IS CATEGORY MANAGEMENT IN SMALL SUPERMARKETS WORTH THE EFFORT?

Category management (CM) is an important tool to strengthen the relationship between manufacturers and retailers. This process has been associated with large corporate retailers; however, some recent researches show that CM is open to companies of any type or size. This possibility is important in emerging markets, where neighborhood supermarkets are still representative and are often considered an alternative for manufacturers to achieve higher margins compared to big chains. In this context, the aim of this research was to analyze the results of a CM initiative in small neighborhood supermarkets from a manufacturer perspective. Data for the study comes from a food manufacturer in Brazil that implemented a CM process with 180 small retailers. A quantitative analysis was conducted in order to analyze the effect of the program on the food manufacturer’s sales and market share. Our analysis suggests an overall positive effect of the program on both, sales and market share.


RESUMEN
La Gestión de categorías (GC) es una herramienta importante para fortalecer la relación entre fabricantes y minoristas. Este proceso ha sido asociado a grandes minoristas corporativos; a pesar de eso, algunos datos recientes muestran que el GC está abierto a empresas de cualquier tipo o tamaño. Esta posibilidad es importante para mercados emergentes, donde los mercados de barrio todavía son representativos, siendo frecuentemente considerados como una alternativa, para que los fabricantes alcancen márgenes más altos que en aquellas alcanzadas en las redes de supermercados mayores. En ese contexto, el objetivo de esta investigación fue la de analizar los resultados de la iniciativa del GC, en minimercados de barrio, desde una perspectiva manufacturera. Los datos usados para la elaboración del estudio vienen de un fabricante de alimentos en Brasil, que implantó un proceso de GC junto a 180 pequeños comerciantes. Se utilizó un análisis cuantitativo, con el fin de analizar el efecto del programa en las ventas y en la participación de mercado, por parte del fabricante de alimentos. En nuestro análisis se sugiere un efecto final positivo del programa, tanto en términos de ventas cuanto sobre la participación de mercado. PALABRAS-CLAVE | Gestión de categorías, mercado, canales de distribución, minoristas de alimentos, mercados emergentes. by focusing on delivering consumer value. Even though CM has become popular, especially associated with large companies in developed markets, Arkader and Ferreira (2004) conclude that its practices can be suited to companies of any type and size.
This finding is important when observing the movement of global firms since they are also interested in developing business in new markets, such as emerging markets; specifically, they want to discover effective strategies and marketing practices in emerging markets (SHETH, 2011).
In terms of retailing, as neighborhood supermarkets are more prevalent in emerging markets than developed markets (CHATTERJEE and others, 2011), the challenge faced by consumer packaged goods (CPG) companies will be to identify strategies that are appropriate for this channel. Under this context, CM might be an alternative to be investigated for this channel.
In this sense, we question whether a CM initiative in an emerging market is worth the effort from a large supplier perspective, once this company has employed some of the CM tools in order to increase sales by establishing a better relationship with family-owned neighborhood supermarkets in Brazil.
This investigation is important for three main reasons.
First, researches on CM often analyze results only from the perspective of large corporate retailers, but neglect the manufacturer's role in this process (ARKADER and FERREIRA, 2004, p. 44) or its implementation in another channel type.
The second reason is because independent stores will tend to play an important role over coming years in emerging markets (CHATTERJEE and others, 2011;KERTESZ and others, 2011;DÍAZ, LACAYO, SALCEDO, 2007). However, it is important to note that independent retailers that own small supermarkets in emerging markets, such as Brazil, are under increasing pressure. Ailawadi, Borin and Farris (1995) found that some large retailers are gaining power over other retailers. As global retailers are also growing their business in emerging markets, independent supermarkets are struggling. This can be seen as an oppor-tunity for CPGs to provide them with some support, such as CM practices, under this competitive environment.
Finally, CPGs can achieve higher margins selling to independent stores in these markets (DÍAZ, LACAYO, SALCEDO, 2007), and also due to the fact that small retailers have a con- In this research, a discussion is proposed in order to bring category management to the attention of researchers and practitioners, considering the analysis of sales and market share outcomes from a field experiment after the implementation of a CM program by a global food manufacturer along with 180 small supermarkets in Brazil.
In the article, the next section outlines the conceptual background based on relevant literature related to the proposal of this paper, followed by a review of the aspects regarding channel environment in emerging markets, methods, procedures, results and conclusions.

CONCEPTUAL BACKGROUND AND HYPOTHESES
The literature review for this paper is based on the following topics: category management, shopper marketing and sales force execution.

Category management
Companies are trying to adapt their demand chain management in order to protect their sales and market share; thus, adapting marketing and supply chain competencies can be a source of competitive advantage for suppliers and retailers to succeed in a highly competitive market by establishing positive channel relationships (JUTTNER, CHRISTOPHER, BAKER, 2007).
Aspects of channel relationships have received considerable amount of attention from researchers (ETGAR, 1978;FRA-ZIER and SUMMERS, 1984;ANDERSON and NARUS, 1990;AI-LAWAD, BORIN, FARRIS, 1995;GANESAN and others, 2009;RUNYAN, STERNQUIST, CHUNG, 2010). Specifically, there are a lot of types of actions that have a negative repercussion and can threaten the supplier-retailer relationship (HIBBARD, KUMAR, STERN, 2001). In order to avoid this type of situation and a negative repercussion in the supplier-retailer relationship (e.g., activation of point of sales materials and retail alignment), category management can contribute by aligning the interests of both parties in order to stimulate sales and a better point-of-sale shopping experience for consumers (DUSSART, 1998;GOONER, MORGAN, PERREAULT JR, 2011).
According to Gooner, Morgan and Perreault Jr. (2011, p. 18) "in increasingly competitive markets, managers and scholars must rethink traditional approaches to creating and capturing value, including how inter-organizational relationships are handled". For CPGs, category management is often considered a breakthrough in trade practices (DUSSART, 1998) and offers conditions to strengthen the relationship between a food supplier and its marketing channels, bringing positive elements to supplier-retailer interaction (LINDBLOM and OLKKONEN, 2008). Dussart (1998) analyzes that a shift from brand management to a category management approach is of importance because consumer decisions for one specific brand (e.g., coca-cola) can have an influence in the role category offered by one store (e.g, beverages). As a category management process offers conditions to manage product categories as business units, the interests from food manufacturers, retailers and shoppers can be combined and integrated in order to deliver value for both of them.
Moreover, CM tactics also deal with assortment, space allocation, and in-store promotional activity (LINDBLOM and OLKKONEN, 2008). As these elements have a large influence on sales and market share, a CPG should implement this process to improve its competitiveness. Moreover, this practice has been largely associated with the relationship between suppliers and large retailers, improving in-store marketing. Arkader and Ferreira (2004) have investigated the importance of a category management initiative in Brazil, from the retailer perspective. The authors conclude that this practice is of importance in the context of some changes that took place in the 90s, and "there is an indication from this study that the cat-egory management process is open to companies of any type or size" (ARKADER and FERREIRA, 2004, p. 50).
Considering the findings of Arkader and Ferreira (2004), this research is important because contrary to traditional research on category management, our data comes from a CM process implemented by a food CPG company with small supermarkets that do not belong to big chains, which are still important in this kind of market; that is, we propose to analyze the CM process from a different application, because its practices can be suited to companies of any type or size.
On the other hand, in spite of its importance, there is still Finally, from previous researchers it is possible to conclude that the category management process plays an important role in this context, because it can strengthen the supplier-retail relationship in order to generate more sales by also improving instore marketing. As a consequence, it is also important to consider the literature on shopper marketing and in-store marketing.

Shopper marketing, in-store marketing and sales force execution
A traditional belief states that more than half of the purchases are unplanned (KOLLAT and WILLETT, 1967;UNDERHILL, 2000) and researchers have agreed that unplanned purchases are more representative due to consumer susceptibility to in-store stimuli and due to some situational factors such as time available for shopping (PARK, IYER, SMITH, 1989;STILLEY, INMAN, WAKEFIELD, 2010). Beatty and Ferrell (1998)  shoppers. Finally, the last stage is product experience, when consuming the product after the purchase.
Therefore, it is important to address the concept of shopper marketing. In the Journal of Marketing, this concept has started to be considered (BELL, CORSTEN, KNOX, 2011) and in the Journal of Retailing this concept has been reviewed, confirming its importance. According to Shankar and others (2011, p. 29), shopper marketing "refers to the planning and execution of all marketing activities that influence a shopper along, and beyond, the entire path-to-purchase". Considering the importance of shopper marketing, the first step of the analyzed category management process with small retailers in the experimental group was to understand the shopper, and then moving forward to the next steps as it is later mentioned in the method section.
Additionally, concerning the execution of marketing activities, the analyzed experiment was executed through the supplier's sales force. In this way, sales force is strongly responsible for the execution of business strategy (CHURCHILL and others, 2000). In a CPG, they can be responsible for implementing CM program among their tasks. Also, sales force plays an important role by establishing closer relationships with retailers.
In the situation we have analyzed, the company's sales force has provided store owners of the experimental group with support to organize its category's assortment, activate point of sale materials, and shelf-space in order to increase sales.

Hypotheses
In accordance with the theoretical review, we expect to validate our proposed hypotheses from the experiment described in this paper. A group of factors motivated us to propose the following hypotheses.
First, some evidences show that category management is open to companies of any type or size (ARKADER and FERREIRA, 2004) and sales lift is expected from the channel relationship established by means of a CM project (DUSSART, 1998;LINDBLOM and OLKKONEN, 2008). Therefore, CM can be an alternative to be applied to a different type of retailer, in this case, small neighborhood supermarkets. But previous studies have not yet tested this specific possibility. Also, sales force is important in the context of a CPG company for implementing business programs with retailers (QUELCH and CANNON-BONVENTRE, 1983). In this context, one manufacturer could use its sales force to implement a CM project in order to bring positive results. Thus: H1: The implementation of a category management initiative from the supplier's sales force has a positive effect on the manufacturer's sales for the specific category of the project.
Second, in-store marketing and shopper marketing practices offer conditions for a supplier to overcome competitors in the same store environment (BELL, CORSTEN, KNOX, 2011;SHANKAR and others, 2011). Taking into account that in-store marketing has a positive effect on sales (CHANDON and others 2009), and considering previous literature that also de-In the analyzed experiment, the manufacturer has also manipulated a second category in which they are not the market leader. Thus, its main competitor in this specific category is the market leader. Knowing that higher-market-share brands have more awareness and preference, having less impact from in-store movements (CHANDON and others 2009), we expect that the market share growth for a manipulated category where the manufacturer is not the market leader will be higher compared to the secondary brands instead of to the main competitor, because secondary brands generally have less awareness and preference than the main competitor. Thus: H3c: Market share growth from the implementation of a category management initiative for a category in which the sponsored manufacturer is not the market leader is higher compared to secondary brands.

CHANNEL ENVIRONMENT IN EMERGING MARKETS
According to a contribution to the study of emerging markets, Sheth (2011)  and local competition in these countries is stronger that it appears. However, channels' features for consumer products are different in emerging markets. As a consequence, it is important to address some specific features of grocery channels in these markets.
In developed markets, the earlier expansion of the supercenter format has boosted significant changes, bringing the so-called "one-stop-shopping" approach, when a retailer offers food items, non-food items and ancillary services. However, in emerging markets the entry of global retailers has happened af- However, it is important to note that some specific contrasts must be considered for CPGs when comparing their approach to developed and developing markets, which also helps to justify the category management employed by a food company in one emerging market.  from classes D/E. Thus, the importance of this channel in this emerging market also relies on the growing middle-class population because this channel has a high penetration in this class.
The same authors conclude that having different actions for the different channel formats is increasingly important in Brazil, such as for neighborhood stores (KERTESZ and others, 2011). Gupta and Steenburgh (2008) conclude that in marketing literature, experiments are a powerful method for understanding the market response to new marketing activities that the firm has not tried historically. In this paper, we analyze the category management process, which was applied as a new marketing activity developed by one manufacturer and a sample of small retailers, allowing us to set up test-and-control samples to assess the effectiveness of the referred category management program.

METHODOLOGY AND PROCEDURES
Thus, this research uses a quantitative approach after a field experiment has been conducted by a large CPG, in which 180 small family-owned supermarkets participated in the company's category management program, called the experimental group in this paper.. The other group of similar store features served by the same company that were not included in the company's category management program is called the control group.
For research development, the following stages were taken: i) we first conducted a review of academic literature on category management, in-store marketing and sales force execution aimed at extracting potential ideas in order to tailor a frame-work of category management that could be applied to small independent supermarkets, considering their features in Brazil; ii) training the company's sales force team to execute the program in the stores they had selected, according to concepts and practices revised during the previous stage; iii) pre-testing the field, called the pilot project in this paper, in order to adjust and validate the CM process according to a smaller-store environment; iv) implementation of the company's program with 180 small family-owned supermarkets through the sales force team; v) separation between the experimental group and the control group, according to the company's customer database for the analyzed type of store, in which the experimental group and the control group had similar features as the other group (e.g. store size, shopper profile, geographic location); vi) sales and market-share data collection related to the experimental and control group from the company's internal system; vii) analysis of the sales and market-share of the company in the experimental and control group stores; viii) finally, a quantitative analysis was conducted in order to analyze available data and to assess the effect of the program on the food manufacturer's sales and market share, as proposed in this research.

Data and framework
The data available for this research includes sales and market share variables from the analyzed CPG company in Brazil, according to monthly data ranging from July 2008 to June 2010. The experiment was implemented in the southeast of Brazil. Some steps were considered by the CPG's sales force in order to develop a category management review in each experiment store, including store layout, assortment, product shelf position, product's in-store visibility, promotion and price. In this sense, the sales force followed steps from the CM framework for small supermarkets proposed by Guissoni, Consoli and Castro (2010). According to the related framework implemented in this experiment after the shopper and retailer understanding, the sales force team from the analyzed CPG should suggest to each experiment store: (i) the categories to be offered; (ii) the store layout to place the categories according to their role; (iii) the assortment, that is, the most appropriate SKUs to be offered considering previous steps; (iv) the in-store materials to be used to promote the categories in the project stores.
In the implementation, due to a decision from the CPG company, this initiative was implemented according to different phases, including a different set of stores for each phase. All sample stores were small independent supermarkets with 1 to 4 checkout lanes. Table 2 presents the main information for implementing the initiative.
Data analysis was also broken down by product category where Category 1 refers to non-alcoholic beverages, and Category 2 to alcoholic ones. Both categories were manipulated in the field experiment. Additionally, we evaluated results considering the focal company's data (Company), the main competitor (Competitor 1) and secondary brandcompetitors (Competitor 2). It is important to note that for category 1, the focal company is the market leader. However, for category 2 the company is not the market leader, since competitor 1 is the market leader. In order to have a comparison base, all data was transformed into a 100 base sequence after the first month considered in the period range analysis.
We analyzed the data from the project stores (i.e., experimental group) consisting of 180 stores and the control group, including 2,500 stores with similar features as the experimental group as described in this paper, which also have 1-4 checkouts and are served by the company in the same regions as the other group. Following the period of analysis, available data from each store shared by the company for this paper were: i) company sales volume for experimental and control groups for category 1 and company market share for both, categories 1 and 2.

Procedures
Since this paper questions whether a CM initiative with small retailers is worth the effort, it is important to analyze sales and market share effects as a result of this experiment. To make this possible and considering the available data, we performed a variance analysis (ANOVA) in three different applications. In the first one, sales differences in the project stores considered pre-and post-implementation data. Secondly, we focused on comparing the means between project stores and the control group's sales. In order to check the control group, we have also compared sales differences in the control group stores considering pre-and post-implementation. Finally, specifically for market-share, another set of variance analysis was also performed comparing company market share evolution to competing companies in the same period.

RESULTS
Before analyzing the effect of category management programs on sales and market-share, a Kolmogorov-Smirnov test, which is a normality test, was performed for each sample regarding sales and market-share variables. This test is important because in order to conduct an analysis of variance (i.e., ANOVA), the distribution of the values must be close to a normal distribution.
The Kolmogorov-Smirnov tests of the sales and market share variables for both the experimental and control groups allowed us to observe that the distribution is normal since the p-value of samples is greater the 0.05. Thus, the analysis of variance can be used as a way to compare the variables from the groups, validating the procedures presented and applied to this research.
Following our analysis, Table 3 illustrates that groups 1 and 4 presented better sales performance for non-alcoholic beverages. For group 1, in terms of sales volume, project store Is category management in small supermarkets worth the effort? sales increased 39.37% compared with pre-project data, and 5.36% compared with control group; although, control group performance was 6.14% higher post x pre-project. As only 8 stores are in this group, we still considered the result as positive. It is important to note that group 1 has the longest data set and having its period analysis starting and finishing in the winter, no inverse relation due to seasonality was identified (as it is a beverage category). The Group 2 data set on the other hand, started late summer and finished in the winter. Thus, company sales are expected to decrease in the winter; so, better results in this case are those with lower sales reductions. We can see that control group sales decreased in the period 11.95% while project stores fell 13.67%, contrary to what we expected, possibly due to inventory, seasonality and the implementation process. The results for group 3 present inverse relations as data sets started late winter, have a full summer and finished in the winter again. We also verified that total sales pre versus post-project increased more in the control group (9.95%) than in the project stores (6.07%). Finally, group 4 results presented positive effects on sales, where project store sales increased 31.80% compared with pre-project data and the biggest difference between project stores and control group was identified (23.46%).
Following the ANOVA test, specifically for group 1, we considered there was a difference in terms of sales performance for project stores, pre x post experiment, at 10% of significance level (p-value = 0.0674). In the other cases, sales differences were significant at 5%, as shown in Table 4. Moreover, significant differences happened in group 2 between project stores versus control group (p-value = 0.0006) as well as control group pre and post-project (p-value = 0.0351). Similarly, group 4 presented a significant difference in sales performance pre and post-project (p-value = 0.02912) as well project stores versus control group (p-value = 0.03568).

TABLE 3. Sales performance for non-alcoholic beverages for the category management program
Sales performance x baseline (Initial period = 100) Group 1 Group 2 Group 3 Group 4 Pre-project x post-project (project stores) 39.37% -13.67% 6.07% 31.80% Project stores (post) x control group (post) 5.36% 0.49% -2.50% 23.46% Control group (pre-project x post-project) 6.14% -11.95% 9.95% 11.02% Based on the results, a positive overall effect of the project on manufacturer category sales can be found, excepting for groups 2 and 3. Therefore, we have partially confirmed H 1 as explained in the sequence of this article, because in order to consistently validate and explain results from tables 3 and 4, after performing this analysis we interviewed six company sales supervisors responsible for implementing the project for all groups. The motivation was to better understand sales results for groups 2 and 3, once the analysis of variance suggested that some of the sales differences were not significant.
The interviews allowed us to understand any evidence from the implementation that might have interfered in the sales results for the experimental group. According to the sales super-visors, when they tried to scale up the project (groups 2 and 3), they did not perform the first steps from the applied framework (i.e., shopper and retailer understanding) well while planning the category management program for stores in groups 2 and 3, once the number of stores had increased after the pilot project.
However, as a learning process to implement CM under a different retail environment, they were able to improve the process in order to scale up the project. This could be seen in the results of group 4, when they achieved better results.
In terms of market share growth, the analysis also shows an overall positive effect across the groups. However, as we did not have market share historical data for group 1 (project pilot), its evolution could not be evaluated. Interest-ingly, groups 2 and 4 presented the best performance in market share analysis. All variance analyses presented in Table 5 were significant at the 5% level. In this way we found results to be positive, which corroborates hypothesis H 2 . As previously stated in the procedures, absolute market share at the beginning of the project was used as 100 base and further analysis considers the relative increases over this base. Taking group 2 to illustrate the analysis presented in However, for all groups the market share increase was high-er in the category in which the manufacturer is not the market leader, allowing us to confirm H 3b . Also, some similarities were found across all groups with regards to market share analysis. For example, overall performance of category 2 was quite positive, mainly against secondary brand competitors. That is, despite the growth of market share against both the main competitor and secondary brands within the category, this project contributed for the manufacturer to further increase its market share compared to competing secondary brands in the category in which they are not market leader, allowing us to confirm H 3c .
Based on the previous analysis, we confirmed both hypotheses, since evidences of an overall positive effect on both sales and market share were found across the groups. Results show that the implementation of category management as a service offered by the manufacturer's sales force to small neighborhood supermarkets besides increasing category sales, seems to have a positive effect for a manufacturer to overcome competitors in terms of market share.
Is category management in small supermarkets worth the effort?

CONCLUSIONS AND MANAGERIAL IMPLICATIONS
Considering that "suppliers must move away from purely transactional relationships with their retail customers and more to- From the results, we can look at the performed experiment as category management giving firms the ability to improve on basic positive trends, but not able to stop negative ones due to several factors (e.g., seasonality). Also, it is interesting to note a kind of side effect of category management for manufacturers that carry both leader and non-leader brands. In this case, H 3b was confirmed. In category 2, the analyzed manufacturer does not have the same brand power as in category 1, as they are not market leader. However, the growth of market-share for this non-leader category was higher in the experiment than for the category in which they are market leader. In this case, for this experiment to happen, the manufacturer took advantage of its brand power in category 1, as a market leader, in order to persuade small retailers to join the CM program in which their two categories would be manipulated. This strategy allowed the manufacturer to further stimulate better management of its brands in category 2 of alcoholic beverages in the stores' projects, in accordance with presented results.
In a nutshell, our findings suggest that a manufacturer can use its influence in one category with leader brands to improve results for its other category with non-leader brands by managing both categories and following a CM process in a small neighborhood store environment. Besides, market share increase was consistently higher against secondary brands than compared to the main competitor in the category, which corroborates H 3c . The same result was verified for both manipulated categories.
In this research, we propose to contribute to the literature by showing the worth of a CM initiative from a manufacturer perspective in a representative channel that is, small neighborhood supermarkets, in emerging markets. Theses perspectives can contribute to the literature, since most previous researches have investigated their effects from a retailer implementation perspective, mostly taking place in developed markets and in big supermarket channels.
Considering however the CM as a process that can contribute to the supplier-retailer relationship and benefit both parties, if retailers in this kind of project are not benefited, they tend to not accept CM programs proposed by a specific manufacturer. In this way, future research should investigate the joint effect of a CM program from both parties, that is, manufacturers and retailers.
In this sense, examples of research questions might include: What are the effects of a specific category management initiative for both, manufacturers and retailers? Does CM improve total category sales in the stores' project or only the share of supplier's sales? Is CM an alternative to improve profits for one specific category for retailers? Are manufacturers' decisions of assortment and in-store marketing biased to protect only its interest during a CM program? Also, by comparing the results of each store project would it be possible to investigate what the main drivers for developing an effective CM program are? Answers to these questions have the potential to contribute to academics and practitioners. Some limitations of our research can be related to some occasional events that affected the market in which we applied our research, as well limitations of data and procedures used to implement the project, as it was the company's sales force's responsibility. Also, one of the limitations is the lack of more complete data coming from retailers.
Considering the importance of a channel push activity under the context of this study, and the importance of emerging markets, this paper shows that providing services to independent retailers, such as category management support in order to help them organize their stores in terms of having an appropriate assortment, layout and in-store marketing, can create overall positive effects on manufacturer sales and market share. Therefore, allocating efforts for a category management initiative with small supermarkets can be an interesting strategy in an emerging market environment, such as Brazil.