U.S. patent application number 16/579630 was filed with the patent office on 2020-01-16 for methods and systems for predicting consumer behavior from transaction card purchases.
The applicant listed for this patent is MASTERCARD INTERNATIONAL INCORPORATED. Invention is credited to Marc Del Bene, Po Hu, Anant Nambiar, Daniel G. Salazar.
Application Number | 20200019974 16/579630 |
Document ID | / |
Family ID | 42267400 |
Filed Date | 2020-01-16 |
![](/patent/app/20200019974/US20200019974A1-20200116-D00000.png)
![](/patent/app/20200019974/US20200019974A1-20200116-D00001.png)
![](/patent/app/20200019974/US20200019974A1-20200116-D00002.png)
![](/patent/app/20200019974/US20200019974A1-20200116-D00003.png)
![](/patent/app/20200019974/US20200019974A1-20200116-D00004.png)
![](/patent/app/20200019974/US20200019974A1-20200116-D00005.png)
![](/patent/app/20200019974/US20200019974A1-20200116-D00006.png)
![](/patent/app/20200019974/US20200019974A1-20200116-D00007.png)
![](/patent/app/20200019974/US20200019974A1-20200116-D00008.png)
![](/patent/app/20200019974/US20200019974A1-20200116-M00001.png)
United States Patent
Application |
20200019974 |
Kind Code |
A1 |
Bene; Marc Del ; et
al. |
January 16, 2020 |
METHODS AND SYSTEMS FOR PREDICTING CONSUMER BEHAVIOR FROM
TRANSACTION CARD PURCHASES
Abstract
A computer-based method for predicting consumer behavior is
provided. The method is performed using a computer system coupled
to a database. The method includes recording consumer data in the
database for each consumer of a global population of consumers
including historical purchases made by each consumer using a
transaction card, defining a life event by assigning spending
variables to the life event, determining a sample group of
consumers that are experiencing the life event based on the
consumer data stored within the database with respect to the
spending variables, generating a predictive model based on
historical purchases made by consumers within the sample group, and
applying the predictive model to predict each consumer within the
global population that will experience the life event. The
predictive model is applied using the computer system. A list of
consumers predicted to experience the life event within a
predetermined time period is output.
Inventors: |
Bene; Marc Del; (Darien,
CT) ; Hu; Po; (Norwalk, CT) ; Nambiar;
Anant; (Larchmont, NY) ; Salazar; Daniel G.;
(Ridgefield, CT) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
MASTERCARD INTERNATIONAL INCORPORATED |
Purchase |
NY |
US |
|
|
Family ID: |
42267400 |
Appl. No.: |
16/579630 |
Filed: |
September 23, 2019 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
12342925 |
Dec 23, 2008 |
10430803 |
|
|
16579630 |
|
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G06Q 30/02 20130101;
G06Q 30/0204 20130101; G06Q 10/06375 20130101; G06Q 30/0202
20130101 |
International
Class: |
G06Q 30/02 20060101
G06Q030/02 |
Claims
1.-22. (canceled)
23. A computer-based method for predicting consumer behavior, said
method performed using a computer device coupled to a database,
said method comprising: storing consumer data in the database for
consumers of a global population of consumers including historical
purchases made by the consumers using a transaction card; defining
a life event by assigning spending variables to the life event;
generating an anticipated spend for each spending variable assigned
to the life event by extrapolating a historical actual spend based
on the historical purchases stored in the database, wherein the
anticipated spend represents anticipated spending in each spending
variable assigned to the life event; creating a sample group of
consumers from the global population of consumers that represents
consumers that are experiencing the life event; generating a
predictive model based on the historical purchases made by the
consumers within the sample group; applying the predictive model to
predict each consumer within the global population and outside of
the sample group that will experience the life event; and
outputting a list of consumers outside of the sample group that are
predicted to experience the life event within a predetermined time
period.
24. A computer-based method in accordance with claim 23 wherein
determining a sample group of consumers that are experiencing the
life event further comprises: analyzing historical purchases for
each consumer within the global population having an account in the
database during a historical time period for each spending variable
in a consumption bundle of the life event; determining a historical
actual spending curve for each consumer within the global
population during the historical time period using the historical
purchases made in the consumption bundle; determining an
anticipated spending curve for each consumer within the global
population during a second time period after the historical time
period using the historical actual spending curve, wherein the
anticipated spending curve is determined before the second time
period occurs; collecting purchase data for each consumer within
the global population over the second time period; and determining
an actual spending curve during the second time period for each
consumer within the global population using the collected purchase
data, wherein the actual spending curve is determined after the
second time period occurs.
25. A computer-based method in accordance with claim 24 wherein
creating the sample group of consumers that are experiencing the
life further comprises: comparing the actual spending curve and the
anticipated spending curve for each consumer within the global
population during the second time period; determining a variance
between the actual spending curve and the anticipated spending
curve for each consumer within the global population; and assigning
each consumer within the global population to the sample group
using the determined variance.
26. A computer-based method in accordance with claim 23 further
comprising: causing a customized offer to be sent to at least one
consumer on the output list for a product related to the life event
being experienced by the at least one consumer.
27. A computer-based method in accordance with claim 23 further
comprising refining the predictive model using actual spending by
at least one consumer on the output list of consumers based on
purchases made on a transaction card.
28. A computer-based method in accordance with claim 23 wherein
storing consumer data in the database for each consumer of a global
population further comprises: storing transaction card purchase
data for each consumer within the global population to the
database; and storing third party data about each consumer within
the global population to the database.
29. A computer-based method in accordance with claim 23 wherein
defining a life event by assigning spending variables to the life
event further comprises: selecting a life event a consumer may
experience; and assigning at least one spending variable to the
selected life event to form a consumption bundle for the selected
life event.
30. A computer-based method in accordance with claim 29 wherein
defining a life event by assigning spending variables to the life
event further comprises assigning at least one demographic variable
to the life event, the consumption bundle comprising the at least
one assigned demographic variable and the at least one assigned
spending variable.
31. A computer-based method in accordance with claim 23 wherein
generating a predictive model based on historical purchases made by
consumers within the sample group further comprises: analyzing each
consumer in the sample group during a historical time period to
determine spending trends that are common to consumers within the
sample group based on the historical purchases, wherein a spending
trend represents at least one of an increase, a decrease or no
change in an amount of spending in a spending variable; and
generating a predictive model that includes the determined spending
trends.
32. A computer-based method in accordance with claim 23 wherein
applying the predictive model to predict each consumer within the
global population that will experience the life event further
comprises: applying the predictive model to all consumers within
the global population having an account in the database; modeling
actual spending trends of each consumer within the global
population using the predictive model, the predictive model
including a set of spending trends; and when the actual spending
trends of a consumer within the global population are similar to
the set of spending trends included within the predictive model,
predicting that the consumer will experience the life event.
33. A computer for predicting behavior of a consumer, said computer
comprising a processor, computer-readable instructions executable
by the processor, and a database, said computer configured to:
store consumer data in the database for consumers of a global
population of consumers including historical purchases made by the
consumers using a transaction card; define a life event by
assigning spending variables to the life event; generate an
anticipated spend for each spending variable assigned to the life
event by extrapolating a historical actual spend based on the
historical purchases stored in the database, wherein the
anticipated spend represents anticipated spending in each spending
variable assigned to the life event; create a sample group of
consumers from the global population of consumers that represents
consumers that are experiencing the life event; generate a
predictive model based on the historical purchases made by the
consumers within the sample group; apply the predictive model to
predict each consumer within the global population and outside of
the sample group that will experience the life event; and output a
list of consumers outside of the sample group that are predicted to
experience the life event within a predetermined time period.
34. A computer in accordance with claim 33 further configured to:
analyze historical purchases for a consumer within the global
population having an account in said database during a historical
time period for each spending variable in a consumption bundle of
the life event; determine a historical actual spending curve during
the historical time period for the consumer using the historical
purchases made in the consumption bundle; determine an anticipated
spending curve during a second time period after the historical
time period for the consumer using the historical actual spending
curve, wherein the anticipated spending curve is determined before
the second time period occurs; collect purchase data for the
consumer over the second time period; determine an actual spending
curve during the second time period for the consumer using the
collected purchase data, wherein the actual spending curve is
determined after the second time period occurs; compare the actual
spending curve and the anticipated spending curve for the consumer
during the second time period; determine a variance between the
actual spending curve and the anticipated spending curve for the
consumer; and assign the consumer to the sample group using the
determined variance.
35. A computer in accordance with claim 33 further configured to:
map anticipated consumer needs to at least one offer stored within
the database based on the defined life event; and provide the least
one offer to the consumer.
36. A computer in accordance with claim 33 further configured to
refine the predictive model using actual spending of the consumers
within the sample group based on purchases made on transaction
cards by at least one of adding a spending variable to a definition
of the life event and removing a spending variable from the
definition of the life event.
37. A computer in accordance with claim 33 further configured to:
select a life event a consumer may experience; and assign at least
one spending variable to the selected life event to form a
consumption bundle for the selected life event.
38. A computer in accordance with claim 33 further configured to:
analyze each consumer in the sample group during a historical time
period to determine spending trends that are common to consumers
within the sample group based on the stored consumer data, wherein
a spending trend represents at least one of an increase, a decrease
or no change in an amount of spending in a spending variable; and
generate a predictive model that includes the determined spending
trends.
39. A computer in accordance with claim 38 further configured to:
apply the predictive model to each consumer within the global
population; model actual spending trends of each consumer within
the global population with the predictive model; and when actual
spending trends of a consumer within the global population are
similar to spending trends within the predictive model, predict
that the consumer will experience the life event.
40. A network based system for predicting behavior of a consumer,
said system comprising: a client computing device; a database for
storing information; and a server computing device comprising a
processor and computer-readable instructions executable by said
processor, said server computing device configured to be coupled to
said client computing device and said database, said server
computing system further configured to: store consumer data in the
database for consumers of a global population of consumers
including historical purchases made by the consumers using a
transaction card; define a life event by assigning spending
variables to the life event; generate an anticipated spend for each
spending variable assigned to the life event by extrapolating a
historical actual spend based on the historical purchases stored in
the database, wherein the anticipated spend represents anticipated
spending in each spending variable assigned to the life event;
create a sample group of consumers from the global population of
consumers that represents consumers that are experiencing the life
event; generate a predictive model based on the historical
purchases made by the consumers within the sample group; apply the
predictive model to predict each consumer within the global
population and outside of the sample group that will experience the
life event; and output a list of consumers outside of the sample
group that are predicted to experience the life event within a
predetermined time period.
41. A network based system in accordance with claim 40, wherein
said server computing device is further configured to: analyze
historical purchases of a consumer within the global population
having an account in said database during a historical time period
for each spending variable in a consumption bundle of the life
event; determine a historical actual spending curve during the
historical time period for the consumer using the historical
purchases made in the consumption bundle; determine an anticipated
spending curve during a second time period after the historical
time period for the consumer using the historical actual spending
curve, wherein the anticipated spending curve is determined before
the second time period occurs; collect purchase data for the
consumer over the second time period; determine an actual spending
curve during the second time period for the consumer using the
collected purchase data, wherein the actual spending curve is
determined after the second time period occurs; compare the actual
spending curve and the anticipated spending curve for the consumer
during the second time period; determine a variance between the
actual spending curve and the anticipated spending curve; and
assign the consumer to the sample group using the determined
variance.
42. A network based system in accordance with claim 40, wherein
said server computing device is further configured to: analyze each
consumer in the sample group during a historical time period to
determine spending trends that are common to consumers within the
sample group, wherein a spending trend represents at least one of
an increase, a decrease or no change in an amount of spending in a
spending variable; generate a predictive model that includes the
determined spending trends; apply the predictive model to each
consumer within the global population; model actual spending trends
of each consumer within the global population with the spending
trends included within the predictive model; and when actual
spending trends of a consumer within the global population are
similar to spending trends within the predictive model, predict
that the consumer will experience the life event.
Description
BACKGROUND OF THE INVENTION
[0001] The field of the invention relates generally to predicting
consumer behavior from transaction card purchases and, more
particularly, to network-based methods and systems for predicting
whether a consumer will experience a life event, and predicting
future purchases of the consumer based on the predicted life
event.
[0002] Historically, the use of "charge" or transaction cards or
payment cards for consumer transaction payments was at most
regional and based on relationships between local credit or debit
card issuing banks and various local merchants. The transaction
card industry has since evolved with the issuing banks forming
associations or networks (e.g., MasterCard.RTM.) and involving
third party transaction processing companies (e.g., "Merchant
Acquirers") to enable cardholders to widely use transaction cards
at any merchant's establishment, regardless of the merchant's
banking relationship with the card issuer. (MasterCard is a
registered trademark of MasterCard International Incorporated
located in Purchase, N.Y.).
[0003] For example, FIG. 1 shows an exemplary multi-party payment
card industry system for enabling payment-by-card transactions in
which the merchants and issuer do not need to have a one-to-one
special relationship. Yet, various scenarios exist in the
payment-by-card industry today, where the card issuer has a special
or customized relationship with a specific merchant, or group of
merchants. These special or customized relationships may, for
example, include private label programs, co-brand programs,
proprietary card brands, rewards programs, and others.
[0004] Further, many merchants spend large amounts of money on
marketing. Because marketing to a large general audience may be
expensive, it may be advantageous to determine consumer interest in
advance and attempt to target marketing toward consumers who are
more likely to be interested in the product or products that a
particular merchant sells. In this way merchants may attempt to
better utilize their marketing budget to improve sales. In
addition, consumers will be less likely to receive irrelevant or
uninteresting offers.
[0005] At least some known systems and methods for determining
consumer interest have relied on demographic information, such as
age, income, and/or occupation. However, a consumer's needs may
change quickly based on the consumer's current circumstances and/or
the consumer's knowledge of future circumstances. For example, a
consumer may decide to purchase a house in the near future. As
such, the circumstances driving the consumer's decision to purchase
the house may change faster than the ability of any of the known
systems to determine such a change in demographic information. In
other words, the demographic data of the consumer may change
resulting in many purchases by the consumer, but by the time the
change is detected by the known systems, the consumer has already
made many of the purchases. In these cases, many of these purchases
are then missed by the marketer.
[0006] In some other cases, there may be little or no change in
demographic data although circumstances surrounding the consumer
have changed and, accordingly, the consumer's needs have changed.
For example, it may be more likely that a consumer who is planning
to buy a house will need a mortgage and/or new furniture. As such,
even though the consumer's demographic data, such as age, income,
and/or occupation, may have remained constant, it is probable that
the consumer will take out a mortgage and/or buy new furniture in
the near future because of an impending home purchase.
[0007] However, in some cases, demographic data does help to
predict some purchases. For example, if a person's job includes
frequent relocation, that person may be a frequent home purchaser.
Based on the demographic data alone, it may be possible to predict
that the person is likely to buy a home.
[0008] At least some known targeting models have been known to
achieve lifts of 1.3 times to 1.6 times on large populations. As
used herein, the term "lift" refers to a ratio of positive
responses to an offer by a consumer included within a target
subgroup as compared to positive responses to the same offer made
to the population as a whole. The target subgroup is usually
selected to include those members of the whole population that are
more likely to respond. It may, however, be difficult to determine
the demographic information. This may be especially likely when
changes have recently occurred in the person's demographic
data.
[0009] Additionally, even when correct demographic information is
known, it may be difficult to determine when a purchase, such as a
home purchase, will actually occur. Timing is important in
advertising and/or marketing because an unneeded advertisement
and/or coupon might be thrown away, thus, wasting marketing money
spent by the merchant. In one example, a discounted mortgage offer
from bank X might be discarded by a consumer this month, while a
discounted mortgage offer from bank Y might be used by the consumer
next month because it is received near the time of a home purchase.
As such, if bank Y is better at predicting when a mortgage is
needed by a consumer, bank Y may be able to get more business than
bank X. Further, the resources of bank X may be wasted by sending
mortgage offers to consumers not planning on buying a new home, and
the consumers may ignore possibly relevant offers after receiving
many irrelevant offers. As such, matching offers to a consumer or a
specific group of consumers in a timely fashion may be beneficial
for the parties involved.
[0010] Accordingly, it is desirable to have the ability to identify
indicators and/or signals that suggest a change in a consumer's
needs or behaviors. By determining a consumer's changing needs and
behaviors more accurately a merchant may, for example, be able to
better predict what promotion, offers, and/or coupons to send to a
consumer, and when these promotions, offers, and/or coupons should
be sent to the consumer.
BRIEF DESCRIPTION OF THE INVENTION
[0011] In one aspect, a computer-based method for predicting
consumer behavior is provided. The method is performed using a
computer system coupled to a database. The method includes
recording consumer data in the database for each consumer of a
global population of consumers including historical purchases made
by each consumer using a transaction card, defining a life event by
assigning spending variables to the life event, determining a
sample group of consumers that are experiencing the life event
based on the consumer data stored within the database with respect
to the spending variables, generating a predictive model based on
historical purchases made by consumers within the sample group, and
applying the predictive model to predict each consumer within the
global population that will experience the life event. The
predictive model is applied using the computer system. A list of
consumers predicted to experience the life event within a
predetermined time period is output.
[0012] In another aspect, a computer for predicting behavior of a
consumer based on the consumer's purchases made using a transaction
card is provided. The computer is coupled to a database. The
computer is configured to record consumer data in the database for
each consumer of a global population of consumers including
historical purchases made by each consumer using a transaction
card, define a life event by assigning spending variables to the
life event, determine a sample group of consumers that are
experiencing the life event based on the consumer data stored
within the database with respect to the spending variables,
generate a predictive model based on historical purchases made by
consumers within the sample group, apply the predictive model to
predict each consumer within the global population that will
experience the life event, and output a list of consumers predicted
to experience the life event within a predetermined time
period.
[0013] In still another aspect, a network based system for
predicting behavior of a consumer based on the consumer's purchases
made using a transaction card is provided. The system includes a
client system, a centralized database for storing information, and
a server system configured to be coupled to the client system and
the database. The server system is configured to record consumer
data in the centralized database for each consumer of a global
population of consumers including historical purchases made by each
consumer using a transaction card, define a life event by assigning
spending variables to the life event, determine a sample group of
consumers that are experiencing the life event based on the
consumer data stored within said database with respect to the
spending variables, generate a predictive model based on historical
purchases made by consumers within the sample group, apply the
predictive model to predict each consumer within the global
population that will experience the life event, and output a list
of consumers predicted to experience the life event within a
predetermined time period.
[0014] In still another aspect, a computer program embodied on a
computer readable medium for predicting consumer behavior is
provided. The program includes at least one code segment that
records consumer data in the database for each consumer of a global
population of consumers including historical purchases made by each
consumer using a transaction card, defines a life event by
assigning spending variables to the life event, determines a sample
group of consumers that are experiencing the life event based on
the consumer data stored within the database with respect to the
spending variables, generates a predictive model based on
historical purchases made by consumers within the sample group,
applies the predictive model to predict each consumer within the
global population that will experience the life event, and outputs
a list of consumers predicted to experience the life event within a
predetermined time period.
[0015] The embodiments described herein facilitate achieving higher
lifts as compared to other known targeting methods and system. For
example, the embodiments described herein achieve lifts of
approximately 2.4 times to approximately 4.3 times, as compared to
the lifts of 1.3 times to 1.6 times for known targeting methods for
a large population. In other words, the methods and systems
described herein provide better accuracy in predicting future
purchases of a consumer.
BRIEF DESCRIPTION OF THE DRAWINGS
[0016] FIG. 1 is a schematic diagram illustrating an exemplary
multi-party payment card industry system for enabling ordinary
payment-by-card transactions in which the merchants and issuer do
not need to have a one-to-one special relationship.
[0017] FIG. 2 is a simplified block diagram of an exemplary
embodiment of a server architecture of a system in accordance with
one embodiment of the present invention.
[0018] FIG. 3 is an expanded block diagram of an exemplary
embodiment of a server architecture of a system in accordance with
one embodiment of the present invention.
[0019] FIG. 4 is a flowchart illustrating an exemplary method
utilized by the system shown in FIG. 2 for predicting a consumer's
behavior.
[0020] FIG. 5 is a flowchart illustrating an exemplary method for
recording consumer data in a database that may be used with the
method shown in FIG. 4.
[0021] FIG. 6 is a flowchart illustrating an exemplary method for
defining a life event that may be used with the method shown in
FIG. 4.
[0022] FIG. 7 is a flowchart illustrating an exemplary method for
determining a sample group of consumers that may be used with the
method shown in FIG. 4.
[0023] FIG. 8 is a graph of historical actual spending, anticipated
actual spending, and actual spending for a consumer in a
consumption bundle that may be used with the method shown in FIG.
7.
[0024] FIG. 9 is a flowchart illustrating an exemplary method for
generating a predictive model that may be used with the method
shown in FIG. 4.
[0025] FIG. 10 is a flowchart illustrating an exemplary method for
predicting whether a consumer will experience a life event that may
be used with the method shown in FIG. 4.
DETAILED DESCRIPTION OF THE INVENTION
[0026] The embodiments described herein are directed to systems and
methods for predicting consumer behavior based on the consumer's
purchases using transaction cards, such as a credit card, debit
card, membership cards, promotional cards, frequent flyer cards,
identification cards, prepaid cards, gift cards, and/or any other
devices that may hold payment account information, such as mobile
phones, personal digital assistants (PDAs), and key fobs. Such
cards and/or devices are referred to herein as "a transaction card"
or "transaction cards." These cards can all be used as a method of
payment for performing a transaction. For example, a transaction
card franchiser, transaction card provider, bank, and/or credit
union may capture and store purchasing data for account holders.
The purchasing data for each transaction may include an account
number, a merchant identification, a transaction amount, a
transaction date, and/or any other suitable information related to
the transaction.
[0027] In the exemplary embodiment, consumer behavior is predicted
using, for example, a transaction date, a transaction amount, and a
spending variable designation for purchases made using a
transaction card. The transaction date, as referred to herein, may
be the purchase date, the posting date, and/or any date associated
with the transaction. The same type of date, such as the purchase
date or the posting date, can be for all transactions.
Alternatively, different types of dates can be used for different
transactions. As such, in the alternative embodiment, any date
available and associated with a transaction can be used, even if
purchase dates and posting dates are mixed together in the same
data set. By using transaction data, such as the date and the
amount, the transaction card provider can identify consumer
spending behaviors and enable merchants to design offers and
targeted marketing campaigns.
[0028] The consumer behavior modeling systems and methods described
herein are based on an historical spending behavior of each
consumer to predict whether a consumer will experience a predefined
life event (e.g., purchasing a new home, having a new baby, and
sending a child to college), and thus, predict the future spending
habits of each consumer that is predicted to experience such a life
event. The example embodiment includes defining a plurality of life
events that a consumer may experience. Each life event includes a
consumption bundle or spending variables. Each consumption bundle
includes different services or products that a consumer purchases
using a transaction card when that consumer is experiencing a
particular life event. The example embodiment, therefore,
determines whether the spending habits of a consumer changes or
varies by a predetermined amount within a consumption bundle. If,
for example, a consumer purchases more of a certain product
included within a consumption bundle (or spending variables) as
compared to that consumer's past purchasing data, then that
consumer may be experiencing the life event assigned to that
consumption bundle. Such information can then be used to identify
other consumers that will also experience the same life event, and
predict those consumers that will increase spending within the
consumption bundle.
[0029] More specifically, the systems and processes described
herein enable a user to predict whether a consumer will experience
a life event, and then predict future purchases of the consumer
based on the predicted life event. A technical effect of the
systems and processes described herein include at least one of (a)
recording consumer data in a database for each consumer having made
a purchase using a transaction card including a transaction date
and a transaction amount, wherein the consumer is included within a
global population of consumers; (b) defining a life event by
assigning spending variables to the life event, wherein the life
event may include such happenings as purchasing a new home, having
a new baby, and sending a child to college; (c) for each consumer
stored within the database, predicting an amount the consumer will
spend on products or services included within the spending
variables assigned to the life event for a predetermined period of
time in the future, wherein the spend prediction is based on
consumer data stored within the database; (d) calculating a
variance for each consumer stored within the database, the variance
comprising a difference between an actual amount spent by the
consumer on products or services included within the spending
variables assigned to the life event for the predetermined period
of time in the future and the predicted amount; (e) identifying the
consumers having a positive variance as consumers experiencing the
life event and including the identified consumers in a sample
consumer group; (f) analyzing consumer data for each consumer
included within the sample consumer group; (g) generating a
predictive model based on the consumer data analysis, wherein the
predictive model uses consumer data to predict whether a
corresponding consumer will experience the life event; (h) applying
the predictive model to all consumers stored within the database;
and (i) outputting a list of consumers satisfying the predictive
model, wherein the outputted list of consumers includes each
consumer that will experience the life event within a predetermined
period of time. By determining the consumers that will experience
the life event and by assigning the spending variables to the life
event, a user can then predict an increase in spending by these
consumers in the spending variable assigned to the life event. This
system and method can be used for a plurality of life events,
wherein each life event has certain spending variable assigned
thereto.
[0030] In one embodiment, a computer program is provided, and the
program is embodied on a computer readable medium and utilizes a
Structured Query Language (SQL) with a client user interface
front-end for administration and a web interface for standard user
input and reports. In an exemplary embodiment, the system is web
enabled and is run on a business-entity intranet. In yet another
embodiment, the system is fully accessed by individuals having an
authorized access outside the firewall of the business-entity
through the Internet. In a further exemplary embodiment, the system
is being run in a Windows.RTM. environment (Windows is a registered
trademark of Microsoft Corporation, Redmond, Wash.). In yet another
embodiment, the system is run on a mainframe environment and a
UNIX.RTM. server environment (UNIX is a registered trademark of
AT&T, New York, N.Y.). The application is flexible and designed
to run in various different environments without compromising any
major functionality.
[0031] The systems and processes are not limited to the specific
embodiments described herein. In addition, components of each
system and each process can be practiced independent and separate
from other components and processes described herein. Each
component and process also can be used in combination with other
assembly packages and processes.
[0032] FIG. 1 is a schematic diagram 20 illustrating an exemplary
multi-party payment card industry system for enabling ordinary
payment-by-card transactions in which the merchants 24 and issuer
30 do not need to have a one-to-one special relationship. The
present invention relates to a payment card system, such as a
credit card payment system using the MasterCard.RTM. interchange
network. The MasterCard.RTM. interchange network is a set of
proprietary communications standards promulgated by MasterCard
International Incorporated.RTM. for the exchange of financial
transaction data and settlement funds between financial
institutions that are members of MasterCard International
Incorporated.RTM.. (MasterCard International Incorporated is a
registered trademark of MasterCard International Incorporated
located in Purchase, N.Y.).
[0033] In a typical payment card system, a financial institution
called the "issuer" issues a payment card, such as a credit card,
to a consumer 22, who uses the card to tender payment for a
purchase from a merchant 24. To accept payment with the card, the
merchant 24 must normally establish an account with a financial
institution that is part of the financial payment system. This
financial institution is usually called the "merchant bank" or the
"acquiring bank" or "acquirer bank." When a consumer 22 tenders
payment for a purchase with a card, the merchant 24 requests
authorization from the merchant bank 26 for the amount of the
purchase. The request may be performed over the telephone, but is
usually performed through the use of a point-of-sale terminal,
which reads the consumer's account information from the magnetic
stripe or chip on the card and communicates electronically with the
transaction processing computers of the merchant bank 26.
Alternatively, a merchant bank 26 may authorize a third party to
perform transaction processing on its behalf. In this case, the
point-of-sale terminal will be configured to communicate with the
third party. Such a third party is usually called a "merchant
processor" or an "acquiring processor" or a "third party
processor."
[0034] Using the interchange network 28, the computers of the
merchant bank 26 or the merchant processor will communicate with
the computers of the issuer bank 30 to determine whether the
consumer's account 32 is in good standing and whether the purchase
is covered by the consumer's available credit line. Based on these
determinations, the request for authorization will be declined or
accepted. If the request is accepted, an authorization code is
issued to the merchant 24.
[0035] When a request for authorization is accepted, the available
credit line of consumer's account 32 is decreased. Normally, a
charge for a credit transaction is not posted immediately to a
consumer's account 32 because bankcard associations, such as
MasterCard International Incorporated.RTM., have promulgated rules
that do not allow a merchant 24 to charge, or "capture," a
transaction until goods are shipped or services are delivered.
However, with respect to at least some debit card transactions, a
charge may be posted at the time of the transaction. When a
merchant 24 ships or delivers the goods or services, the merchant
24 captures the transaction by, for example, appropriate data entry
procedures on the point-of-sale terminal. This may include bundling
of approved transactions daily for standard retail purchases. If a
consumer 22 cancels a transaction before it is captured, a "void"
is generated. If a consumer 22 returns goods after the transaction
has been captured, a "credit" is generated.
[0036] After a transaction is captured, the transaction is settled
between the merchant 24, the merchant bank 26, and the issuer 30.
Settlement refers to the transfer of financial data or funds
between the merchant's account, the merchant bank 26, and the
issuer 30 related to the transaction. Usually, transactions are
captured and accumulated into a "batch," which are settled as a
group. More specifically, a transaction is typically settled
between the issuer 30 and the interchange network 28, and then
between the interchange network 28 and the merchant bank 26 (also
known as the acquirer bank), and then between the merchant bank 26
and the merchant 24.
[0037] FIG. 2 is a simplified block diagram of an exemplary system
100 in accordance with one embodiment of the present invention. In
one embodiment, system 100 is a payment card system used for
predicting consumer behavior, and is operable to implement the
modeling techniques and transaction database described herein. In
addition, system 100 is operable as a payment card system, which
can be utilized by users for management of accounts and payment
transactions.
[0038] More specifically, in the example embodiment, system 100
includes a server system 112, and a plurality of client
sub-systems, also referred to as client systems 114, connected to
server system 112. In one embodiment, client systems 114 are
computers including a web browser, such that server system 112 is
accessible to client systems 114 using the Internet. Client systems
114 are interconnected to the Internet through many interfaces
including a network, such as a local area network (LAN) or a wide
area network (WAN), dial-in-connections, cable modems and special
high-speed ISDN lines. Client systems 114 could be any device
capable of interconnecting to the Internet including a web-based
phone, personal digital assistant (PDA), or other web-based
connectable equipment. A database server 116 is connected to a
database 120 containing information on a variety of matters, as
described below in greater detail.
[0039] In one embodiment, centralized database 120 is stored on
server system 112 and can be accessed by potential users at one of
client systems 114 by logging onto server system 112 through one of
client systems 114. In an alternative embodiment, database 120 is
stored remotely from server system 112 and may be non-centralized.
Database 120 stores transaction data generated as part of sales
activities conducted over the bankcard network including data
relating to merchants, account holders or customers, and purchases.
Database 120 may also be utilized to store survey results and
results of the modeling processes described herein.
[0040] FIG. 3 is an expanded block diagram of an exemplary
embodiment of a server architecture of a system 122 in accordance
with one embodiment of the present invention. Components in system
122, identical to components of system 100 (shown in FIG. 2), are
identified in FIG. 3 using the same reference numerals as used in
FIG. 2. System 122 includes server system 112 and client systems
114. Server system 112 further includes database server 116, an
application server 124, a web server 126, a fax server 128, a
directory server 130, and a mail server 132. A disk storage unit
134 is coupled to database server 116 and directory server 130.
Servers 116, 124, 126, 128, 130, and 132 are coupled in a local
area network (LAN) 136. In addition, a system administrator's
workstation 138, a user workstation 140, and a supervisor's
workstation 142 are coupled to LAN 136. Alternatively, workstations
138, 140, and 142 are coupled to LAN 136 using an Internet link or
are connected through an Intranet.
[0041] Each workstation, 138, 140, and 142 is a personal computer
having a web browser. Although the functions performed at the
workstations typically are illustrated as being performed at
respective workstations 138, 140, and 142, such functions can be
performed at one of many personal computers coupled to LAN 136.
Workstations 138, 140, and 142 are illustrated as being associated
with separate functions only to facilitate an understanding of the
different types of functions that can be performed by individuals
having access to LAN 136.
[0042] Server system 112 is configured to be communicatively
coupled to various individuals, including employees 144 and to
third parties, e.g., account holders, customers, auditors, etc.,
146 using an ISP Internet connection 148. The communication in the
exemplary embodiment is illustrated as being performed using the
Internet, however, any other wide area network (WAN) type
communication can be utilized in other embodiments, i.e., the
systems and processes are not limited to being practiced using the
Internet. In addition, and rather than WAN 150, local area network
136 could be used in place of WAN 150.
[0043] In the exemplary embodiment, any authorized individual
having a workstation 154 can access system 122. At least one of the
client systems includes a manager workstation 156 located at a
remote location. Workstations 154 and 156 are personal computers
having a web browser. Also, workstations 154 and 156 are configured
to communicate with server system 112. Furthermore, fax server 128
communicates with remotely located client systems, including a
client system 156 using a telephone link. Fax server 128 is
configured to communicate with other client systems 138, 140, and
142 as well.
[0044] FIG. 4 shows a flowchart illustrating an exemplary method
200 for predicting a consumer's behavior. Method 200 is performed
using system 100 (shown in FIG. 2) and/or system 122 (shown in FIG.
3). Method 200 includes recording 202 consumer data in a database,
such as database 120, and defining 204 a life event using at least
one spending variable. As used herein, the term "life event" refers
to an event experienced by a consumer during his/her life time.
Examples of a life event include, but are not limited to including,
purchasing a new home, having a new baby, sending a child to
college, starting a business, marriage, a promotion, retirement,
getting a new job, graduating, receiving an inheritance, starting
post-grad education, having new grandchildren, getting divorce,
purchasing a car, and/or experiencing a medical event.
[0045] Further, as used herein, the term "spending variable" refers
to spending at a type of merchant and/or on a type of good or
service. Examples of spending variables include, but are not
limited to including, spending at the following types of merchants
and/or on the following types of goods and/or services: maternity,
child/baby goods/services, toys, home goods/services, education,
travel, financial instruments, office supplies, electronics,
vehicles, books/media, home repair/remodeling, clothing,
appliances, grocery, restaurants, phone/utility bills, and/or
sub-categories of the above-listed types. The spending variable for
kid/baby goods/services includes, for example, spending at
child/baby stores, spending on a babysitter/nanny, purchases of
child/baby clothing, and/or purchases of baby food, diapers,
bottles, etc. A plurality of spending variables are defined within
system 100 and/or 122, however, in the exemplary embodiment, a
sub-set of all spending variables are used to define 204 a life
event.
[0046] After data is recorded 202 and at least one life event is
defined 204, method 200 includes determining 206 a sample group of
consumers that are experiencing the defined life event, as
described in more detail below. The sample group is used to
generate 208 a predictive model using spending trends that are
common to consumers within the sample group. The predictive model
is used to predict 210 which consumers in a global population of
consumers will experience the life event. Such consumers are
referred to herein as "predicted consumers." Further, as used
herein, the term "global population" refers to all consumers having
an account on the interchange network 28. In an alternative
embodiment, the term "global population" refers to a selected set
of all the consumer accounts on the interchange network 28. For
example, the selected set of accounts includes accounts having at
least a predetermined number of purchases per month and that have
been open for at least a predetermined time period. Moreover, the
term "spending trends" as used herein refers to an increase, a
decrease, or no change in the amount of spending in a spending
variable, whether or not used to define 204 the life event, over a
predetermined time period. Examples of spending trends include,
without limitation, decreased spending at restaurants for the past
six months, a purchase of a new home, and/or trading in an older
car on a newer car. Steps 202, 204, 206, and 208 are described in
more detail below.
[0047] In the exemplary embodiment, by predicting 210 which
consumers in the global population will experience the life event,
the spending behavior of such predicted consumers can be predicted.
For example, predicted behaviors of the predicted consumers for the
life event of having a new baby include spending more on baby
supplies and/or maternity clothing, buying a family car, starting a
college savings plan, and/or spending less on dining out and/or
travel. The step of predicting 210 consumers is described in more
detail below. A list, table, file, and/or other suitable
compilation of the predicted consumers are then output 212 to, for
example, and referring to FIGS. 2 and 3, server system 112, client
system 114, database server 116, disk storage unit 134,
workstation(s) 138, 140, 142, 154, and/or 156, a printer, a
removable storage device, and/or to any other suitable location. In
one embodiment, the output compilation of consumers can be used by
the interchange network 28, issuer bank 30, and/or a merchant 24 to
provide 214 an offer to a predicted consumer.
[0048] More specifically, in the exemplary embodiment, to provide
214 offers, anticipated consumer needs are mapped to offers based
on the predicted change in needs and/or behaviors. Such mapping is
used to recommend products to consumers predicted to experience the
life event. For example, the list of predicted consumers can be
offered value propositions from the interchange network 28.
Further, the list of predicted consumers may be used for
cross-selling products and/or services, such as, without
limitation, banking (checking, savings, money market/CDs), lines
and loans (mortgages/home equity, personal, student, small
business), investing, planning (retirement, tax, education), and/or
insurance. Additionally, a communication channel may be used to
communicate a recommended promotion to the participating banks
and/or directly to a predicted consumer. The communication channel
can be any suitable communication channel, such as, without
limitation, e-mail, mail, the internet, and/or in-person. In some
embodiments, when timing may be important, offers, coupons, and/or
promotions are provided with the consumer's transaction card
monthly statement. In other embodiments, when timing is important,
offers, coupons and/or promotions are sent to the consumer using an
e-mail address associated with the card.
[0049] In one embodiment, when the predictive model determines that
a consumer is or will experience the life event, the consumer be
provided 214 with offers and/or promotions related to the life
event. More specifically, the system 100 and/or 122 transmits
information related to the predicted consumers and the life event
to a bank and/or merchant, such that bank and/or merchant can
provide 214 an offer, a suggestion, and/or promotion to a predicted
consumer. Further, the bank and/or merchant can transmit available
offers and/or promotions to the database 120 such that system 100
and/or 122 can match an offer and/or promotion to a predicted
consumer based on the life event. In an alternative embodiment, the
predicted model suggests products and/or promotions that are likely
to appeal to a predicted customer based on the predicted change in
needs or behavior associated with the life event. For example, if
the predictive model predicts, based on a customer's historical
spending, that a customer may be pregnant or have a pregnant
spouse, it may be likely that promotions for baby related products
may appeal to the customer.
[0050] Further, future actual spending of the predicted consumers
can optionally be monitor and/or analyzed to refine 216 the
predictive model by adding, verifying, or removing spending trends
from the predictive model. More specifically, when a predicted
consumer accepts or rejects an offer and/or promotion, data is
transmitted to a bank, a merchant, and/or the interchange network
indicating the acceptance or rejection by the consumer. For
example, when the predicted consumer uses the card to accept the
offer and/or promotion by making a purchase using the card, data is
fed back to the predictive model such that the predictive model can
be refined 216. Feedback on offer uptake can be provided to the
predictive model by the bank, merchant, and/or consumer.
[0051] When the predicted consumer is experiencing the life event
as predicted by the predictive model, the predictive model is
verified and/or modified. In one embodiment, actual spending trends
of predicted consumers accepting offers are used to add, verify, or
remove spending trends within the predictive model. More
specifically, by analyzing actual spending of consumers predicted
to experience the life event, spending variables defining the life
event can be added, verified, or removed from a consumption
bundle.
[0052] Further, as predicted consumers begin to experience the life
event, the consumers will be added to the sample group, as
described in more detail below. As such, the predicted consumers
will become the consumers used to generate 208 the predictive
model. As actual spending by a consumer in the sample group
achieves a new average spending profile, the consumer will be
removed from the sample group. More specifically, a variance, as
described in more detail below, for at least some consumers within
the sample group reduces, and such consumers are removed from the
sample group. As such, as consumers start exhibiting different
spending trends before experiencing a life event, the predictive
model is refined 216 to reflect such changes in spending behaviors.
Accordingly, the predictive model gains or removes spending trends
as actual consumer spending changes.
[0053] Moreover, by analyzing a group of the predicted consumers
for whether predicted consumers exhibit the variance, the accuracy
of the predictive model can be verified and/or refined 216. More
specifically, the variance indicates that the life event is
occurring. As such, predicted consumers that exhibit a variance
verify that the predictive model was accurate for those consumers.
Historical spending of such consumers can be analyzed to determine
additional spending trends for inclusion in the predictive model.
Predicted consumers that do not exhibit a variance can be used to
refine 216 the predictive model to exclude such consumers from
being predicted in the future.
[0054] FIG. 5 is a flowchart illustrating an exemplary method 300
for recording 202 consumer data in a database that may be used with
method 200 (shown in FIG. 4). Method 300 includes recording 302
card transaction data for each card purchase made by each consumer
having an account in the interchange network 28. Method 300
optionally includes recording 304 third party data about a consumer
having an account in the network 28. Card transaction data and/or
third party data are recorded 302 and/or 304 in database 120 (shown
in FIG. 2). Cards transaction data includes, but is not limited to
including, merchant name/type, transaction time, transaction date,
such as a purchase date or a post date, and/or the amount spent. In
one embodiment, the transaction date, transaction amount, and
spending variable designation for a purchase are recorded 302. Card
transaction data can be signals derived directly from transaction
data captured at a point of sale device and/or other suitable
device and recorded 302 to the database 120. Captured card
transaction data may, when necessary, be converted into numerical
form.
[0055] More specifically, the captured data can be in numerical
form and/or any other suitable form. Examples of numerical form
data include, without limitation, the amount of a transaction, a
date of a transaction, and/or any other number or series of
numbers. An example of non-numerical data includes, without
limitation, categorical data, such as an identity of the merchant,
a location of the merchant, a type of item purchased, a description
of the item purchased, retail category, retail industry, and/or any
other suitable data. When the captured data is not in numeric form,
the data may need to be converted into numerical form. For example,
different merchants and/or different types of merchants may be
designated by a merchant number and/or a numeric merchant type
designation. As such, the merchant and/or the merchant type may be
processed using the systems and methods described herein.
[0056] Optionally, additional data provided by at least one third
party, such as a bank, merchant, consumer, and/or other third
party, is also recorded 304 in database 120. For example,
demographic information and/or potentially psychographic data, such
as age, occupation, education, income, gender, home address, and/or
other demographic data, is recorded 304 in the database 120.
[0057] FIG. 6 is a flowchart illustrating an exemplary method 400
for defining 204 a life event that may be used with method 200
(shown in FIG. 4). Method 400 includes selecting 402 a life event a
consumer may experience, such as "new baby," "child to college," or
"new home." For the selected life event, at least one spending
variable is assigned 404 to the life event to form a consumption
bundle for the life event. The spending variables assigned 404 to
the life event include categories of spending that are related to
the life event. In one embodiment, the spending variables assigned
404 to the life event are intuitively selected. In an alternative
embodiment, the spending variables assigned 404 to the life event
are empirically selected.
[0058] In the exemplary embodiment, spending variables are selected
to be in a consumption bundle such that the consumption bundle
substantially proxies the selected life event. More specifically, a
consumption bundle is a group of spending variables, such as
merchants and/or merchant categories, associated with the onset of
a life-event. For example, a consumption bundle for new baby
includes merchants that sell maternity clothes and baby
products.
[0059] In one embodiment, in which spending variables are assigned
404 empirically to the life event, merchants and/or industries are
recursively grouped by lift by: (1) from any starting seed
industry, industry A (e.g. industry A=maternity clothes); (2) find
a second industry B which is mostly purchased together with
industry A (e.g. industry B=toys); and (3) group industries A and B
together as the new seed and repeat steps 1 and 2 until a lower
lift threshold (e.g., a maximum number of industries) is reached. A
conservative definition of a target life event is used to minimize
classification error or false target error. More specifically, a
consumption bundle is limited to those cases most likely to be
indicative of the target life event. The construction of a
consumption bundle associated with a life event is unique to each
life event being modeled.
[0060] Optionally, at least one demographic variable is assigned
406 to the life event. As used herein, the term "demographic
variable" refers to a value for a demographic category, such as
age, occupation, education, income, gender, home address, and/or
other demographic characteristic. For example, the life event "new
bay" may be refined to "new child" or "new grandchild" by assigning
406 an age variable to the life event. In the exemplary embodiment,
more than one life event can be defined 204 by assigning 404 and/or
406 respective variables to each of a plurality of life events.
[0061] FIG. 7 is a flowchart illustrating an exemplary method 500
for determining 206 a sample group of consumers that may be used
with method 200 (shown in FIG. 4). FIG. 8 is a graph 600 of
historical spending, anticipated spending, and actual spending for
a consumer in terms of money spent with respect to time. Money
spent may be in any suitable currency, such as U.S. dollars, and
time may be in any suitable measurement of time, such as days.
Graph 600 shows a sum of spending in all spending variables
assigned to a consumption bundle of a life event. Alternatively,
spending in each spending variable of the consumption bundle may be
graphed separately.
[0062] Method 500 includes, for a consumer of the global
population, analyzing 502 actual spending during a historical time
period 602 in each spending variable of the consumption bundle for
the life event. More specifically, in one embodiment, the global
population is defined as including consumers having accounts with
multiple transactions per month across multiple industries and that
have been active for at least a year. In a particular embodiment,
the global population includes consumers that have an account that
is active in the current month, that has been open for thirteen or
more months, and that averages five or more transactions per month
since opening the account. Further, in one embodiment, each
consumer's spending for the past three years in each spending
variable is analyzed 502. In the exemplary embodiment, the actual
spending for the whole consumption bundle for the life event is
summed 504 for each day of the historical time period 602 and
plotted 506 with respect to time. Such a plot is shown in FIG. 8 as
historical actual spending curve 604.
[0063] From the historical actual spending curve 604, anticipated
spending for a predetermined time period 606 into the future, or an
extrapolation time period 606, is determined for each consumer of
the global population by extrapolating 508 from the historical
actual spending for a respective consumer. More specifically, the
historical actual spending curve 604 is extrapolated 508 to
generate an anticipated spending curve 608. In one embodiment,
historical actual spending curve 604 is extrapolated one month to
three months into the future to generate anticipated spending curve
608. In the exemplary embodiment, anticipated spending curve 608
represents anticipated spending in all spending variables assigned
to the consumption bundle of the life event.
[0064] The consumer's actual spending during the extrapolation time
period 606 is determined 510. More specifically, based on recorded
transaction data, the consumer's spending in the spending variables
of the consumption bundle are summed for each day during the
extrapolation time period 606 and are plotted on graph 600 as an
actual spending curve 610. In the exemplary embodiment, each
consumer of the global population will have a respective graph 600
showing each consumer's historical actual spending, anticipated
spending, and actual spending. Further, in the exemplary
embodiment, after the extrapolation time period 606 expires, steps
504-510 are repeated for a subsequent extrapolation time period. As
such, each consumer of the global population is continuously being
monitored for an occurrence of the life event. In one example,
consumers are re-evaluated every month to determine if a consumer
is experiencing the life event.
[0065] After historical actual spending curve 604, anticipated
spending curve 608, and actual spending curve 610 are determined
506, 508, and 510 for each consumer of the global population, each
consumer's actual spending curve 610 is compared 512 his/her
anticipated spending curve 608 for the extrapolated time period 606
to determine 514 a variance 612 between actual spending and
anticipated spending. In one embodiment, the variance 612 is
determined 514 by comparing 512 an average daily anticipated spend
with an average daily actual spend. In an alternative embodiment,
the variance 612 is determined 514 by comparing 512 by total
anticipated spending with total actual spending during the
extrapolated time period 606. In still another embodiment, any
suitable statistic method and/or technique is used to determine 514
the variance 612 between anticipated spending curve 608 and actual
spending curve 610. In one example, a consumer who has never had a
child and has never purchased baby products or maternity clothes
will have fairly low anticipated spending in the new baby purchase
bundle while someone who is the mother of five young children will
likely have higher anticipated spending in the new baby purchase
bundle. As such, by comparing each consumer to his/her own
historical actual spending, consumers experiencing the life can be
determined.
[0066] One example of comparing 512 anticipated spending to actual
spending is to use residual analysis. Residual analysis, as
referred to herein, is the identification of significant
differences, beyond some threshold, between anticipated and actual
spending behavior. If a consumer's, or an account holder's, actual
spending in a given consumption bundle deviates significantly from
anticipated spending in that consumption bundle, then it is assumed
that the consumer is experiencing the relevant life event. More
specifically, in residual analysis a normalized residual R(t) is
defined as the difference in actual spending S(t) and anticipated
spending P(t), divided by the anticipated spending P(t) in month
t.
R ( t ) = S ( t ) - P ( t ) P ( t ) , ( Eq . 1 ) ##EQU00001##
The onset of an event is indicated the first time a calculated
residual R(t) surpasses, or falls below, a predetermined threshold
c. For example, given c=10 the onset of an event occurs when R(t)
is greater than 10. The results of the residual analysis include
the determined variance 612 between the anticipated spending curve
608 and the actual spending curve 610 of the consumer in the
spending variables of the consumption bundle.
[0067] As such, in one embodiment of the systems and methods
described herein, the normal spending habits of a transaction card
user are tracked to determine if the consumer is experiencing the
life event. These normal spending habits, such as how much a person
spends and/or how often the person spends, are based on historical
use of the card to make purchases. A determination of experiencing
the life event is made when deviations in spending occur. Such
deviations, as indicated by the variance 612, can include, without
limitation, increases in spending amount, decreases in spending
amount, increases in spending frequency, and/or decreases in
spending frequency. Further, such deviations indicate a change in
consumer needs and/or behavior, including possibly predicting
future needs or behavior. Accordingly, in one embodiment, when
deviations occur as indicated by the variance 612, it is assumed
that something out of the ordinary has occurred for a particular
consumer. The meaning of one or more deviations can then be
determined, for example, based on additional data, such as, without
limitation, merchants shopped and/or direct consumer surveys.
[0068] In the exemplary embodiment, when the variance 612 is below
a predetermined threshold and/or within predetermined criteria, the
consumer continues to be monitored for the occurrence of the life
event by repeating steps 502-514. When the variance 612 is above
the predetermined threshold and/or not within predetermined
criteria, it is determined 516 that the consumer is experiencing
the life event. Optionally, when demographic variables have been
assigned 406 (shown in FIG. 6) to the life event, the consumer's
demographic data is also analyzed 518 to determine if the consumer
is experiencing the life event.
[0069] Consumers who are determined 516 and/or 518 to be
experiencing the life event are assigned 520 to a sample group. As
such, the sample group includes consumers who, based on a
respective variance 612, are experiencing or have recently
experienced the life event. Based on continued monitoring of each
consumer's anticipated spending and actual spending, consumers are
added to or removed from the sample group. In the exemplary
embodiment, only consumers who have had a variance 612 in the
consumption bundle exceeding a predetermined threshold for a
predetermined time period are assigned 520 to the sample group. In
one example, consumers having a variance 612 in the consumption
bundle for the life event for two consecutive months wherein the
variance 612 indicates actual spending that is at least three times
the anticipated spending are assigned 520 to the sample group.
[0070] FIG. 9 is a flowchart illustrating an exemplary method 700
for generating 208 a predictive model that may be used with method
200 (shown in FIG. 4). Method 700 includes analyzing 702 each
consumer in the sample group to determine spending trends that are
common among the consumers in the sample group. More specifically,
card purchases for each consumer over a historical time period are
analyzed 702 for spending trends. In one embodiment, each
consumer's card purchases for the last three months to one year are
analyzed 702 to determine spending trends for each consumer. In the
exemplary embodiment, mathematical techniques are used to analyze
over six-hundred variables and to identify the changes in spending
behavior, or spending trends, that are most predictive of the life
event. More specifically, the spending trends for each consumer are
compared to the spending trends of other consumers within the
sample group, as described in more detail below. When a
predetermined percentage and/or other threshold of consumers within
the sample group have the same spending trend, the spending trend
is considered to be a common spending trend.
[0071] Using the common spending trends, a predictive model is
generated 704. In the exemplary embodiment, the predictive model
uses a logistic technique in which scores for missing targets rank
similarly to detected targets. Further, in the exemplary
embodiment, the predictive model includes the common spending
trends such that consumers having similar spending trends can be
predicted to experience the life event. As such, the predictive
model is a set of spending trends and logic.
[0072] FIG. 10 is a flowchart illustrating an exemplary method 800
for predicting 210 whether a consumer will experience a life event
that may be used with method 200 (shown in FIG. 4). Method 800
includes applying 802 the predictive model to each consumer having
an account on the interchange network 28 (shown in FIG. 1) to
determine if a consumer is going to experience the life event. More
specifically, by comparing 804 spending trends within the
predictive model to consumers' actual spending trends over a
historical time period, the occurrence of the life event for a
consumer can be determined 806 before a variance 612 (shown in FIG.
8) in the consumer's actual spending and anticipated spending
occurs.
[0073] In the exemplary embodiment, the numeric captured
transaction card information that has been recorded 202 (shown in
FIG. 4) is processed to determine if any changes in purchasing
behavior are predicted for each consumer having a card account.
More specifically, the system 100 and/or 122 (shown in FIGS. 2 and
3) captures and records 202 information on customer spending using
a card, and the predictive model uses the recorded data to identify
signals that suggest a change in need and/or behavior of a consumer
before the change occurs. In one embodiment available purchasing
data is analyzed such that changes in consumer needs are
anticipated and changes in consumer needs and/or behaviors are
predicted. As such, past spending behavior may be used to predict
future spending behaviors, preferences, and/or needs. Further, the
past spending behavior may be used to identify products and/or
services that may be best suited for specific consumers.
[0074] Referring to FIG. 4, examples of method 200 are described
below.
[0075] Child to College Example
[0076] In this example, the predictive model is configured to
predict which account holders are most likely to send a child to
college in the coming months. For example, consumer purchasing data
is used to identify a family whose children are about to leave for
college. In the exemplary embodiment, the sample group is
determined by using a consumption bundle including the following
spending variables: purchases in the university category, purchases
at merchants on college campuses, spending occurring in August
and/or September, spending is less than 50% of total card spending
for a month, purchases of college prep courses, and purchase made
in college towns.
[0077] Spending trends exhibited by consumers in the sample group
for inclusion in the predictive model include: increases in
consumer electronics purchases, family apparel purchases, book
store purchases, airline purchases, sporting good purchases, and/or
software and network service purchases. Consumers within the global
population exhibiting similar spending trends, whether or not a
variance in the consumption bundle has occurred, are predicted to
send a child to college in the near future, such as the next three
months to one year.
[0078] Based on the determination that a family has at least one
child who is about to leave for college, the system 100 and/or 122
(shown in FIGS. 2 and 3) may recommend products or services aimed
at "empty nesters" and young students. It is likely that the needs
of the parents and children will soon be changing and the
predictive model is configured to predict these changes before the
change actually happens. As such, valuable products or services can
be offered in anticipation of these changes. Further, feedback into
the predictive model based on which promotions or coupons are used
by the consumer is used to refine the model. For example, if
families sending children to college tend to accept offers for
student cards but reject "empty nester" cards, it may make sense to
continue providing the former and discontinue offering the
latter.
[0079] By adding the optional demographic variables to the
consumption bundle, the predictive model may be able to distinguish
between first or last child leaving for college, and/or between
grandchild or child leaving for college. In one embodiment, the
"child to college" predictive model is 4.3 times more likely to
predict a child leaving for college than other known marketing
models, and has a lead time of about six months for the
prediction.
[0080] New Baby Example
[0081] In this example, the predictive model is configured to
predict which account holders are most likely to have a new baby in
the coming months. In the exemplary embodiment, the sample group is
determined by using a consumption bundle including the following
spending variables: specialty merchants selling maternity wear,
merchants selling baby products, specialty children's merchants,
and spending variance lasts for at least two consecutive
months.
[0082] Spending trends exhibited by consumers in the sample group
for inclusion in the predictive model include: increases in
children's apparel purchases, toy store purchases, overall card
spending and number of purchases, and family apparel purchases, and
a decrease in restaurant purchases. Consumers within the global
population exhibiting similar spending trends, whether or not a
variance in the consumption bundle has occurred, are predicted to
have a new baby in the near future, such as the next three months
to one year.
[0083] By adding the optional demographic variables to the
consumption bundle, the predictive model may be able to distinguish
between first or subsequent child, and/or between grandchild or
child. In one embodiment, the "new baby" predictive model is 4.2
times more likely to predict a consumer having a new baby than
other known marketing models, and has a lead time of about three or
more months for the prediction.
[0084] New Home Example
[0085] In this example, the predictive model is configured to
predict which account holders are most likely to change their
residential zip code in the coming months. In the exemplary
embodiment, the sample group is determined by using a consumption
bundle including the following spending variables: change in
residential zip code (measured using bundle of everyday spend
merchant categories), five or more months of stable spending in one
zip code followed by five or more months of stable spending in
another zip code.
[0086] Spending trends exhibited by consumers in the sample group
for inclusion in the predictive model include: increases in
restaurant purchases, a number of public administration
transactions, a number of telecommunication transactions, a number
of legal and accounting service transactions, and a number of
jewelry and giftware transactions, and a decrease in a number of
grocery transactions. Consumers within the global population
exhibiting similar spending trends, whether or not a variance in
the consumption bundle has occurred, are predicted to purchase a
new home in the near future, such as the next three months to one
year.
[0087] In the exemplary embodiment, card transaction data is used
to identify a young couple about to purchase a home. By using the
couple's transaction data, the predictive model can predict that
there is a 60% chance the couple will purchase a house in the next
twelve months. The predictive model may suggest that the couple be
offered a transaction card product best suited for new home owners
(for example, a card that provides discounts or cash back at home
furnishings and home improvement stores). Uptake of offers can be
monitored and used to refine the model.
[0088] In one embodiment, the "new home" predictive model is 2.3
times more likely to predict a purchase of a new home than other
known marketing models, and has a lead time of about three months
for the prediction.
[0089] Other examples of predictive models include: a model
targeted at top of wallet cardholders having thirteen months of
consecutive spending with an average of five or more transaction
per month, a model that can accurately predict a cardholder's
overall spending for next month and/or next quarter to identify
emerging affluent, a model that can accurately predict cardholder
spending by merchant category for the next month and/or the next
quarter to target for commerce coalition, and a model that predicts
cardholder inactivity next month and/or next quarter to identify
"at risk" cardholders for retention efforts.
[0090] The above-described methods and systems facilitate
proactively refreshing value propositions offered to consumers to
increase card relevance and reduce attrition. As such, card holders
will receive less irrelevant offers and more relevant offers, which
benefits both the cardholders and the offerors. The predictive
models described herein provide valuable information to issuers and
merchants to help them cross-sell to consumers. Issuers and/or
merchant may gain a competitive advantage over competitors because
the embodiments described herein enable issuers and/or merchants to
reach consumers before they experience the life event and,
accordingly, before the competition has the chance to market to
these consumers.
[0091] According to some embodiments of the systems and methods
described herein, consumer needs and/or behaviors are predicted
based on historical, anticipated, and actual customer spending
using transaction cards. Transactions using the cards may provide a
large amount of data about consumer spending. Further, spending
data from cards may, in some cases, be more readily available than
demographic data, which may be incomplete, inaccurate, and/or
infrequently updated. In the embodiments described herein,
transaction data is used to predict which consumers will experience
a life event and how consumer spending behavior may change when the
life event is experienced. Further, modeling and/or analytical
algorithms are used in the above-described embodiments to identify
trends and make recommendations. For example, when transaction date
and amount data predicts that a change has occurred, or will occur,
additional data, such as merchant name, type, demographic data,
and/or any other suitable data, can be used to predict what type of
change has occurred. As more predictions are made by the predictive
model, it may be possible to predict what change in customer need
has occurred based on, for example, transaction date and amount
without additional transaction or demographic data.
[0092] Exemplary embodiments of methods and systems for predicting
consumer behavior from transaction card purchases are described
above in detail. The methods and systems are not limited to the
specific embodiments described herein, but rather, components of
the systems and/or steps of the methods may be utilized
independently and separately from other components and/or steps
described herein. For example, the methods may also be used in
combination with other targeting systems and methods, and are not
limited to practice with only the targeting systems and methods
based on transaction card purchases as described herein. Rather,
the exemplary embodiment can be implemented and utilized in
connection with many other target marketing applications.
[0093] Although specific features of various embodiments of the
invention may be shown in some drawings and not in others, this is
for convenience only. In accordance with the principles of the
invention, any feature of a drawing may be referenced and/or
claimed in combination with any feature of any other drawing.
[0094] This written description uses examples to disclose the
invention, including the best mode, and also to enable any person
skilled in the art to practice the invention, including making and
using any devices or systems and performing any incorporated
methods. The patentable scope of the invention is defined by the
claims, and may include other examples that occur to those skilled
in the art. Such other examples are intended to be within the scope
of the claims if they have structural elements that do not differ
from the literal language of the claims, or if they include
equivalent structural elements with insubstantial differences from
the literal language of the claims.
* * * * *