A Cluster Analysis Study of Small & Medium Enterprises Menna Sharma & Pawan Wadhawan Tita Borshalina 29009008.

A Cluster Analysis Study of Small & Medium

EnterprisesMenna Sharma & Pawan Wadhawan

Tita Borshalina

29009008

Purpose

The study attempts to cluster the successful Small & Medium Enterprises (SMEs) in India according to their growth mode & strategies

Utility

Identifying the conditions of success in the SME sector is very important for the acting & nascent entrepreneurs, organizations fostering SME development, financiers, public policy makers & other stakeholders of SMEs.

Using the results obtained, organizations fostering entrepreneAurship & SME

development can direct their actions & develop their products, education & advisory

services.

Applying Cluster Analysis to Categorize Enterprises

Stuart & Abetti (1990) Birley & Westhead (1994)

Applied factor analysis followed by cluster analysis to quantitatively

measure. Results of the analysis made by grouping the factors revealed that

entrepreneurial experience- the number of previous new venture involvements and the level of the management role played in such ventures was the most

significant factor. Other experience factors such as age, years of business, management and technical experience,

various dimensions of the entrepreneurial team’s experience, etc., were not

significantly related to performance.

Hierarchical cluster analysis has been applied to group the

attitudes that entrepreneurs hold towards their regional &

economic environment in Great Britain. 9 elements were valuated for evidence that perception of

the same region differs according to the metropolitan location, &

that location images are related to specific types of manufacturing

activity.

Applying Cluster Analysis to Categorize Enterprises (con’t)

Cunningham & Maloney (2001) McMahon (2001)

Segmented the Mexican microfirms by applying factor & cluster analysis. The analysis was based on 4 broad classes of variables: characteristics of entrepreneurs, characteristics of firms, entrepreneurial dynamics &

their participation in formal market & legal institutions. They were divided

into 6 clusters which provided information on a variety of firm & entrepreneur characteristics by

cluster.

Employed empirically-based development taxonomy for SMEs in

the manufacturing sector using panel data recently made available from Australia’s business longitudinal

survey. Exploratory cluster analysis was used with key enterprise age,

size, & growth variables to discover if there appeared to be any stable

development pathways evident in the data.

Applying Cluster Analysis to Categorize Enterprises (con’t)

Pasenen (2003) Ariyawardana & Bailey (2003)

Identify the factors affecting SMEs in Eastern Finland &

classify these firms in different clusters. An analysis of all the successful SMEs using cluster

analysis revealed that they costitute a heterogeneous group

with a large variety of characteristics.

Applied cluster analysis to determine the existence of strategic groups. The three hierarchical methods: average linkage,

centroid and Ward’s methods resulted in three different cluster solutions. However,

the Ward’s method outperformed both the average linkage and centroid

methods. Based on the dendrogram of the Ward’s method, the existence of

three different clusters was identified. These three clusters generated by the

cluster analysis were referred to as strategic groups.

Applying Cluster Analysis to Categorize Enterprises (con’t)

Leger et al. (2005)

applied cluster analysis for profiling the firms conducting electronic commerce at the international level. The paper illustrates how firms in a

highly technological industry may be categorized according to their levels of both international and electronic commerce. Based on the

different firm and market

characteristics, the profile of each cluster was established. Each group corresponded to a specific behavior pattern with respect to

internationalization and electronic commerce usage.

Research Questions & Objectives

Research Questions1. How can successful SMEs be characterized?2. How can successful SMEs be clustered?

Research Objectives1. To identify the common characteristics

between different SMEs.2. To derive the taxonomy of successful SMEs.

Research MethodologyScope of Study & Sample Selection

The study is focused on the established SMEs (i.e., more than four years old) with roughly 5 to 249 employees located in the Ambala district of

Haryana (India).This study attempts to find the common characteristics between different types of SMEs. Ambala is known for scientific and

surgical equipments. There are others firms also which are engaged in manufacturing kitchen mixer grinder, submersible motor pump and metal

casting. There are currently more than 250 SMEs in Ambala scientific instrument cluster. Manufacturers have their association namely Ambala Scientific Instruments Manufacturers Association (ASIMA) consisting of 191 members. Random sampling was done and every second member

was interviewed with the help of a structured questionnaire. The study was conducted during June 2006 to March 2007. Primary data was collected with the help of questionnaires. Information to be collected consisted of

information related to the characteristics of the firm, characteristics of the entrepreneur and entrepreneurial dynamics. The questionnaire comprised

of questions regarding age of the firm, number of founders, number of personnel, type of business, number of establishments etc. To obtain the

profile of the entrepreneurs, questions were asked about gender, age, educational level, etc.

Research Methodology (con’t)

Research Methodology (con’t)In this paper, using Ward’s Method to measure distance between 2 clusters

Research Methodology (con’t)

Using an agglomerative hierarchial method

Results & Discussions

Results

Results (con’t)

Verifying the Accuracy of the Clustering Process

Results (con’t)

In order to ensure that the clustering is appropriate, a canonical discriminant analysis was performed on the three clusters and 11 variables. Two canonical discriminant functions were significant in differing among the clusters (p < 0.0005).

The analysis revealed that the discriminant functions had eigenvalues of 4.930 and 1.183 with canonical correlations of 0.912 and 0.736. Such eigenvalues can be regarded as good, and such correlations indicate high efficiency of discriminant functions in discrimination. The map of discriminant functions was clearly divided into sectors, which indicates that both functions add new information to the classification (Table 5).

Results (con’t)

Results (con’t)

Wilks’ lambda value for both functions was 0.077, and 0.458 for the second function alone (p < 0.0005). Thus, the discriminant model explained 97.8% of the total variance between the clusters (p < 0.0005). Standardized discriminant coefficients indicate the predictive power of single variables of the model.

Variables with over 0.3 values of standardized canonical discriminant function coefficients had the highest predictive power: growth in turnover and its nature, goals for growth and firm size. Examining the correlations of the discriminant function and original variables, the most important factors for discrimination in addition to those mentioned above were the growth in demand in the main markets, local market’s share in the firm’s sales, and the uniqueness of the products in the market.

Another indicator of the applicability of the discriminant model is the degree of predictive accuracy measured by the percentage of cases classified correctly. Overall, 97.8% of the cases were correctly classified, considerably greater than could be achieved by chance alone.

Results (con’t)

Results (con’t)

It is clear from the Table 6 that the most important factor in discriminating the three clusters was respondent’s satisfaction with the firm’s performance with value of 0.510, followed by market demand change. The factors which contributed least towards discriminating the clusters were goals and objectives of the firm with a discriminating power of 0.164.

Discussions

Cluster 1: Growth-Oriented & Network Intensive SMEs

The total number of SMEs in this cluster are 66. Had medium number of employees in the range of 20-40. They were growth oriented, transition in the ownership to the next

generation was being or about to be made in these organizations. Their present business is different from their original business. Necessary modifications were made in management principles &

practices to adapt themselves on par with the changing environment. The growth rate was not that steep and same was the trend with market

demand of their products. They were satisfied with their performance and their goal was to earn

basic livelihood for the owner. Performance of the members of these clusters was not good compared

to the most important competitors and the market demand was not that encouraging for their growth.

Discussions (con’t)Cluster 2: Innovators with Continuous Growth

Out of a total of 135 SMEs only 14 were clustered in this group. The number of employees in this cluster were less than 20, it means

the size of the SME is not that big in this cluster. These firms were growth oriented and were always in search for new

business opportunities. In most of the firms, transition had already taken place, the present

business was totally different from their earlier business. To capitalize on the opportunities available to them, the firms shifted to new business. The new generation in these SMEs is more proactive. Their management principles and practices were also changed considerably. These firms can also be called as adaptive because of the change in their line of business and management principles and practices to suit the changing environment. As a result of these practices, their turnover increased in spite of not much increase in the demand for their products. They place themselves at a slightly better position than their competitors but are not satisfied with what they have achieved because they intend to grow at much higher rate.

Discussions (con’t)

Cluster 3: Independent Survivors

The total number of SMEs in this cluster are 55 and the average number of employees is more than 20 per SME. It consists of SMEs which are comparatively bigger in size. But these firms are not growth oriented. Their only aim is to provide reasonable livelihood to the entrepreneur. In most of the firms in this cluster, transition in the ownership to the next generation is being or about to be made. Their present business is somewhat similar to their earlier business. This indicates that they have not changed as per the changing business environment. Also their management principles and practices have not changed.

The turnover has not changed much and is almost constant in comparison to other clusters. Their market demand grew to some extent. They place themselves at a slightly better position than most competitors of their business (which they have not changed).

Moreover, they are quiet satisfied with what they have achieved. Respondents are not much educated in this group. They are typically between SSC and graduate level.

Conclusion The results of this research can help SME policy makers and organizations fostering

entrepreneurship and SME development. First of all, in planning public interventions and other actions, the differences between the types of SMEs should be taken into account. This means that an action which may be appropriate for SMEs in one cluster may not be appropriate for SMEs in another cluster. In other words, the actions should be modified according to the needs of the target firms.

A high proportion of successful SMEs are family firms. Consideration should be given to the ways in which the development potential bound

up with family firms could be utilized, and how family firms could be encouraged to grow.

Moreover, in a significant number of family firms, the transfer of business from one generation to another has not yet happened.

Particularly in these firms, the role of founders ay be crucial for firm development. The time of succession might be a natural moment to check the firm’s course of action.

Also, more attention should be paid to fostering the success of successions. On the basis of the age distribution of the entrepreneurs, it can be concluded that in many successful SMEs the management of the firm will change during the next 10 years, and this may have an important impact on the future development of the region, so this issue should be given serious thought by the local and regional organizations supporting SME development.

Conclusion (con’t)

The study focuses on successful independent SMEs only in one particular region of the country.

A separate study covering broader geographical area with a larger sample can be carried out.

However, studying SMEs within homogenous clusters seems to be a promising way to proceed and the present study constitutes a starting point for the construction of appropriate models.