Statistical Approaches in Accounting for Optimal Business Choice and Location Decisions

Introduction

The process of selecting a business type and optimal location is a complex decision that requires thorough consideration of various factors. To Perez-Benitez et al. (2021), decision on the physical place where to locate an entity is a strategic one that must be made not only at the inception of business activities but also as the business is expanding. In addition, choosing the type of business to establish is a key and crucial strategic decision, as it lays the foundation for the long-term survival and financial stability of any company. Factors which are considered in choosing the type and location of a company include population density, energy costs, labour cost and quality, local economic conditions, security in the environment, raw material supplies, agglomeration economies, as well as access to technology and digital infrastructure (Perez-Benitez et al., 2021; Arauzo-Carod et al., 2010; Che’ Man & Timmerman, 2016; Button, 2019)

The location of a business is a critical factor influencing a company’s overall strategy and success. The choice of business type and location can affect various aspects of a business, including customer reach, operational efficiency, cost structure, and competitive advantage (Thum & Wolf, 2018; Piacentini, 2016; Sweeney, 2015; Van Dijk & Leeuwen, 2014). In this regard, the decision to locate a business in an urban or rural area or to deal in consumer goods, industrial goods, foodstuff materials, or run a supermarket outlet and how this choice would impact the turnover of the business becomes critical to the strategic positioning, growth, and survival of the organization. For this reason, among others, Perez-Benitez et al. (2021, p. 1) submitted that business choice and location decisions have become “complicated, turning them into multi-criteria decision problems.”

Drawing from neoclassical theory, location theory, value-maximizing firm theory, discrete choice models, and Radom Utility Models (RUM), this study argues that in order to position any company in its industry so as to enable it gain a competitive advantage, there is a need to apply adequate statistical and accounting tools in the optimal choice of its location and type of activities it engages in. Emphasizing the importance of a holistic approach, this study seeks to provide a comprehensive understanding of the viability and profitability of potential business choices and locations.

There has been limited research on using statistical approaches as an accounting aid to guide business type and location choices in sub-Saharan African countries, such as Nigeria. This gap inadvertently contributes to business survival volatility (Kim et al., 2025; Ukamaka, 2021; Enyi, 2018; Sefiani et al., 2016). This study aims to bridge this gap by exploring the significance of integrating statistical and accounting considerations into business type and location decisions, and highlighting the implications for strategic planning and business performance. In doing this, the study has the main objective of exploring how statistical methods can enhance accounting practices in determining the choice of type of business and the ideal business location that can guarantee operational stability and maximum financial performance. The sub-objectives are to:

Following this introduction is a literature review (section 2). Section 3 presents the methodology; Section 4 gives the analysis, results, and discussions, while Section 5 concludes.

Literature review

The choice of business type usually occupies the greater aspects of the concerns of budding entrepreneurs because it builds on the experience, knowledge, and passion the business proposer attaches to the proposed business. Closely aligned to this is the choice of location that is critical in attracting patronage and eventual profitability. These decisions are important for customer reach, market access, cost savings, competitive advantage, and strategic flexibility.

Methods

This study adopted an empirical survey research design utilizing a structured questionnaire validated using Cronbach’s alpha reliability test, a public domain research instrument for testing the reliability of questionnaire in a pilot study of a sub-sample. The pilot study for this research was conducted using 21 respondents and the results for the five turnover sub-groups ranged from 0.8935 to 0.9264 as presented in Table 1.

The Cronbach alpha statistics requires no permission or license for deployment. In addition to the use of questionnaire, telephone interviews were scheduled for physically unreachable respondents through phone calls with the consent of the respondents obtained through the same. A total of 1,275 copies of the questionnaire were distributed, and only 1200 were returned valid and usable. The questionnaire administered clearly left out personal details of respondents. Consents of all respondents (both telephone and questionnaire) were obtained verbally because most respondents averred that they had little or no time to give written consent; and since much of the questionnaire were distributed on a person-to-person and telephone calls made on same basis which gave prospective respondents the opportunity to decline or accept to answer the questions posed the verbal consents were considered appropriate. No form of physical contact, exercise or taking of samples was required from any of our respondents. Other respondents contacted by phone were only asked if they could be reached through email or WhatsApp, which majority agreed to.

The responses from the questionnaire were analysed, codified, and arranged according to the factors and levels of turnover to be tested for influence. The predictor options were identified as urban and rural locations, supermarket outfits (supmkts), industrial goods (indgoods), consumer goods (consgood), and foodstuff materials, while five turnover ranges of below 2 million (Below2m), 2 to 5 million (B2T5m), 5 to 50 million (B5T50m), 50 to 100 million (B50T100m), and above 100 million (Above100m) were used as the options for the outcome or dependent variable “Turnover.”

The data were analysed using ValuStats (VSP 2.0) Ordinal Logistics Regression Analysis software.

Analysis, results and discussions

Results and discussions

The results of the ordinal logistic regression analysis for the data collected from the 1,200 respondents during the survey period are presented in Tables 2 and 3.

The MN Logit Regression Results for the dependent variable “Turnover” are interpreted as follows:

Discussion

The choice of the type of business to be considered and the location of such business depend heavily on factors that guarantee operational success and business sustainability. These factors include proximity to the source of supply, availability of cheap labor, regulatory environment, government policies, customer access, low operating costs, and a friendly tax regime (Kim, et al., 2025; Ukamaka, 2021; Sefiani et al., 2016). The choice of business type must be based on the competence, experience, funding capacity, and passion of business promoters. The choice of an incorrect business or location is a clear invitation to business failure. The empirical literature reviewed in this study affirms this assertion. Other empirical evidence unearthed within this study points to the fact that employing accounting metrics and statistical analysis as support in aiding business choice and location decisions will greatly improve organizational effectiveness and profitability (Horngren et al., 2015; Shim & Siegel, 2015; Hwang & Jang, 2012; Ying & Chen, 2019).

The general model information shows that the data were collected from 1,200 observations, providing a residual degree of freedom of 1172 and a model degree of freedom of 24. It also indicated that the maximum likelihood estimation (MLE) was used to estimate the parameters of the model. Log-Likelihood of -1414.4 indicates that the model was not perfectly fitted, and this was also indicated by the Pseudo R-squared of 0.2268, which means that approximately 22.68% of the variance in the dependent variable (Turnover) is explained by the model. Nevertheless, a Log-Likelihood Ratio (LLR) p-value of 0.0000 indicates that the model is highly significant, which compensates for the low LLR value. This means that the fitted model is a much better fit than the null model value of -1829.2; it also means that the predictors in the model significantly contribute to explaining the variance in the dependent variable.

Gleaning from the outcome of the ordinal logistic regression analysis on the research data collected, it is evident that factors such as urban, rural, supermarkets, consumer goods, industrial goods, and foodstuff exerted their own levels of influence on the turnover range. Table 5 shows how.

Given a 95% confidence level, urban as a predictor variable was significant for generating between the 2 million and 5 million turnover range and was marginally significant for generating between the 50 million and 100 million range. The analysis assumed that the baseline range of 100 million and above was used as the major comparator and that the urban variable was more related to that range than any other predictor. This is revealed by the descriptive statistics of the 246 occurrences of the “Above 100 million” turnover category in Table 4, which shows that the urban predictor has the highest mean of 8.98 compared to other predictors with less than 2. In addition, the maximum value of 45.66 for the urban predictor, which is 38.05 points higher than the nearest predictor (supermarkets) at 7.61 points, clearly proved that the urban predictor was mostly responsible for turning out the “Above 100 million” category turnover. The reason for this is not far-fetched because businesses with sales outlets located in urban areas can easily generate a turnover of up to 100 million naira and above annually.

Surprisingly, the rural predictor was only significant in the 5–50 million turnover categories. This could be attributed to the misconceptions held by respondents in rural areas who were largely uninformed about the reasons behind the questions posed to them. The results for the supermarket predictor are highly significant across all turnover categories, indicating that supermarkets remain a vital sales outlet that should never be overlooked as a business choice. The consumer goods results were highly significant for four of the five turnover categories, including the “Above 100 million” category. The reason for the insignificance in the 50 million to 100 million category is not entirely clear; however, it might be attributed to misjudgment by many respondents in that category, as consumer goods encompass the most basic family needs.

The results for industrial goods, on the other hand, were significant in the “Below 2 million,” “2 million to 5 million,” and “Above 100 million” turnover categories but were inexplicably insignificant in the “5 million to 50 million” and “50 million to 100 million” categories. The foodstuff predictor (including groceries) was significant across all turnover categories, highlighting the importance of food in human communities. This further demonstrates that well-managed investments in foodstuffs and grocery businesses can consistently succeed. See Figures 1 and 2 for pictorial clarity.

Figure 1. Bar chart of turnover predictors.

Source: ValuStats (VSP 2.0) February 2025.

Figure 2. Line graph of turnover predictors.

Source: ValuStats (VSP 2.0) February 2025.

These findings are in line with the findings of the works of Anderson (2019), Ying and Chen (2019), Dubé et al. (2016), Bryman and Bell (2015), Horngren et al. (2015), and Hwang and Jang (2012) on the decision criteria for the business type and location choice.

Conclusion

In conclusion, the importance of business location in the overall strategy and success of a business cannot be overlooked. A well-chosen location enhances customer reach, operational efficiency, cost structure, competitive advantage, and strategic flexibility. The selection of an optimal business location is a critical decision that requires careful consideration of both statistical and accounting perspectives. By employing regression analysis, factor analysis, cluster analysis, cost-benefit analysis, financial forecasting, and risk assessment, businesses can make more informed business choices and strategic location decisions. This integrated approach enhances the likelihood of operational efficiency and long-term profitability.

The implication of the findings of this study is that entrepreneurs who choose to ignore the importance of data analytics in business choice and locational decisions are likely to flounder and lose value, while the efforts of others with more discerning minds are crowned with great successes.

This study is limited in scope to the six locational and choice variables of consumer goods, industrial goods, supermarkets, foodstuffs, and rural and urban areas used in the study and covers only the southern and middle belt regions of Nigeria. Also, it covers only turnover as a measure of performance and did not consider profitability or any of its proxies. Future researchers can broaden the research horizon and include other factors not addressed in this study.

Ethical approval and consent to participate

In line with the Babcock University Health, Research, and Ethics guidelines, the telephone interviews conducted were scheduled through phone calls with the consent of the respondents obtained through the same. The Babcock University Health, and Research Ethics Committee (BUHREC) does not require researchers using questionnaire to obtain special written permission to conduct the study so long as the questionnaire does not contain any intimidating, coercive, or personalized questions. This implicit approval of BUHREC requires that no personal details of the respondents were to be collected or published in accordance with its guidelines. The execution of the data collection aspect of this research followed the BUHREC guidelines strictly.

Consent for data collection and publication

Informed consent of every respondent sampled was duly obtained orally during the data collection stage and the consent of those surveyed with phone calls were also obtained prior to every data collection conversation. No personalized details, images, or videos of any individual were used in the preparation of this document; thus, no consent for publication was required.