PDD Chapter 9
PDD Chapter 9
PDD Chapter 9
Verbal description
Sketch
Photos and renderings
Storyboard
Video
Simulation
Interactive multimedia
Physical appearance models
Working prototypes
Matching the Survey Format with the Means of Communicating the Concept
The choice of survey format is tightly linked to the means of communicating
the product concept
Most concept test surveys first communicate the
product concept and then measure customer response.
When a concept test is performed early in the concept
development phase, customer response is usually
measured by asking the respondent to choose from two or
more alternative concepts. Additional questions focus on
why respondents react the way they do and on how the
product concepts could be improved. Concept tests also
generally attempt to measure purchase intent. The most
commonly used purchase-intent scale has five response
categories:
Definitely would buy.
Probably would buy.
Might or might not buy.
Probably would not buy.
Definitely would not buy.
If the team is simply interested in comparing two or more concepts,
interpretation of the results is straightforward. If one concept dominates
the others and the team is confident that the respondents understood
the key differences among the concepts, then the team can simply
choose the preferred concept. If the results are not conclusive, the team
may decide to choose a concept based on cost or other considerations,
or may decide to offer multiple versions of the product. Note that care
must be applied in making this judgment for cases in which
manufacturing costs are dramatically different among the concepts
under comparison and in which no price information is communicated to
the respondents. In such cases, respondents may be biased to select the
most costly alternative.
In many cases the team is also interested in estimating the demand for a
product in the period following launch, usually one year. Here we
present a model for estimating the sales potential of durables. By
durables we mean products that last several years, and for which there
is, therefore, a negligible repeat-purchase rate. These products are in
contrast to consumer packaged goods, like razor blades, toothpaste, or
frozen food, for which forecasting models must consider rates of trial
and subsequent repeat purchase.
We estimate Q, the quantity of the product expected to be sold
during a time period, as
Q=N xAxP
N is the number of potential customers expected to make
purchases during the time period. For an existing and stable
product category N is the expected number of purchases to be
made of existing products in the category over the time period.
A is the fraction of these potential customers or purchases for
which the product is available and the customer is aware of the
product. (In situations where awareness and availability
are assumed to be separate independent factors, they are
multiplied together to generate A.)
P is the probability that the product is purchased if available and if
the customer is aware of it. P is estimated in turn by
Fdefinitely is the fraction of survey respondents
indicating in the concept test survey that they would
definitely purchase (often called the “top box” score).
Fprobably is the fraction of survey respondents
indicating that they would probably purchase (often
called the “second box” score).
Cdefinitely and Cprobably are calibration constants usually
established based on the experience of a company
with similar products in the past. Generally the values
of Cdefinitely and Cprobably fall in these intervals: 0.10 <
Cdefinitely < 0.50, 0 < Cprobably < 0.25. Absent prior
history, many teams use values of Cdefinitely = 0.4 and
Cprobably = 0.2.
Note that these values reflect the typical bias of respondents to
overestimate the probability that they would actually purchase the
product.
The primary benefit of the concept test is in getting
feedback from real potential customers. The qualitative
insights gathered through open-ended discussions with
respondents about the proposed concepts may be the
most important result of concept testing, especially early
in the development process.
The team benefits from thinking about the impact of the
three key variables in the forecasting model:
(1) the overall size of the market, (2) the availability and
awareness of the product, and (3) the fraction of
customers who are likely to purchase. In reflecting on the
results of the concept test, the team should ask two key
diagnostic questions:
First, was the concept communicated in a way that is likely
to elicit customer response that reflects true intent?
Second, is the resulting forecast consistent with observed
sales rates of similar products?