S29 F12
- PI statements in 3
samples - 6 weeks later check of
purchase, where 19%
actually bought
product at least once
Relation Between PI and Actual Purchase (II)
- PI is not a good predictor of repeat purchase
- requires substantial change of habit
- does not account for variety seeking behavior
Relation Between PI and Actual Purchase (III)
S31
Relation Between PI and Actual Purchase (IV)
S32
Relation Between PI and Actual Purchase (IV)
S33
Summary
- Concept testing is crucial research process for new products
- Key tool for setting development priorities prior to major
investments to create product/service - Key guidelines:
- Select tone/mode appropriate to situation
- Interpret PI scores in light of tone/mode selection and appropriate
benchmarks - Design the test to afford diagnostic as well as predictive
information
