FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENTS RULES, 2003
COMPLETE SPECIFICATION
(See section 10, rule 13)
1. Title of the invention: STOCK KEEPING UNIT RATIONALIZATION
2. Applicant(s)
NAME NATIONALITY ADDRESS
TATA CONSULTANCY Indian Nirmal Building, 9th Floor, Nariman
SERVICES LIMITED Point, Mumbai, Maharashtra 400021,
India
3. Preamble to the description
COMPLETE SPECIFICATION
The following specification particularly describes the invention and the manner in which it
is to be performed.

TECHNICALFIELD
[0001] The present subject matter relates, in general, to stock keeping units and, in
particular, to systems and computer-implemented methods for stock keeping unit rationalization.
BACKGROUND
[0002] There are ongoing efforts in the retail sector to reduce unnecessary investments
in inventory. Assortment planning is one of the ways to reduce unnecessary investments in inventory. Assortment can be defined as the combination of all products available in a retail store. A set of products may be combined together to define a product category and each product may be associated with a number of brands. For example, liquor may be a product category, and beer, whiskey, and vodka may be a set of products. Further, each brand of a product may be assigned one or more unique identification numbers based on its attributes, such as size, color, packaging style, etc. This unique identification number is referred to as stock keeping unit (SKU).
[0003] There has been a sharp increase in the number of products sold in retail stores
due to consumer tastes and preferences, competition in the market, introduction of new products, and multiple selling channels. With this increase in the number of products, inventory and shelf cost has increased. Further, shelf space for displaying and storing the products has reduced. In such a situation, assortment planning helps retailers to overcome these challenges by periodically reviewing and revising the assortment they carry, to keep an account of changes in consumer demand over time as well as new products introduced by suppliers. Assortment planning can also improve shelf space availability and state the budget for each product followed by sales goals.
[0004] Several assortment planning techniques have been developed in the past few
years. One such technique is a SKU rationalization. SKU rationalization is an inventory management technique that helps retailers optimize their assortments by eliminating one or more SKUs based on one or more parameters, such as sales, shelf space and performance. Conventional techniques of SKU rationalization involve elimination of one or more SKUs based on survey data. However, such techniques do not take into account factors like

consumer choice. This calls for a better SKU rationalization technique so as to overcome the shortcomings of the existing techniques.
SUMMARY
[0005] This summary is provided to introduce concepts related to stock keeping unit
(SKU) rationalization. These concepts are further described below in the detailed description. This summary is not intended to identify essential features of the claimed subject matter nor is it intended for use in determining or limiting the scope of the claimed subject matter.
[0006] In one embodiment, the method for SKU rationalization comprises
determining a most preferred SKU and at least one alternative SKU from amongst the plurality of SKUs for each of a plurality of consumers based on a sales data set. The method further comprises, comparing a current satisfaction score and a new satisfaction score for each of the plurality of consumers upon purchase of the most preferred SKU and the at least one alternative SKU respectively. Once the current satisfaction score and the new satisfaction score are compared, for each of the plurality of SKUs, a total number of store-switchers from amongst the plurality of consumers are estimated. Based on the estimation, at least one SKU from amongst the plurality of SKUs is delisted.
BRIEF DESCRIPTION OF THE DRAWINGS
[0007] The detailed description is described with reference to the accompanying
figure(s). In the figure(s), the left-most digit(s) of a reference number identifies the figure in
which the reference number first appears. The same numbers are used throughout the figure(s)
to reference like features and components. Some embodiments of systems and/or methods in
accordance with embodiments of the present subject matter are now described, by way of
example only, and with reference to the accompanying figure(s), in which:
[0008] Fig.1 illustrates a network environment implementing a stock keeping unit
(SKU) rationalization system, according to an embodiment of the present subject matter.
[0009] Fig. 2a illustrates components of the SKU rationalization system, according to
an embodiment of the present subject matter.
[0010] Fig. 2b illustrates an example graphical representation of store-switchers
estimation, using a store-switching condition.

[0011] Fig. 3 illustrates a method for SKU rationalization, according to an
embodiment of the present subject matter.
DETAILED DESCRIPTION
[0012] Conventionally, various assortment planning techniques, such as SKU
rationalization, are available, that help retailers optimize their assortments by eliminating one or more SKUs. Such assortment planning techniques, however, suffer from numerous drawbacks. For example, such assortment planning techniques simply eliminate low selling SKUs without prior experimentation. In such a scenario, certain group of consumers that demand the eliminated or delisted low selling SKUs, switch to other retail stores due to nonavailability of the preferred SKUs. Therefore, this may have a negative impact on profitability of the retail stores.
[0013] Other advanced assortment planning techniques delist the SKUs based on
survey data by considering individual level choices of consumers. However, survey data does not reflect repeated purchase behavior of consumers. Therefore, the survey data might fail to reflect actual consumer taste and preferences. Moreover, such assortment planning techniques based on survey data are costly.
[0014] In accordance with the present subject matter, a system and a computer-
implemented method for stock keeping unit (SKU) rationalization are described. The system as described herein is a SKU rationalization system, hereinafter referred to as rationalization system. The term SKU rationalization may be understood as developing an optimal assortment plan by delisting one or more SKUs, which have minimal impact on consumer satisfaction and profitability, from an assortment and reduce a number of store-switchers as a result of the SKU rationalization.
[0015] Initially, a database for storing panel data associated with one or more retail
stores is maintained according to one implementation. The database can be an external repository associated with the rationalization system, or an internal repository within the rationalization system. The panel data may include data associated with at least one product category of one or more retail stores. The product category may include grocery, beverages, or durables, such as home goods, personal care products, liquor, baby care products, food,

home cleaning products and the like. The product category may be segmented into one or
more products, and each product may be associated with at least one brand.
[0016] In one implementation, the panel data may also include identification codes of
a plurality of consumers, a plurality of SKUs associated with each brand of the product,
number of trips made by each consumer to one or more retail stores, etc. In an example, panel
data may include data associated with a product category, such as liquor and product may be a
beer. The beer may have a plurality of brands, such as beer A, beer B, and beer C, and one or
more SKUs may be associated with each beer brand. In said example, beer A may be
associated with 2 SKUs, such as SKU1 and SKU2. Similarly, beer B may be associated with
SKU3, and beer C may be associated with SKU4 and SKU5. The panel data may be retrieved
whenever the SKU rationalization is to be performed. Further, the panel data contained within
such database may be updated periodically or whenever required. For example, new data may
be added into the database, existing data can be modified, or non-useful data may be deleted
from the database. In one example, the panel data may be obtained from a retailer.
[0017] In one implementation, a sales data set is generated based on the panel data. In
one implementation, sales data set for a plurality of SKUs associated with the at least one product category is generated. In another implementation, sales data set for a plurality of SKUs associated with the one or more products of the at least one product category is generated. In yet another implementation, sales data set for a plurality of SKUs associated with at least one brand of one or more products is generated. To generate the sales data set, panel data of one or more retail stores for a predefined period, say, past one year are retrieved from the database. The generated sales data set, for example, may include sales data of one or more SKUs by a plurality of consumers. As indicated in previous example, beer A has SKU1 and SKU2, beer B has SKU3, and beer C has SKU4 and SKU5. Accordingly, the sales data set may include number of times each of the SKU1, SKU2, SKU3, SKU4, and SKU5 are purchased by a plurality of users. The purchase decision of a SKU may be driven by various factors, such as type of brand, packaging style, display or feature, size, popularity etc. For example, if a consumer drinks small quantity of beer, then he may prefer a can of beer over a bottle.
[0018] Thereafter, for each of the plurality of consumers, a most preferred SKU and at
least one alternative SKU from amongst the plurality of SKUs are determined based on the

generated sales data set. In one implementation, the determination of the most preferred SKU
and at least one alternative SKU from amongst the plurality of SKUs is based on historical
data of repeat purchase behavior. A most preferred SKU may be understood as a SKU from
which a consumer derives maximum utility. When a consumer does not find his most
preferred SKU in a retail store, he may switch to another SKU, referred to as alternative SKU
or may switch the retail store. In one implementation, SKU that may have been purchased for
maximum number of times amongst the plurality of SKUs is determined as the most preferred
SKU and at least one of the remaining SKUs may be determined as an alternative SKU. In an
example, SKU1, SKU2, SKU3, SKU4, and SKU5 may have been purchased 50, 20, 70, 80,
and 45 number of times respectively. Therefore, SKU4 may be the most preferred SKU and at
least one of the SKU1, SKU2, SKU3, and SKU5 may be the alternative SKU.
[0019] Further, a current satisfaction score and a new satisfaction score are computed
for each of the plurality of consumers upon purchase of the most preferred SKU and the at
least one alternative SKU respectively. For example, upon purchase of the most preferred
SKU by the plurality of consumers, a current satisfaction score is computed for each of the
plurality of consumers. The computation of the current satisfaction score, in one
implementation, may take place using a conventionally known Bayesian Hierarchical Logit
(BHL) Model. Based on taste and preference of each consumer and repeat purchase behavior
of the consumers, the BHL model generates a plurality of coefficients through iterative
process. Using these coefficients, the BHL model computes the current satisfaction score.
These coefficients may include, for example, brand coefficients and attribute coefficients.
[0020] Further, in the said example, when the most preferred SKU is removed from a
retail store, the each consumer may switch to at least one SKU amongst the remaining plurality of SKUs or may switch the retail store. As indicated previously, if a consumer does not find his most preferred SKU in a retail store, he may switch to another SKU, referred to as an alternative SKU. In case the consumers switch to the alternative SKU, a new satisfaction score is computed for each of the plurality of consumers.
[0021] For computation of the new satisfaction score, a conventionally known
segmented regression algorithm is run for each consumer by regressing the computed current satisfaction score on budget and price of the alternative SKU in each purchase occasion of the consumers. For example, if there are total number of 10 SKUs associated with a product, 10

segmented regression algorithm may be run. For each regression algorithm, three sets of coefficients may be generated, first for income, another for price and third for intercept. Based on these coefficients, the new satisfaction score is computed. The computed current satisfaction score and the new satisfaction score are stored into the database as a score data. The database, thus, contains the panel data and the score data.
[0022] Subsequently, the current satisfaction score and a new satisfaction score are
compared to estimate, for each of the plurality of SKUs, a total number of store-switchers from amongst the plurality of consumers. In one implementation, the total number of store-switchers is estimated based on a store-switching condition. For example, for a consumer, if the new satisfaction score is less than the current satisfaction score, the consumer may switch the retail store. Alternatively, if the new satisfaction score exceeds or closely matches the current satisfaction score, the consumer may switch to the alternative SKU. In case more than one alternative SKUs are present, the consumer may switch to that alternative SKU which has highest satisfaction score. Therefore, based on comparing, the total number of store-switchers is estimated. A consumer who switches the retail store if the new satisfaction score is less than the current satisfaction score is referred to as store-switcher. Based on estimating, one or more SKUs from amongst the plurality of SKUs are delisted from the retail store. The SKUs for which number of store-switchers is minimum may be delisted.
[0023] According to the system and the computer-implemented method of the present
subject matter, one or more SKUs are delisted from an assortment based on taste and
preference of each consumer and repeat purchase behavior of the consumers. Thus, SKUs are
delisted with minimal impact on consumer satisfaction and profitability is achieved. Further,
the system and the computer-implemented method develop an optimal assortment plan to
delist one or more SKUs which further reduces the number of store-switchers.
[0024] The following disclosure describes a system and a method for stock keeping
unit (SKU) rationalization. While aspects of the described system and method can be implemented in any number of different computing systems, environments, and/or configurations, and embodiments for the SKU rationalization are described in the context of the following exemplary system(s) and method(s).
[0025] Fig. 1 illustrates a network environment 100 implementing a stock keeping
unit (SKU) rationalization system 102 (hereinafter referred to as rationalization system 102),

in accordance with an embodiment of the present subject matter. In one implementation, the network environment 100 can be a public network environment, including thousands of personal computers, laptops, various servers, such as blade servers, and other computing devices. In another implementation, the network environment 100 can be a private network environment with a limited number of computing devices, such as personal computers, servers, laptops, and/or communication devices, such as mobile phones and smart phones.
[0026] The rationalization system 102 is communicatively connected to a plurality of
user devices 104-1, 104-2, 104-3...104-N, collectively referred to as user devices 104 and individually referred to as a user device 104, through a network 106. In one implementation, a plurality of users, such as retailers may use the user devices 104 to communicate with the rationalization system 102.
[0027] The rationalization system 102 and the user devices 104 may be implemented
in a variety of computing devices, including, servers, a desktop personal computer, a notebook or portable computer, a workstation, a mainframe computer, a laptop and/or communication device, such as mobile phones and smart phones. Further, in one implementation, the rationalization system 102 may be a distributed or centralized network system in which different computing devices may host one or more of the hardware or software components of the rationalization system 102.
[0028] The rationalization system 102 may be connected to the user devices 104 over
the network 106 through one or more communication links. The communication links between the rationalization system 102 and the user devices 104 are enabled through a desired form of communication, for example, via dial-up modem connections, cable links, digital subscriber lines (DSL), wireless, or satellite links, or any other suitable form of communication.
[0029] The network 106 may be a wireless network, a wired network, or a
combination thereof. The network 106 can also be an individual network or a collection of many such individual networks, interconnected with each other and functioning as a single large network, e.g., the Internet or an intranet. The network 106 can be implemented as one of the different types of networks, such as intranet, local area network (LAN), wide area network (WAN), the internet, and such. The network 106 may either be a dedicated network or a

shared network, which represents an association of the different types of networks that use a variety of protocols, for example, Hypertext Transfer Protocol (HTTP), Transmission Control Protocol/Internet Protocol (TCP/IP), etc., to communicate with each other. Further, the network 106 may include network devices, such as network switches, hubs, routers, for providing a link between the rationalization system 102 and the user devices 104. The network devices within the network 106 may interact with the rationalization system 102, and the user devices 104 through the communication links.
[0030] The network environment 100 further comprises a database
108communicatively coupled to the rationalization system 102. In one implementation, the database 108 may store panel data (denoted as panel data 110), of one or more retail stores. As indicated previously, the panel data 110 include, but is not limited to, data associated with at least one product category of one or more retail stores. The product category may include grocery, beverages, or durables, such as home goods, personal care products, liquor, baby care products, food, home cleaning products and the like. The product category may be segmented into one or more products and each product may be associated with at least one brand. The panel data 110 may also include identification codes of a plurality of the consumers, number of SKUs associated with each brand of the product, number of trips made by each consumer to one or more retail stores, etc.
[0031] Although the database 108 is shown external to the rationalization system 102,
it will be appreciated by a person skilled in the art that the database 108 can also be implemented internal to the rationalization system 102, wherein the panel data 110 may be stored within a data component of the rationalization system 102.
[0032] According to an implementation of the present subject matter, the
rationalization system 102 retrieves panel data 110 for a predefined period, say past one year from the database 108. As indicated earlier, the panel data 110 includes data associated with at least one product category of one or more retail stores. The product category may include grocery, beverages, or durables, such as home goods, personal care products, liquor, baby care products, food, home cleaning products and the like. The product category may be segmented into one or more products, and each product may be associated with at least one brand. The panel data may also include identification codes of a plurality of consumers, a plurality of

SKUs associated with each brand of the product, number of trips made by each consumer to one or more retail stores, etc.
[0033] Based on the panel data 110, the rationalization system 102 generates a sales
data set. The sales data set may include sales data of the plurality of SKUs associated with at
least one product category. The purchase decision of a SKU may be driven by various factors,
such as type of brand, packaging style, display or feature, size, popularity etc. For example, if
a consumer drinks small quantity of beer, then he may prefer can of beer over bottle.
[0034] Further, for each of a plurality of consumers, a most preferred SKU and at least
one alternative SKU from amongst the plurality of SKUs is determined based on the generated sales data set. A most preferred SKU may be understood as the SKU from which a consumer derives maximum utility. When a consumer does not find his most preferred SKU in a retail store, he may switch to another SKU, referred to as an alternative SKU or may switch the retail store. In one implementation, SKU that may have been purchased for maximum number of times amongst the plurality of SKUs is determined as the most preferred SKU and at least one of the remaining SKUs may be determined as the alternative SKU. In an example, a product may have a plurality of SKUs, such as SKU1, SKU2, SKU3, SKU4, and SKU5 and these may have been purchased 50, 20, 70, 80, and 45number of times respectively. Therefore, SKU4 may be the most preferred SKU and at least one of the SKU1, SKU2, SKU3, and SKU5 may be the alternative SKU.
[0035] Upon determination of the most preferred SKU and at least one alternative
SKU, a current satisfaction score and a new satisfaction score are computed for each of the
plurality of consumers upon purchase of the most preferred SKU and the at least one
alternative SKU respectively. For example, upon purchase of the most preferred SKU by the
plurality of consumers, a current satisfaction score is computed for each of the plurality of
consumers. In case the most preferred SKU is not available in the retail store, consumers may
switch to one or more alternative SKU. In such a case, a new satisfaction score is computed
for each of the plurality of consumers. In one embodiment, the rationalization system 102
includes a score computation module 112, which is configured to compute the current
satisfaction score and a new satisfaction score for each of the plurality of consumers.
[0036] The computation of the current satisfaction score, in one implementation, may
take place using a conventionally known Bayesian Hierarchical Logit (BHL) Model. Based on

taste and preference of each consumer and repeated purchase behavior, the BHL model generates a plurality of coefficients through iterative process. These coefficients may include, for example, brand coefficients and attribute coefficients. Using these coefficients, the score computation module 112 computes the current satisfaction score.
[0037] Further, the score computation module 112 computes the new satisfaction
score for each of the plurality of consumers using a conventionally known segmented regression algorithm. The segmented regression algorithm is run for each consumer by regressing the computed current satisfaction score on budget and price for the alternative SKU in each purchase occasion of the consumers. For each regression algorithm, three sets of coefficients may be generated, first for income, another for price and third for intercept. Based on these coefficients, the new satisfaction score is computed.
[0038] Once the current satisfaction score and the new satisfaction score are
computed, the scores are compared to estimate for each of the plurality of SKUs, a total number of store-switchers from amongst the plurality of consumers. A consumer who switches the retail store is referred to as a store-switcher. For example, for a consumer, if the new satisfaction score is less than the current satisfaction score, the consumer may switch the retail store. Alternatively, if the new satisfaction score exceeds or closely matches the current satisfaction score, the consumer may switch to the alternative SKU. Therefore, based on comparing, the total number of store-switchers is estimated. In one implementation, the number of store-switchers is estimated based on a store-switching condition. Based on the total number of store-switchers, the rationalization system 102 delists one or more SKUs from amongst the plurality of SKUs from one or more retail stores.
[0039] The manner in which the SKU rationalization takes place is explained in
greater detail according to the Fig. 2a.
[0040] Fig. 2a illustrates various components of the rationalization system 102,
according to an embodiment of the present subject matter.
[0041] In said embodiment, the rationalization system 102 includes one or more
processor(s) 202, interface(s) 204, and a memory 206 coupled to the processor(s) 202. The processor(s) 202 may be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, state machines, logic

circuitries, and/or any devices that manipulate signals based on operational instructions. Among other capabilities, the processor(s) 202 are configured to fetch and execute computer-readable instructions and data stored in the memory 206.
[0042] The functions of the various elements shown in the figure, including any
functional blocks labeled as “processor(s)”, may be provided through the use of dedicated hardware as well as hardware capable of executing software in association with appropriate software. When provided by a processor, the functions may be provided by a single dedicated processor, by a single shared processor, or by a plurality of individual processors, some of which may be shared. Moreover, explicit use of the term “processor” should not be construed to refer exclusively to hardware capable of executing software, and may implicitly include, without limitation, network processor, application specific integrated circuit (ASIC), field programmable gate array (FPGA), read only memory (ROM) for storing software, random access memory (RAM), non-volatile storage. Other hardware, conventional and/or custom, may also be included.
[0043] The interface(s) 204 may include a variety of software and hardware
interfaces, for example, interface for peripheral device(s), such as a keyboard, a mouse, an external memory, and a printer. Further, the interface(s) 204 may enable the rationalization system 102 to communicate over the network 106, and may include one or more ports for connecting the rationalization system 102 with other computing devices, such as web servers and external databases. The interface(s) 204 may facilitate multiple communications within a wide variety of protocols and networks, such as a network, including wired networks, e.g., LAN, cable, etc., and wireless networks, e.g., WLAN, cellular, satellite, etc.
[0044] The memory 206 may include any computer-readable medium known in the art
including, for example, volatile memory, such as static random access memory (SRAM) and dynamic random access memory (DRAM), and/or non-volatile memory, such as read only memory (ROM), erasable programmable ROM, flash memories, hard disks, optical disks, and magnetic tapes. The rationalization system 102 also includes module(s) 208 and data 210.
[0045] The module(s) 208 include routines, programs, objects, components, data
structures, etc., which perform particular tasks or implement particular abstract data types. The module(s) 208 further include, in addition to the score computation module 112, a stock

keeping unit (SKU) determination module 212, a SKU delisting module 214, and other module(s) 216. The other module(s) 216 may include programs or coded instructions that supplement applications and functions, for example, programs in the operating system of the user device 104 and the rationalization system 102.
[0046] The data 210 serves, amongst other things, as a repository for storing data
processed, received and generated by one or more of the module(s) 208. The data 210 includes the panel data 110, score data 218, and other data 220. The panel data 110 includes data associated with at least one product category of one or more retail stores. The product category may include grocery, beverages, or durables, such as home goods, personal care products, liquor, baby care products, food, home cleaning products and the like. The product category may be segmented into one or more products and each product may be associated with at least one brand. The panel data 110 may also include identification codes of a plurality of the consumers, a plurality of SKUs associated with each brand of the product, number of trips made by each consumer to one or more retail stores etc. The score data 218 includes current satisfaction score and new satisfaction score. The other data 220 includes data generated as a result of the execution of one or more other modules 216. Although it has been described, the panel data 110 includes a plurality of SKUs associated with each brand of the product, however, it would be understood by those skilled in the art that the plurality of SKUs may also be associated with each product or the product category.
[0047] In the present embodiment, the panel data 110 and the score data 218 are
depicted to be stored within the data 210, which is a repository internal to the rationalization system 102. However, as described in the previous embodiment, the panel data 110 and the score data 218 may also be stored in the database 108 that is external to the rationalization system 102.
[0048] According to the present subject matter, the SKU determination module 212 is
configured to retrieve panel data 110 for a predefined period, for example, past one year from the data 210.As indicated previously, the panel data110 may include data associated with at least one product category of one or more retail stores. The product category may include grocery, beverages, or durables, such as home goods, personal care products, liquor, baby care products, food, home cleaning products and the like. The product category may be segmented

into one or more products and each one or more products may be associated with at least one brand. The panel data 110 may also include identification codes of the plurality of the consumers, number of SKUs associated with each brand of the product, number of trips made by each consumer to one or more retail stores etc.
[0049] In an example, the panel data 110 may include data associated with a product
category, such as liquor, of a retail store. The product category may further be segmented into a product, such as beer. There may be three brands of beer, i.e., beer A, beer B, and beer C. Each beer brand may have a plurality of SKUs. In said example, beer A may have 5 SKUs, beer B may have 5 SKUs, and beer C may have 4 SKUs.
[0050] Based on the retrieved panel data 110, the SKU determination module 212
further generates a sales data set. The sales data set may include sales data of a plurality of SKUs associated with at least one brand of each one or more products. The description hereinafter is explained with reference to the plurality of SKUs associated with at least one brand of each one or more products only for the purpose of explanation, and it should not be construed as a limitation, it is well appreciated that the sales data set for the plurality of SKUs associated with one or more products or at least one product category can also be generated for SKU rationalization.
[0051] According to an example, sales data of a plurality of SKUs associated with a
product, such as beer is depicted in Table 1(provided below). According to said example, sales data of the plurality of SKUs associated with three beer brands are considered. The three beer brands may be beer A, beer B, and beer C.

Table 1

PRODUCT BRAND NAME SKU NUMBER SKU NAME NUMBER OF PURCHASES
Beer Beer A 1 Beer A 6 pack bottle 31

2 Beer A 6 pack can 217

3 Beer A 12 pack bottle 128

4 Beer A 12 pack can 40

5 Beer A 24 pack can 62

6 Beer B 6 pack bottle 22

Beer B 7 Beer B 6 pack can 72

8 Beer B 12 pack bottle 90

9 Beer B 12 pack can 200

10 Beer B 24 pack can 74

Beer C 11 Beer C 6 pack bottle 35

12 Beer C 6 pack can 37

13 Beer C 12 pack bottle 109

14 Beer C 12 pack can 40
[0052] As shown in the Table 1 above, 5 SKUs are associated with beer A, 5 SKUs
are associated with beer B, and 4 SKUs are associated with beer C. As is evident from the above table, SKU2, SKU3, SKU9, and SKU13 have been purchased more than 100 times, i.e., SKU2, SKU3, SKU9, and SKU13 have been purchased 217, 128, 200, and 109 times respectively. The purchase decision of an SKU may be driven by various factors, such as type

of brand, packaging style, display or feature, size, popularity etc. For example, if a consumer drinks small quantity of beer, then he may prefer can of beer over bottle.
[0053] According to one implementation, the SKU determination module 212 further
determines for each of a plurality of consumers, a most preferred SKU and at least one alternative SKU from amongst the plurality of SKUs based on the generated data set. The term most preferred SKU may be understood as the SKU from which a consumer derives maximum utility. When the consumer does not find his most preferred SKU in a retail store, he may switch to another SKU, referred to as an alternative SKU or may switch the retail store. In one implementation, the determination of the most preferred SKU and at least one alternative SKU from amongst the plurality of SKUs is based on historical data of repeated purchase behavior. Taking the previous example, as shown in previous table 1, SKU9 has been purchased for maximum number of times. Therefore, SKU9 is the most preferred SKU and SKU2, SKU3, and SKU13 may be alternative SKUs for a group of consumer. For the sake of brevity, only those SKUs are taken into considerations which have been purchased more than 100 times.
[0054] Subsequent to determination of the most preferred SKU and at least one
alternative SKU from amongst the plurality of SKUs, the score computation module 112 of the rationalization system 102 computes a current satisfaction score and a new satisfaction score for each of the plurality of consumers upon purchase of the most preferred SKU and the at least one alternative SKU respectively. For example, upon purchase of the most preferred SKU by the plurality of consumers, a current satisfaction score is computed for each of the plurality of consumers. Further, in the said example, when the most preferred SKU is removed from a retail store, the consumers may switch to at least one alternative SKU or may switch the store. In case when the consumers switch to the alternative SKU, a new satisfaction score is computed for each of the plurality of consumers.
[0055] The current satisfaction score may be computed using a conventionally known
Bayesian Hierarchical Logit (BHL) Model. Based on taste and preference of each consumer and repeated purchase behavior, the BHL model generates a plurality of coefficients through an iterative process. These coefficients may include brand coefficients, outside good coefficient, and attribute coefficients, such as display coefficient, feature coefficient, and packaging style coefficient. For example, if a consumer purchases a beer along with grocery

items, then grocery items may be the outside product. Using these coefficients, the score computation module 112 computes the current satisfaction score using the equation (1) provided below:

where, uch represents the current satisfaction score for hth consumer,
p represents total number of consumers,
n represents total number of brands,
Bih represents ith brand purchased by hth consumer,
Bih coeff represents brand coefficient of ith brand purchased by hth consumer,
A1h, A2h, and A3h represents attributes, such as packaging style, display, and feature
respectively, of a product,
A1h coeff , A2h coeff, and A3hcoeff represents attributes coefficients of above
mentioned attributes for store level promotions,
αh represents expenditure on outside good by hth consumer, where the outside good may be any regular grocery item, and
αhcoeff represents coefficient for expenditure on the outside good by hth consumer.
[0056] The score computation module 112 is further configured to compute the new
satisfaction score for each of the plurality of consumers upon purchase of the at least one alternative SKU. For computation of the new satisfaction score, a conventionally known segmented regression algorithm is run for each consumer by regressing the computed current satisfaction score on budget and price for the alternative SKU in each purchase occasion of the consumers. For example, if there are total number of 10 SKUs associated with a product, 10 segmented regression algorithm may be run. For each regression algorithm, three sets of coefficients may be generated, first for income, another for price and third for intercept. Based on these coefficients, the new satisfaction score is computed. According to one implementation, the score computation module 112 computes the new satisfaction score for each of the plurality of consumers using the equation (2) provided below:
ukh=αokh-α1khlnE(Pk)+γkhlnB+εkh,hЄ{1 ,…,p},ke{1 ,…,m}
…. (2)

where, ukh represents new satisfaction score for hth consumer,
m represents total number of alternative SKUs,
p represents total number of consumers,
α1kh represents impact of price on utility,
αokh represents intercept of the equation,
єkhrepresents the disturbance term,
γkh represents impact of income on utility,
E(Pk) represents price of the alternative SKUs averaged over all purchase
occasions, and
B represents budget.
[0057] In one implementation, the score computation module 112 may store the
computed current satisfaction score and the new satisfaction score within the score data 218.
[0058] Upon computation of the current satisfaction score and the new satisfaction
score, the SKU delisting module 214 estimates for each of the plurality of SKUs, a total number of store-switchers from amongst the plurality of consumers based on a store-switching condition. The number of switchers can be estimated based on one or more rules. In another implementation, the number of switchers can be based on a mapping between the comparative difference between the current and the new satisfaction store. According to the store-switching condition, the current satisfaction score and the new satisfaction score are compared to estimate for each of the plurality of SKUs, the total number of store-switchers from amongst the plurality of consumers.
[0059] According to one implementation, the SKU delisting module 214 derives the
store-switching condition for the plurality of consumers using the equation (3) provided below:

where, ukh represents the new satisfaction score,
uch represents the current satisfaction score,
m represents total number of alternative SKUs, and
p represents total number of consumers.

[0060] In an example, four SKUs are associated with a product, i.e., SKU1, SKU2,
SKU3, and SKU4 and SKU, such that SKU2 is the most preferred SKU and remaining SKUs are alternative SKUs for a consumer. Upon purchasing of SKU2, the current satisfaction score for the consumer may be 5. For alternative SKUs, i.e., SKU1, SKU3, and SKU4, the new satisfaction score may be 4.82, 4.98, and 3.4 respectively. Since the new satisfaction score of any of the alternative SKUs does not match with the current satisfaction score, the consumer may switch the store.
[0061] As would be appreciated, a consumer who switches the retail store if the new
satisfaction score is less than the current satisfaction score is referred to as store-switcher. For
example, for a consumer, if the new satisfaction score is less than the current satisfaction
score, the consumer may switch the retail store. Alternatively, if the new satisfaction score
exceeds or closely matches the current satisfaction score, the consumer may switch to the
alternative SKU. In case more than one alternative SKUs are present, the consumer may
switch to that alternative SKU which has highest satisfaction score. Therefore, based on
comparing, the total number of store-switchers is estimated. Based on estimating, the SKU
delisting module 214 is further configured to delist one or more SKUs from amongst the
plurality of SKUs from one or more retail stores. The SKUs for which number of store-
switchers is minimum may be delisted for the assortment. Since the one or more SKUs are
delisted from the assortment based on taste and preference of each consumer and repeat
purchase behavior of the consumers, minimal impact on consumer satisfaction and
profitability is achieved. Further, number of store-switchers is minimized.
[0062] Fig. 2b illustrates an example graphical representation of store-switchers
estimation using store-switching condition. In an example, the store-switchers are estimated for two SKUs, that is SKU1 and SKU2; and two consumers, that is consumer1 and consumer2. If the most purchased SKU for the consumers is SKU1, then current satisfaction scores for consumer1 and consumer2 are u11 and u12 respectively. Further, if SKU1 is removed from an assortment, then SKU2 is an alternative SKU for consumer1 and consumer2. The new satisfaction score for consumer1 and consumer2 upon purchase of SKU2 are u21 and u22, respectively. As shown in graph 250, the current satisfaction score (u11) and the new satisfaction score (u21) for consumer1, when SKU1 is delisted are plotted on the Y-axis. The expected price E(P2) of SKU2 when SKU1 is dropped and price (P21) at which the new

satisfaction score (u21) upon purchase of SKU2 equals the current satisfaction score (u11) for consumer1 are plotted on the X-axis. As shown in the graph 250, the new satisfaction score (u21) (when SKU1 is delisted) is greater than the current satisfaction score (u11) for the consumer1. Also, the expected price E(P2) is less than the price (P21). Therefore, the consumer1 may switch to SKU2. Similarly, as shown in graph 260, the current satisfaction score (u12) and the new satisfaction score (u22) for consumer2, when SKU1 is delisted are plotted on the Y-axis. The expected price E(P2) of SKU2 when SKU1 is dropped and price (P22) at which the new satisfaction score (u22) upon purchase of SKU2 equals the current satisfaction score (u12) for consumer2 is plotted on the X-axis. As shown in the graph 260, the new satisfaction score (u22) (when SKU1 is delisted) is less than the current satisfaction score (u12) for the consumer2. Also, the expected price E(P2) is greater than the price (P22) at which the new satisfaction score (u22) by purchase of SKU2 equals the current satisfaction score (u12). Therefore, the consumer2 may switch the store.
[0063] Fig. 3 illustrates a method 300 for stock keeping unit (SKU) rationalization, in
accordance to an embodiment of the present subject matter. The method 300 is implemented in computing device, such as a SKU rationalization system 102. The method may be described in the general context of computer executable instructions. Generally, computer executable instructions can include routines, programs, objects, components, data structures, procedures, modules, functions, etc., that perform particular functions or implement particular abstract data types. The method may also be practiced in a distributed computing environment where functions are performed by remote processing devices that are linked through a communications network.
[0064] The order in which the method 300 is described is not intended to be construed
as a limitation, and any number of the described method blocks can be combined in any order to implement the method 300, or alternative methods. Additionally, individual blocks may be deleted from the methods without departing from the spirit and scope of the subject matter described herein. Furthermore, the method 300 can be implemented in any suitable hardware, software, firmware, or combination thereof.
[0065] Referring to Fig. 3, at block 302, the method 300 includes retrieving panel data
110 from a database 108 for generating a sales data set. The panel data 110 may include data associated with at least one product category of one or more retail stores. The generated sales

data set may include sales data of a plurality of SKUs associated with the at least one product category. In accordance with one implementation of the present subject matter, the SKU determination module 212 generates the sales data set comprising sales data of a plurality of SKUs associated with at least one product category.
[0066] At block 304, the method 300 includes determining for each of a plurality of
consumers, a most preferred SKU and at least one alternative SKU from amongst the plurality of SKUs based on the dataset. A most preferred SKU may be understood as the SKU from which a consumer derives maximum utility. When a consumer does not find his most preferred SKU in a retail store, he may switch to another SKU, referred to as alternative SKU. In one example, a most preferred SKU and an alternative SKU may be determined based on number of purchases of each SKU by a plurality of consumers. In said example, SKU which is purchased for maximum number of times may be the most preferred SKU. In one implementation, the SKU determination module 212 determines the most preferred SKU and at least one alternative SKU from amongst the plurality of SKUs for each of the plurality of consumers.
[0067] At block 306, the method 300 includes comparing a current satisfaction score
and a new satisfaction score for each of the plurality of consumers upon purchase of the most preferred SKU and the at least one alternative SKU respectively. The current satisfaction score is computed for each of the plurality of consumers upon purchase of the most preferred SKU. In case when the consumers switch to the alternative SKU due to non-availability of the most preferred SKU, a new satisfaction score is computed for each of the plurality of consumers. The current satisfaction score and the new satisfaction score are then compared to estimate for each of the plurality of SKUs a total number of store-switchers from amongst the plurality of consumers. In one implementation, the score computation module 112 of the rationalization system 102 is configured to compute the current satisfaction score and the new satisfaction score for each of the plurality of consumers and stores the computed scores in score data 218.
[0068] At block 308, the method 300 includes estimating for each of the plurality of
SKUs, a total number of store-switchers from amongst the plurality of consumers based on the comparing. The number of store-switchers may be estimated based on a store-switching

condition. For example, for a consumer, if the new satisfaction score is less than the current satisfaction score, the consumer may switch the retail store. Alternatively, if the new satisfaction score exceeds or closely matches the current satisfaction score, the consumer may switch to the alternative SKU. In case when more than one alternative SKU is present, the consumer may switch to that alternative SKU which has highest satisfaction score. Further, if a consumer switches the retail store then he is referred to as store switchers. Therefore, based on comparing, the total number of store-switchers is estimation. In one implementation, the SKU delisting module 214 is configured to compare the current satisfaction score and the new satisfaction score to estimate the total number of store-switchers.
[0069] At block 310, the method 300 includes delisting one or more SKUs from
amongst the plurality of SKUs from a retail store based on estimating. The SKUs for which number of store-switchers is minimum may be delisted.
[0070] Although embodiments for the SKU rationalization have been described in
language specific to structural features and/or methods, it is to be understood that the invention is not necessarily limited to the specific features or methods described. Rather, the specific features and methods are disclosed as exemplary implementations for SKU rationalization.

I/We claim:
1. A computer-implemented method for stock keeping unit (SKU) rationalization,
wherein the method comprising:
determining for each of a plurality of consumers, a most preferred SKU and at least one alternative SKU from amongst a plurality of SKUs based on a sales data set;
comparing a current satisfaction score and a new satisfaction score for each of the plurality of consumers upon purchase of the most preferred SKU and the at least one alternative SKU respectively;
estimating for each of the plurality of SKUs, a total number of store-switchers from amongst the plurality of consumers based on the comparing; and
delisting at least one SKU from amongst the plurality of SKUs based on the estimating.
2. The method as claimed in claim 1, wherein the method further comprising retrieving panel data (110) for a predefined period from a database (108), and wherein the panel data (110) comprises data associated with at least one product category of one or more retail stores.
3. The method as claimed in claim 1, wherein the method further comprising generating the sales data set based on the panel data (110), and wherein the sales data set include sales data of each of the plurality of SKUs associated with the at least one product category.
4. The method as claimed in claim 1, wherein the method further comprising computing the current satisfaction score and the new satisfaction score for each of the plurality of consumers upon purchase of the most preferred SKU and the at least one alternative SKU respectively.
5. The method as claimed in claim 1, wherein the method further comprising identifying a minimum number of store-switchers for at least one SKU from amongst the plurality of SKUs.
6. The method as claimed in claim 1, wherein the determining comprises analyzing the sales data set to ascertain number of purchases corresponding to each of the plurality of SKUs.

7. The method as claimed in claim 5, wherein the identifying is based on a store-switching condition.
8. A stock keeping unit (SKU) rationalization system (102) for SKU rationalization, the SKU rationalization system (102) comprising:
a processor (202);
a SKU determination module (212) coupled to the processor (202), the SKU determination module (212) configured to determine for each of a plurality of consumers, a most preferred SKU and at least one alternative SKU from amongst the plurality of SKUs based on a sales data set;
a score computation module (112) coupled to the processor (202), the score computation module (112) configured to compute a current satisfaction score and a new satisfaction score for each of the plurality of consumers upon purchase of the most preferred SKU and the at least one alternative SKU respectively; and
a SKU delisting module (214) coupled to the processor (202), the SKU delisting module (214) configured to:
compare the current satisfaction score and the new satisfaction score for each of the plurality of consumers upon purchase of the most preferred SKU and the at least one alternative SKU respectively;
estimate for each of the plurality of SKUs, a total number of store-switchers from amongst the plurality of consumers based on the comparison; and
delist at least one SKU from amongst the plurality of SKUs based on the estimation.
9. The SKU rationalization system (102) as claimed in claim 8, wherein the SKU determination module (212) is further configured to retrieve panel data (110) for a predefined period from a database (108), and generate the sales data set based on the panel data (110).
10. The SKU rationalization system (102) as claimed in claim 8, wherein the total number of store-switchers is estimated based on a store-switching condition.

11. A computer-readable medium having embodied thereon, a computer program for executing a method comprising:
determining for each of a plurality of consumers, a most preferred SKU and at least one alternative SKU from amongst the plurality of SKUs based on the generated data set;
comparing a current satisfaction score and a new satisfaction score for each of the plurality of consumers upon purchase of the most preferred SKU and the at least one alternative SKU respectively;
estimating for each of the plurality of SKUs, a total number of store-switchers from amongst the plurality of consumers based on the comparing; and
delisting a SKU from amongst the plurality of SKUs based on the estimating