U.S. patent application number 13/600778 was filed with the patent office on 2014-03-06 for methods and apparatus to forecast new product launch sourcing.
The applicant listed for this patent is Kyle A. Gerhart, Yue Xiao. Invention is credited to Kyle A. Gerhart, Yue Xiao.
Application Number | 20140067466 13/600778 |
Document ID | / |
Family ID | 50188711 |
Filed Date | 2014-03-06 |
United States Patent
Application |
20140067466 |
Kind Code |
A1 |
Xiao; Yue ; et al. |
March 6, 2014 |
METHODS AND APPARATUS TO FORECAST NEW PRODUCT LAUNCH SOURCING
Abstract
Methods and apparatus are disclosed to forecast new product
launch sourcing. An example method includes identifying shared
attributes between the new product and a plurality of existing
products in the target market, calculating theoretical
co-penetration values between the attributes shared between the new
product and at least one of the plurality of existing products,
calculating actual co-penetration values between the attributes
shared between the new product and at least one of the plurality of
existing products, calculating an attribute distance value between
corresponding ones of the theoretical and actual co-penetration
values, and calculating a percent volume of the new product
expected to be sourced from one of the plurality of existing
products based on the attribute distance value.
Inventors: |
Xiao; Yue; (Palatine,
IL) ; Gerhart; Kyle A.; (Chicago, IL) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Xiao; Yue
Gerhart; Kyle A. |
Palatine
Chicago |
IL
IL |
US
US |
|
|
Family ID: |
50188711 |
Appl. No.: |
13/600778 |
Filed: |
August 31, 2012 |
Current U.S.
Class: |
705/7.31 |
Current CPC
Class: |
G06Q 10/04 20130101;
G06Q 30/0201 20130101; G06Q 30/0202 20130101 |
Class at
Publication: |
705/7.31 |
International
Class: |
G06Q 30/02 20120101
G06Q030/02; G06Q 10/04 20120101 G06Q010/04 |
Claims
1. A method to determine an effect of introducing a new product in
a target market, comprising: identifying, with a processor, shared
attributes between the new product and a plurality of existing
products in the target market; calculating, with the processor,
theoretical co-penetration values between the attributes shared
between the new product and at least one of the plurality of
existing products; calculating actual co-penetration values between
the attributes shared between the new product and at least one of
the plurality of existing products; calculating an attribute
distance value between corresponding ones of the theoretical and
actual co-penetration values; and calculating a percent volume of
the new product expected to be sourced from one of the plurality of
existing products based on the attribute distance value.
2. A method as defined in claim 1, further comprising calculating a
substitutability index between the new product and the one of the
plurality of existing products based on the attribute distance
value.
3. A method as defined in claim 2, wherein the substitutability
index is based on a degree of polarization associated with at least
one of the shared attributes.
4. A method as defined in claim 3, wherein the substitutability
index weights the degree of polarization by the attribute distance
value.
5. A method as defined in claim 3, wherein the degree of
polarization comprises an inverse Dirichlet parameter associated
with the theoretical co-penetration values.
6. A method as defined in claim 1, wherein calculating the
theoretical co-penetration values comprises estimating a Dirichlet
model associated with attributes of the plurality of existing
products in the target market.
7. A method as defined in claim 1, wherein the shared attributes
are associated with a product category.
8. A method as defined in claim 1, wherein the shared attributes
comprise at least one of a brand, a product type, a size, a feature
or a flavor.
9. A method as defined in claim 1, further comprising calculating a
substitutability index between the new product with the one of the
plurality of existing products based on distance values for all
attributes.
10. An apparatus to determine an effect of introducing a new
product in a target market, comprising: a product category
comparator to identify shared attributes between the new product
and a plurality of existing products in the target market; a
Dirichlet modeling engine to calculate theoretical co-penetration
values between the attributes shared between the new product and at
least one of the plurality of existing products; an empirical
co-penetration engine to calculate actual co-penetration values
between the attributes shared between the new product and at least
one of the plurality of existing products; a distance calculator to
calculate an attribute distance value between corresponding ones of
the theoretical and actual co-penetration values; and a volume
sourcing calculator to calculate a percent volume of the new
product expected to be sourced from one of the plurality of
existing products based on the attribute distance value.
11. An apparatus as defined in claim 10, further comprising a
substitutability engine to calculate a substitutability index
between the new product and the one of the plurality of existing
products based on the attribute distance value.
12. An apparatus as defined in claim 11, wherein the
substitutability engine is to base the substitutability index on a
degree of polarization associated with at least one of the shared
attributes.
13. An apparatus as defined in claim 12, wherein the
substitutability engine is to weight the degree of polarization by
the attribute distance value.
14. An apparatus as defined in claim 12, wherein the degree of
polarization comprises an inverse Dirichlet parameter associated
with the theoretical co-penetration values.
15. An apparatus as defined in claim 10, wherein the Dirichlet
modeling engine is to estimate a Dirichlet model associated with
attributes of the plurality of existing products in the target
market.
16. A tangible machine readable storage medium comprising
instructions that, when executed, cause a machine to, at least:
identify shared attributes between the new product and a plurality
of existing products in the target market; calculate theoretical
co-penetration values between the attributes shared between the new
product and at least one of the plurality of existing products;
calculate actual co-penetration values between the attributes
shared between the new product and at least one of the plurality of
existing products; calculate an attribute distance value between
corresponding ones of the theoretical and actual co-penetration
values; and calculate a percent volume of the new product expected
to be sourced from one of the plurality of existing products based
on the attribute distance value.
17. A machine readable storage medium as defined in claim 16,
wherein the instructions, when executed, cause the machine to
calculate a substitutability index between the new product and the
one of the plurality of existing products based on the attribute
distance value.
18. A machine readable storage medium as defined in claim 17,
wherein the instructions, when executed, cause the machine to base
the substitutability index on a degree of polarization associated
with at least one of the shared attributes.
19. A machine readable storage medium as defined in claim 18,
wherein the instructions, when executed, cause the machine to
weight the degree of polarization by the attribute distance
value.
20. A machine readable storage medium as defined in claim 18,
wherein the instructions, when executed, cause the machine to
associated an inverse Dirichlet parameter with the theoretical
co-penetration values.
21. A machine readable storage medium as defined in claim 16,
wherein the instructions, when executed, cause the machine to
estimate a Dirichlet model associated with attributes of the
plurality of existing products in the target market.
22. A machine readable storage medium as defined in claim 16,
wherein the instructions, when executed, cause the machine to
associate the shared attributes with a product category.
23. A machine readable storage medium as defined in claim 16,
wherein the instructions, when executed, cause the machine to
calculate a substitutability index between the new product with the
one of the plurality of existing products based on distance values
for all attributes.
Description
FIELD OF THE DISCLOSURE
[0001] This disclosure relates generally to market research, and,
more particularly, to methods and apparatus to forecast new product
launch sourcing.
BACKGROUND
[0002] In recent years, market researchers have strived to predict
the success of products introduced into one or more markets. Market
researchers expect that new products that are introduced into a
market will source from other products that already reside in the
market. In some circumstances, new products may source from
existing products of the same manufacturer, which may result in
undesirable cannibalization.
BRIEF DESCRIPTION OF THE DRAWINGS
[0003] FIG. 1 is a schematic illustration of a system to forecast
new product launch sourcing in accordance with the teachings of
this disclosure.
[0004] FIG. 2 is an example theoretical co-penetration matrix
generated by the example system of FIG. 1.
[0005] FIG. 3 is an example actual co-penetration matrix generated
by the example system of FIG. 1.
[0006] FIG. 4 is an example distance matrix generated by the
example system of FIG. 1.
[0007] FIG. 5 is a flowchart representative of example machine
readable instructions which may be executed to forecast new product
launch sourcing.
[0008] FIG. 6 is a schematic illustration of an example processor
platform that may execute the instructions of FIG. 5 to implement
the example systems and apparatus of FIGS. 1-4.
DETAILED DESCRIPTION
[0009] Market researchers seek to gain competitive advantage by
bringing new products to a market. When bringing a product into the
market, a portion of the volume of the new product is expected to
source from (e.g., replace sales of) other products that already
participate in that market. However, when the newly introduced
product diverts sales from another product also produced and/or
otherwise manufactured by the same manufacturer and/or supplier,
some of the net competitive advantage realized by introduction of
the new product is lost. This loss of sales is sometimes referred
to as cannibalization because a manufacturer's own product is
"consuming" (e.g., replacing) sales of another product of that same
manufacturer.
[0010] Market researchers typically expect some degree of
cannibalization to occur when introducing a product (e.g., a new
product having some similarities to existing competitive products,
an existing product introduced to a particular market geography for
the first time, etc.). However, methods, apparatus, systems and/or
articles of manufacture disclosed herein identify consumers that
will likely participate in such cannibalization by buying a new
product in the market. As such, the market researchers may realize
and/or otherwise appreciate whether the introduced product steals
its own buyers (e.g., from its own products) from other products in
the manufacturer's portfolio, or whether the introduced product
steals buyers from competing manufacturers as hoped.
[0011] Typical estimations of sourcing behavior rely on fair share
sourcing estimates that expect a newly introduced product to source
from other products in a manner that is directly proportional to
existing competitive product shares in the market. For example, in
the context of automobile sales in the United States, Ford.RTM. has
the highest share of automobile sales at the time of this writing.
Fair share sourcing logic for a current Mercedes Benz.RTM. owner
would identify that the next car to be purchased by that consumer
would most likely be a Ford.RTM. automobile because Ford.RTM. is
the automobile company having the largest market share of
automobile sales. Market researchers familiar with typical
purchasing behavior of current Mercedes Benz.RTM. owners would
identify that such fair share sourcing logic is not an accurate
representation of expected future behavior. In other words, for
some product types and/or categories, reliance upon fair share
sourcing logic yields estimates that are too generalized for
practical use for market researchers. Accordingly, example methods,
apparatus, systems and/or articles of manufacture disclosed herein
deviate from the generalized errors associated with fair share
sourcing by, in part, focusing on cross purchasing within
attributes. Attributes may include, but are not limited to brands
(e.g., attribute levels of Coke.RTM. vs. Pepsi.RTM.), flavors
(e.g., attribute levels of cherry vs. lime), size, feature (e.g.,
attribute levels of powder, liquid, etc.), and diet classifications
(e.g., attribute levels of diet soda vs. regular soda).
[0012] Products of interest to a market researcher may have any
number of associated attributes. Each such attribute may include
any number of mutually exclusive attribute levels, and every
product will be mapped to only one attribute level within a
corresponding attribute. For example, every product typically has
an associated brand (e.g., an attribute). Example attribute levels
of the brand attribute include, for a soft drink product, a name
brand such as Coke.RTM., Pepsi.RTM., Sprite.RTM., Dr. Pepper.RTM.,
etc. In the event the product of interest is a 12-oz aluminum can
of Coke.RTM., then the attribute level is Coke.RTM., which is
mutually exclusive to all other attribute levels in the brand
attribute. For example, a Coke.RTM. product cannot also be a
Sprite.RTM. or Dr. Pepper.RTM. product. Continuing with the example
product of interest of a 12-oz aluminum can of Coke.RTM., a
container attribute would have an attribute level of aluminum can.
Other attribute levels for the container attribute may contain, but
are not limited to plastic bottle, glass bottle, etc.
[0013] Dirichlet theory utilizes and improves upon pure fair share
sourcing logic to generate theoreticals (estimate values,
parameters) indicative of where an introduced product will source,
as described by Goodhardt, Ehrenberg, and Chatfield (see "The
Dirichlet: A Comprehensive Model of Buying Behavior," Journal of
the Royal Statistical Society, Series A, Number 147, pp. 621-55,
1984), which is hereby incorporated by reference in its entirety.
When provided with consumer purchase summary data (e.g.,
Nielsen.RTM. panelist data, Nielsen.RTM. Homescan.RTM. data, etc.),
the Dirichlet model estimates parameters indicative of the consumer
repeat buying behavior (e.g., such as loyalty and/or purchase
frequency). Additionally, estimation of the Dirichlet model yields
probability functions to facilitate calculation of theoretical
co-penetration. Co-penetration refers to a percentage of a
population that purchases a pair of products during a time period
of interest. When considering a product for introduction into an
existing market (e.g., a line extension of an existing product),
example methods, apparatus, systems and/or articles of manufacture
disclosed herein apply a Dirichlet analysis for each major (e.g.,
statistically relevant, as deemed by businesses, etc.) attribute to
estimate Dirichlet parameters (Dirichlet S parameters) and
parameters that conform to a negative binomial distribution (NBD)
(NBD-k, NBD-a). These parameters may indicate how each attribute
level polarizes a category, and they may be used to calculate a
theoretical cross purchase pattern (e.g., switching patterns)
between pairs of attribute levels. Switching patterns (e.g., cross
purchasing patterns) between pairs of attribute levels may quantify
to what degree each attribute and/or attribute level (e.g., liquid
detergent attribute level, powder detergent attribute level, color
safe attribute level, hypoallergenic attribute level, etc.)
polarizes a product category (e.g., soft drinks, laundry detergent,
etc.).
[0014] Average polarization of levels within an attribute provides
an indication of the polarization of the corresponding attribute.
Some attributes and/or attribute levels are associated with
differing degrees of polarization. For example, attribute levels
associated with brand typically exhibit stronger polarization
behaviors as compared with attribute levels related to container
(e.g., the choice between Coke.RTM. and Pepsi.RTM. (brand) is much
more polarizing than the choice between cans and bottles
(container)). In the event an attribute polarizes consumer behavior
to a relatively high degree (e.g., relatively high when compared to
other attribute parameters, as shown below), then switching (e.g.,
co-purchasing) among different levels of the attribute will be
relatively low. The level of polarization for each attribute may be
identified by the Dirichlet S parameter. A relatively small
Dirichlet S parameter value is indicative of high polarization
(e.g., low switching tendency), while a relatively large Dirichlet
S parameter is indicative of low polarization (e.g., high switching
tendency). A theoretical co-penetration of product attributes may
be calculated in a manner consistent with example Equation 1.
P.sub.inj=P.sub.i+P.sub.j-P.sub.i.andgate.j Equation 1.
[0015] In example Equation 1, P.sub.i.andgate.j reflects a
theoretical co-penetration of product attribute level pair i and j,
P.sub.i reflects a theoretical penetration of product attribute
level i, P.sub.j reflects a theoretical penetration of product
attribute level j, and P.sub.i.andgate.j reflects a theoretical
penetration of product attribute level i or product attribute level
j. Each penetration value in example Equation 1 is derived from
corresponding NBD distributions. While the Dirichlet model
calculates and/or otherwise exposes theoretical co-penetration
values between one or more pairs of attribute levels of interest,
actual co-penetration purchase behavior may deviate from
theoretical calculations. In some circumstances, the underlying
reliance upon fair share sourcing techniques employed by the
Dirichlet model causes deviation from empirical observations. Such
deviations may be more pronounced and/or otherwise significant in
view of particular brands, markets and/or product types. To
determine a degree of substitutability between product attributes
in a more reliable and/or industry-expected manner than is
otherwise generated by the Dirichlet model, example methods,
apparatus, systems and/or articles of manufacture disclosed herein
ground the theoretical co-penetration values with empirical values.
The gap between theoretical and empirical co-penetration may be
leveraged to forecast sourcing of new product launches.
[0016] FIG. 1 is a schematic illustration of an example system 100
to forecast new product launch sourcing. In the illustrated example
of FIG. 1, the system includes a product sourcing engine 102
communicatively connected to a panelist database 104, a product
reference library (PRL) 106 and a new product attribute database
108. The example product sourcing engine 102 also includes a
product selector 110, a product category comparator 112, a
Dirichlet modeling engine 114, an empirical co-penetration engine
116, a distance calculator 118, a substitutability engine 120 and a
volume sourcing calculator 122.
[0017] In operation, the example product sourcing engine 102
invokes the example product selector 110 to identify a product of
interest that is to be considered for entry into a target market in
which it has not previously participated. As used herein, a
"product of interest" or a "new product" refers to a new
combination of previously existing attribute levels, which may be
related to one or more products that have no market presence in a
market of interest (but may have a market presence in a separate
market from the target market location of interest). In some
examples, the example new products may be line extensions of
already existing products, which have a degree of attribute
similarity to one or more products that already participate in the
example target market. As discussed above, the degree of similarity
between such products (e.g., in view of attribute levels and the
interaction therebetween) may determine where the new product will
source in the target market. In the event the new product sources
from other products in the target market that are also provided by
the same manufacturer/supplier, (i.e., cannibalization effects) the
manufacturer/supplier may not realize a net economic benefit in the
target market.
[0018] The example product category comparator 112 automatically
identifies existing products in the target market of interest that
are in a similar or same category as the new product. This
identification is qualified by comparing information characterizing
the new product (e.g., attributes) to information characterizing
existing products (e.g., attributes). Information related to the
new product of interest may be stored in the example new product
attribute database 108 and/or the example PRL database 106. The
example PRL database 106 may include, but is not limited to the
Nielsen.RTM. TDLinx.RTM. database and/or the Nielsen.RTM. PRL.RTM.
that includes product specific attribute information such as, but
not limited to product name, manufacturer name, brand, packaging
type, product size, flavor, lot number, serial number, nutritional
information, features and/or corresponding universal product codes
(UPCs). The Nielsen.RTM. PRL.RTM. presently codes more than 700,000
items, in which each item includes an average of forty (40)
descriptive characteristics/attributes/attribute levels.
[0019] For example, if the new product is a baby diaper brand, then
the example product category comparator 112 identifies one or more
products from the PRL database 106 having the same/similar category
of "baby products," and/or "diapers." Additionally, the example
product category comparator 112 identifies one or more subgroups of
products associated with the category that may be a closer match
when considering target market performance, such as "baby diapers"
rather than "adult diapers." In other examples, a new laundry
detergent product may reside in a "laundry detergent" category
(e.g., a highest level attribute), and the example product category
comparator 112 may further tailor a subgroup of candidate products
with which to compare by identifying an attribute level of "liquid
detergents" versus "powder detergents," depending on the type of
new product form. Depending on the new product of interest to be
considered for a target market, the example product category
comparator 112 may identify any number of attribute levels for use
in identifying candidate comparative products that currently
participate in the target market.
[0020] The example product category comparator 112 obtains
available purchase frequency data and/or other target market data
associated with the identified currently available comparative
products from the example panelist database 104. The example
panelist database 104 may include, but is not limited to the
Nielsen.RTM. Homescan.RTM. database and/or point-of-sale (POS) data
from retailers and/or merchants. The example Dirichlet modeling
engine 114 estimates Dirichlet parameters (Dirichlet S parameters)
for the attribute levels of products identified by the example
product category comparator 112, and calculates theoretical
co-penetration values in a manner consistent with example Equation
1 above. As discussed above, while the Dirichlet analysis provides
an indication of co-penetration between products and/or between
attributes levels of products, the Dirichlet analysis may not
accurately fit marketing expectations for one or more product types
and/or product categories.
[0021] To improve the application of Dirichlet techniques, the
example empirical co-penetration engine 116 of FIG. 1 retrieves
product market activity data from the panelist database 104 to
calculate an actual co-penetration without Dirichlet techniques. In
particular, if the actual co-penetration between
attributes/attribute levels deviates from theoretical predictions
based on fair share sourcing influences associated with Dirichlet
techniques, then attribute interaction effects may be observed. The
degree of such interaction effects may depend upon a direction
and/or magnitude of the deviation, which is indicated by a distance
value. The example distance calculator 118 calculates the distance
value between the theoretical values and the empirical values to
determine a degree of substitutability. Generally speaking, the
closer two attribute levels are to each other, such attribute
levels are deemed to be more substitutable to each other. The
example substitutability engine calculates a substitutability index
between the new and existing products based on the distance values
between attribute levels to generate a weighted average of
distances between a new product for the target market and existing
products. The example volume sourcing calculator determines volume
sourcing value(s) for the new product based on the substitutability
index, as described in further detail below, which provides
information indicative of where sourcing may occur in the target
market.
[0022] In the illustrated example of FIG. 2, a theoretical
co-penetration matrix 200 between different laundry detergent
attribute levels is shown. The example matrix 200 includes a row
201 having an example baby attribute level 202, an example base
attribute level 204, an example cold attribute level 206, an
example color attribute level 208, an example delicate (DELIC)
attribute level 210, an example hypoallergenic (HYPO) attribute
level 212, an example odor attribute level 214, an example pure
attribute level 216, an example scent attribute level 218, an
example softening (SOFT) attribute level 220 and an example stain
attribute level 222. The example matrix 200 also includes a column
224 including the same example attribute levels (202-222) residing
in the example row 201 to facilitate comparisons between the
different attribute levels, in which a zero cross-matrix diagonal
226 is populated with zero as a reflection of attribute levels
compared against themselves. A lower half 228 of the example matrix
200 is not populated in the illustrated example of FIG. 2 because
it is a symmetric matrix. The values in the illustrated matrix 200
of FIG. 2 indicate a theoretical degree of co-penetration (e.g.,
based on the Dirichlet model) indicative of a percentage of a
household in a selected population that purchased both a product
having the row 201 attribute level and a product having the column
224 attribute level.
[0023] In the illustrated example of FIG. 3, an empirical
co-penetration matrix 300 between the different laundry detergent
attribute levels of FIG. 2 is shown. The example empirical matrix
300 of FIG. 3 includes similarly numbered elements as is shown in
FIG. 2, but with a three-hundred series prefix.
[0024] Generally speaking, the example empirical matrix 300 of FIG.
3 includes co-penetration values that differ from the example
theoretical co-penetration values of the matrix 200 of FIG. 2.
Differences between theoretical co-penetration and empirical
co-penetration illustrate details of how different attribute levels
may interact. The deviation between the actual co-penetration
values and the theoretical co-penetration values are characterized
by a distance in a manner consistent with example Equation 2.
D ( A i , A j ) - C T ( A i , A j ) C A ( A i , A j ) . Equation 2
##EQU00001##
[0025] In example Equation 2, D(A.sub.i, A.sub.j) represents an
attribute A having i.sup.th and j.sup.th levels and their relative
distances therebetween, C.sub.T(A.sub.i, A.sub.j) represents the
attribute A in view of its theoretical co-penetration between the
i.sub.th and j.sub.th level, and C.sub.A(A.sub.i, A.sub.j)
represents the attribute A in view of its actual co-penetration
between the i.sub.th and j.sub.th level. The lower the distance,
the higher the degree of interaction. The example diagonal may be
assumed to have a distance of zero (0) because corresponding
attribute levels of such pairs are the same. As described above,
levels of an attribute may refer to descriptive sub-categories
related to a higher level attribute type. For example, a laundry
product may have major attributes of segment and benefit. A segment
attribute may have levels of liquid and powder, while a benefit
attribute may have levels of baby (e.g., to represent baby-safe or
gentle washing), base, cold (e.g., to represent cleaning
capabilities at relatively colder water temperatures), color (e.g.,
to represent color safe washing protection), delicate (e.g., to
represent fabric safe washing capabilities) and odor (e.g., to
represent an ability to remove strong odors). Attribute levels of
an attribute are mutually exclusive and do not intersect in view of
a particular product. For example, a liquid laundry detergent
cannot also be a powder detergent at the same time and, as such,
will not include intersecting data points on one or more
theoretical and/or empirical matrices.
[0026] In the illustrated example of FIG. 4, a distance matrix 400
is shown to reflect relative attribute level differences between
the theoretical co-penetration values of FIG. 2 and the actual
co-penetration values of FIG. 3. The example distance matrix 400 of
FIG. 4 includes similarly numbered elements as is shown in FIGS. 2
and 3, but with a four-hundred series prefix. In the event a first
attribute level pair exhibits a relatively lower distance value
than a second attribute level pair, then the first attribute level
pair is deemed to be more substitutable. In other words, if a new
product to be introduced into the market is more likely to
cannibalize from other market products having those attribute
levels that are closer to those of the new product. In the
illustrated example of FIG. 4, the relative distance between
attribute levels of "color" and "delic" (e.g., delicates) is 0.664
and the relative distance between attribute levels "color" and
"hypo" is 1.750, which indicates that products having the attribute
type "color" could be more easily substituted by "delic" than by
"hypo."
[0027] Generally speaking, because different products in a market
include any number of different attributes and/or attribute levels,
calculating and/or otherwise determining a relative distance
between theoretical Dirichlet co-penetration values and empirical
co-penetration values illustrates a degree of substitutability
based on the distances between attribute levels within a set of
attributes. In the event similar attributes and/or attribute levels
are found in a product to compete with a newly introduced product,
then sourcing is likely to occur therebetween. Further, in the
event that a market researcher can identify that a degree of
sourcing will likely occur from an already existing market product
by the same manufacturer as a candidate new product, then the
market researcher can recommend alternate markets to avoid
undesirable cannibalization effects.
[0028] While the example distance matrix 400 of FIG. 4 represents
relative distances between attribute levels of the "benefit"
attribute, one or more products of interest may have any number of
additional and/or alternate attributes. As such, example methods,
systems, apparatus and/or articles of manufacture disclosed herein
generate any number of distance matrices in view of any number of
candidate attributes of interest because a product of interest
(e.g., a candidate laundry product) in a category of interest
(e.g., laundry detergents) includes any number of combinations of
different attributes and/or attribute levels. In view of the
differing number of attributes and/or attribute levels for each
candidate product of interest, a substitutability index value is
calculated between two products of interest to leverage (a) the
distances between attribute levels of the two products within an
attribute and (b) a degree of polarization of the relevant
attributes. In other words, a substitutability index is calculated
for each existing product in the market as the weighted average, by
attribute polarization of all distances between the candidate
product and the existing product across all attributes. As
described above, a level of polarization is derived from an inverse
of the Dirichlet S parameter for an attribute. If an attribute
highly polarizes consumer behavior, then switching among different
levels of the attribute will be relatively low.
[0029] A substitutability index between two products of interest
may be calculated in a manner consistent with example Equation
3.
SI ( P i , P j ) = [ k = 1 n ( DOP Ak * D ( A ki , A kj ) ) k = 1 n
( DOP Ak ) ] - 1 . Equation 3 ##EQU00002##
[0030] In example Equation 3, SI(P.sub.i, P.sub.j) represents the
substitutability index between product i and product j, in which
product i is indicative of a new product (e.g., a line product
introduced into a new geographical market area) and product j is an
existing product. DOP.sub.Ak represents a degree of polarization
(e.g., an inverse of the Dirichlet S parameter) for attribute k,
D(A.sub.ki, A.sub.kj) represents a distance between the attribute
levels of product i and j for an attribute of interest k, and n
represents a number of attributes of interest. Example Equation 3
applies weights in view of a degree of polarization so that
distances between attribute levels in attributes that are more
polarizing than others count more to a consideration of the
substitutability of products. To determine a percent of volume
shifted from a specific product, example methods, apparatus,
systems and/or articles of manufacture disclosed herein apply one
or more substitutability indexes in a manner consistent with
example Equation 4.
Source ( P i , P j ) = SI ( P i , P j ) * SOPO j k = 1 n ( SI ( P i
, P k ) * SOPO k ) . Equation 4 ##EQU00003##
[0031] In example Equation 4, Source (P.sub.i, P.sub.j) refers to a
percent volume of product i that is sourced from product j, in
which product i refers to the newly introduced product and product
j refers to an existing product. SI(P.sub.i, P.sub.j) refers to the
substitutability index in a manner consistent with example Equation
3, SOPO.sub.k refers to a share of purchase occasions of product k,
and n refers to a number of existing products.
[0032] While an example manner of implementing the system 100 to
dynamically track consumer segments with point-of-sale data has
been illustrated in FIGS. 1-4, one or more of the elements,
processes and/or devices illustrated in FIGS. 1-4 may be combined,
divided, re-arranged, omitted, eliminated and/or implemented in any
other way. Further, the example product sourcing engine 102, the
example panelist database 104, the example PRL database 106, the
example new product attribute database 108, the example product
selector 110, the example product category comparator 112, the
example Dirichlet modeling engine 114, the example empirical
co-penetration engine 116, the example distance calculator 118, the
example substitutability engine 120, the example volume sourcing
calculator 122, the example theoretical co-penetration matrix 200,
the example empirical co-penetration matrix 300, and/or the example
distance matrix 400 of FIGS. 1-4 may be implemented by hardware,
software, firmware and/or any combination of hardware, software
and/or firmware. Thus, for example, any of the example product
sourcing engine 102, the example panelist database 104, the example
PRL database 106, the example new product attribute database 108,
the example product selector 110, the example product category
comparator 112, the example Dirichlet modeling engine 114, the
example empirical co-penetration engine 116, the example distance
calculator 118, the example substitutability engine 120, the
example volume sourcing calculator 122, the example theoretical
co-penetration matrix 200, the example empirical co-penetration
matrix 300, and/or the example distance matrix 400 of FIGS. 1-4
could be implemented by one or more circuit(s), programmable
processor(s), application specific integrated circuit(s) (ASIC(s)),
programmable logic device(s) (PLD(s)) and/or field programmable
logic device(s) (FPLD(s)), etc. When any of the apparatus or system
claims of this patent are read to cover a purely software and/or
firmware implementation, at least one of the example product
sourcing engine 102, the example panelist database 104, the example
PRL database 106, the example new product attribute database 108,
the example product selector 110, the example product category
comparator 112, the example Dirichlet modeling engine 114, the
example empirical co-penetration engine 116, the example distance
calculator 118, the example substitutability engine 120, the
example volume sourcing calculator 122, the example theoretical
co-penetration matrix 200, the example empirical co-penetration
matrix 300, and/or the example distance matrix 400 of FIGS. 1-4 are
hereby expressly defined to include a tangible computer readable
storage medium such as a memory, DVD, CD, Blu-ray, etc. storing the
software and/or firmware. Further still, the example system 100 of
FIG. 1 may include one or more elements, processes and/or devices
in addition to, or instead of, those illustrated in FIG. 1 and/or
may include more than one of any or all of the illustrated
elements, processes and devices.
[0033] Flowcharts representative of example machine readable
instructions for implementing the system 100 of FIG. 1, the
theoretical matrix 200 of FIG. 2, the empirical matrix 300 of FIG.
3 and/or the distance matrix 400 of FIG. 4 are shown in FIG. 5. In
this example, the machine readable instructions comprise a program
for execution by a processor such as the processor 612 shown in the
example processor platform 600 discussed below in connection with
FIG. 6. The program may be embodied in software stored on a
tangible computer readable storage medium such as a CD-ROM, a
floppy disk, a hard drive, a digital versatile disk (DVD), a
Blu-ray disk, or a memory associated with the processor 612, but
the entire program and/or parts thereof could alternatively be
executed by a device other than the processor 612 and/or embodied
in firmware or dedicated hardware. Further, although the example
program is described with reference to the flowcharts illustrated
in FIG. 5, many other methods of implementing the example system
100 to forecast new product launch sourcing may alternatively be
used. For example, the order of execution of the blocks may be
changed, and/or some of the blocks described may be changed,
eliminated, or combined.
[0034] As mentioned above, the example processes of FIG. 5 may be
implemented using coded instructions (e.g., computer readable
instructions) stored on a tangible computer readable storage medium
such as a hard disk drive, a flash memory, a read-only memory
(ROM), a compact disk (CD), a digital versatile disk (DVD), a
cache, a random-access memory (RAM) and/or any other storage media
in which information is stored for any duration (e.g., for extended
time periods, permanently, brief instances, for temporarily
buffering, and/or for caching of the information). As used herein,
the term tangible computer readable storage medium is expressly
defined to include any type of computer readable storage device
and/or storage disc and to exclude propagating signals.
Additionally or alternatively, the example processes of FIG. 5 may
be implemented using coded instructions (e.g., computer readable
instructions) stored on a non-transitory computer readable storage
medium such as a hard disk drive, a flash memory, a read-only
memory, a compact disk, a digital versatile disk, a cache, a
random-access memory and/or any other storage device and/or storage
disc in which information is stored for any duration (e.g., for
extended time periods, permanently, brief instances, for
temporarily buffering, and/or for caching of the information). As
used herein, the term non-transitory computer readable medium is
expressly defined to include any type of computer readable storage
device and/or storage disc and to exclude propagating signals. As
used herein, when the phrase "at least" is used as the transition
term in a preamble of a claim, it is open-ended in the same manner
as the term "comprising" is open ended. Thus, a claim using "at
least" as the transition term in its preamble may include elements
in addition to those expressly recited in the claim.
[0035] The program 500 of FIG. 5 begins at block 502 where the
example product selector 110 identifies a product of interest that
is to be considered for entry into a target market in which it has
not previously participated. To identify one or more existing
products in the target market of interest that are in a similar or
same category as the new product, the example product category
comparator 112 identifies such similar products from the example
PRL database 106 based on new product attribute levels stored in
the example new product attribute database 108 (block 504). The
example product category comparator 112 also identifies one or more
subgroups of products associated with an attribute level of an
attribute of interest within the category that may be a closer
match when considering target market performance (block 506). As
discussed above, a relatively high-level category related to a new
baby diaper product may be "diapers." However, the "diapers"
category may contain both "baby diapers" and "adult diapers," only
one of which (the former) is a suitable category when comparing to
a new category of a baby diaper product to be introduced into the
target market. In some examples, products in the "baby diapers"
sub-category may still be too generalized in the event the new baby
diaper product of interest is associated with "newborns" versus
"toddlers." Depending on the product type and/or number of
sub-attributes within a category, the example product category
comparator 112 identifies suitable products for comparison
purposes.
[0036] The example product category comparator 112 obtains
available purchase frequency data and/or other target market
performance data associated with the identified currently available
comparative products from the example panelist database 104 (block
508). In the event additional attributes and/or sub-attributes are
associated with the category (block 510), control returns to block
506. Otherwise, the example Dirichlet modeling engine 114 estimates
Dirichlet S parameters for attribute levels of products currently
in the target market (block 512), calculates an attribute degree of
polarization (block 513), and calculates theoretical co-penetration
values between pairs of attribute levels (block 514), such as those
shown in the example theoretical co-penetration matrix 200 of FIG.
2.
[0037] The example empirical co-penetration engine 116 retrieves
product market activity data from the panelist database 104 to
calculate an actual co-penetration without Dirichlet techniques
(block 516), such as those shown in the example empirical
co-penetration matrix 300 of FIG. 3. Using the theoretical
co-penetration values (e.g., based on the Dirichlet calculations)
and the actual co-penetration values, which are not based on
Dirichlet calculations, the example distance calculator 118
calculates a distance value (block 518). As described above, the
distance values expose a degree of substitutability between
attributes and/or attribute levels that may indicate
cannibalization could occur by introduction of the new product of
interest in the target market. The example substitutability engine
120 calculates a substitutability index value for the new product
and one of the products that already participate in the target
market (block 520). The substitutability index may be calculated in
a manner consistent with example Equation 3, and the example volume
sourcing calculator 122 calculates a volume sourcing value for the
new product based on the substitutability index and a share of
purchase occasions (block 522). The volume sourcing value may be
calculated by the example sourcing calculator 122 in a manner
consistent with example Equation 4 to reveal a percent volume of
the new product that is sourced from an already existing product in
the target market.
[0038] FIG. 6 is a block diagram of an example processor platform
600 capable of executing the instructions of FIG. 5 to implement
the system 100 of FIG. 1, the theoretical co-penetration matrix 200
of FIG. 2, the actual co-penetration matrix 300 of FIG. 3 and/or
the distance matrix 400 of FIG. 4. The processor platform 600 can
be, for example, a server, a personal computer, an Internet
appliance, or any other type of computing device.
[0039] The system 600 of the instant example includes a processor
612. For example, the processor 612 can be implemented by one or
more microprocessors or controllers from any desired family or
manufacturer.
[0040] The processor 612 includes a local memory 613 (e.g., a
cache) and is in communication with a main memory including a
volatile memory 614 and a non-volatile memory 616 via a bus 618.
The volatile memory 614 may be implemented by Synchronous Dynamic
Random Access Memory (SDRAM), Dynamic Random Access Memory (DRAM),
RAMBUS Dynamic Random Access Memory (RDRAM) and/or any other type
of random access memory device. The non-volatile memory 616 may be
implemented by flash memory and/or any other desired type of memory
device. Access to the main memory 614, 616 is controlled by a
memory controller.
[0041] The processor platform 600 also includes an interface
circuit 620. The interface circuit 620 may be implemented by any
type of interface standard, such as an Ethernet interface, a
universal serial bus (USB), and/or a PCI express interface.
[0042] One or more input devices 622 are connected to the interface
circuit 620. The input device(s) 622 permit a user to enter data
and commands into the processor 612. The input device(s) can be
implemented by, for example, a keyboard, a mouse, a touchscreen, a
track-pad, a trackball, isopoint and/or a voice recognition
system.
[0043] One or more output devices 624 are also connected to the
interface circuit 620. The output devices 624 can be implemented,
for example, by display devices (e.g., a liquid crystal display, a
cathode ray tube display (CRT), a printer and/or speakers). The
interface circuit 620, thus, typically includes a graphics driver
card.
[0044] The interface circuit 620 also includes a communication
device such as a modem or network interface card to facilitate
exchange of data with external computers via a network 626 (e.g.,
an Ethernet connection, a digital subscriber line (DSL), a
telephone line, coaxial cable, a cellular telephone system,
etc.).
[0045] The processor platform 600 also includes one or more mass
storage devices 628 for storing software and data. Examples of such
mass storage devices 628 include floppy disk drives, hard drive
disks, compact disk drives and digital versatile disk (DVD)
drives.
[0046] The coded instructions 632 of FIG. 5 may be stored in the
mass storage device 628, in the volatile memory 614, in the
non-volatile memory 616, and/or on a removable storage medium such
as a CD or DVD.
[0047] Although certain example methods, apparatus and articles of
manufacture have been described herein, the scope of coverage of
this patent is not limited thereto. On the contrary, this patent
covers all methods, apparatus and articles of manufacture fairly
falling within the scope of the claims of this patent.
* * * * *