Patterns of the Distribution of the Demand of End-Consumers among Retailers in the Zone of their Residence

The process of consuming goods is closely linked to the concept of process of consumption. The cost of actions in the process of consumption of goods was estimated using the formula presented in Halkin (2020). Traditional shopping deals with physical movement to store and back and trade service inside of it. In e-commerce option, the physical attendance of retailer is missing. End-consumer chooses goods via Internet and receives delivery of goods in the final stage. Both options can be chosen, especially in today's pandemic circumstances.

The work includes study of the area of demand formation at the initial stage, establishment of areas of residence of consumers, major competitors, and geographic features of the area, in accordance with the method proposed by Galkin, et al. (2020). Based on these data, a matrix of probability of choosing the option of consumption by end-consumers from different areas of residence was generated.

The research interest lies in the conditions under which the end user will choose one option over another, for example, online option instead of traditional shopping. It is also interesting to know how to change the existing probabilities of choosing retailers when adding an Internet option.

The systematic approach allows to distribute material flows across retailers (Fig. 1).

Figure 1

The total consumer demand in the z – service area for the m-th product for the analyzed period is: (1) Qmz=∑J=1J∑w=1W(Nωz⋅(∑m=1MQωzM′)) {Q_{mz}} = \sum\limits_{J = 1}^J {\sum\limits_{w = 1}^W {\left( {{N_{\omega z}} \cdot \left( {\sum\limits_{m = 1}^M {Q_{\omega zM}^\prime } } \right)} \right)} } where N_ωz is the number of end-consumers living in the ω-th building area, people; QωzM′ Q_{\omega zM}^\prime the rate of consumption of goods on average per 1 buyer living in the w-th building area for the analyzed period, kg/inhabitant; ω the number of consumer's zones, ω = 1, 2, …, I; j the number of retailers in the trade urban zone, j = 1, 2, …, J.; and m is the goods type, m = 1, 2, …, M.

In a closed system, the total demand for the m-th product is defined as the sum of demand for all retail outlets and demand that is realized through the Internet: (2) Qmz=Qmzj+Qmze+Qmzo {Q_{mz}} = {Q_{mzj}} + {Q_{mze}} + {Q_{mzo}} where:

Q_mzj is the quantity of goods sold through the retailer, kg;

The quantity of sales of the m-th product through j-th retailer is defined as: (3) Qmzj=∑ω=1W(Pzωj⋅Nωz⋅(∑m=1MQωzM′)) {Q_{mzj}} = \sum\limits_{\omega = 1}^W {\left( {{P_{z\omega j}} \cdot {N_{\omega z}} \cdot \left( {\sum\limits_{m = 1}^M {Q_{\omega zM}^\prime } } \right)} \right)}

The volume of sales of M-th goods via the Internet in the considered shopping area can be determined using the formula: (4) Qmze=∑w=1W(Pmωe⋅Nωz⋅(∑m=1MQωzM′)) {Q_{mze}} = \sum\limits_{w = 1}^W {\left( {{P_{m\omega e}} \cdot {N_{\omega z}} \cdot \left( {\sum\limits_{m = 1}^M {Q_{\omega zM}^\prime } } \right)} \right)}

The quantity of m-th goods sold through other retailers in this trading area for a period of time is determined using the formula: (5) Qmzo=Qmz−Qmzj−Qmez=Qmz⋅Pmωo {Q_{mzo}} = {Q_{mz}} - {Q_{mzj}} - {Q_{mez}} = {Q_{mz}} \cdot {P_{m\omega o}} where P_mωo is the probability of choosing retailers in addition to those located in the shopping area when buying the m-th goods.

Every consumer's choice involves costs. Based on the total losses for each alternative choice of retailer, the market share that will be sold for it is determined. Market share owned by retailer j in the sale of the m-th product is determined as follows: (6) Emzj=QmjzQmz=∑ω=1W(Pmωj⋅Nωz⋅(∑m=1MQωzM′))∑j=1J∑ω=1W(Nωz⋅(∑m=1MQωzM′)) {E_{mzj}} = {{{Q_{mjz}}} \over {{Q_{mz}}}} = {{\sum\limits_{\omega = 1}^W {\left( {{P_{m\omega j}} \cdot {N_{\omega z}} \cdot \left( {\sum\limits_{m = 1}^M {Q_{\omega zM}^\prime } } \right)} \right)} } \over {\sum\limits_{j = 1}^J {\sum\limits_{\omega = 1}^W {\left( {{N_{\omega z}} \cdot \left( {\sum\limits_{m = 1}^M {Q_{\omega zM}^\prime } } \right)} \right)} } }}

Market share owned by Internet sales of m-th goods in the residence zone is found using the formula: (7) Emze=QmezQmz=∑ω=1W(Nωz⋅Pmωe⋅(∑m=1MQωzM′))∑j=1J∑ω=1W(Nωz⋅(∑m=1MQωzM′)) {E_{mze}} = {{{Q_{mez}}} \over {{Q_{mz}}}} = {{\sum\limits_{\omega = 1}^W {\left( {{N_{\omega z}} \cdot {P_{m\omega e}} \cdot \left( {\sum\limits_{m = 1}^M {Q_{\omega zM}^\prime } } \right)} \right)} } \over {\sum\limits_{j = 1}^J {\sum\limits_{\omega = 1}^W {\left( {{N_{\omega z}} \cdot \left( {\sum\limits_{m = 1}^M {Q_{\omega zM}^\prime } } \right)} \right)} } }}

Market share of the producer of the m-th product, which is sold in the residence zone z by other market participants, is given as: (8) Emzo=Qmzo−Qmjz−QmezQmz=∑j=1J∑j=1W(Nωz⋅(∑m=1MQωzM′))−∑j=1W(Pmωe⋅Nω⋅(∑m=1MQωzM′))−∑j=1J∑ω=1W(Pmωj⋅Nω⋅(∑m=1MQωzM′))∑j=1J∑j=1W(Nωz⋅QωzM′) {E_{mzo}} = {{{Q_{mzo}} - {Q_{mjz}} - {Q_{mez}}} \over {{Q_{mz}}}} = {{\sum\limits_{j = 1}^J {\sum\limits_{j = 1}^W {\left( {{N_{\omega z}} \cdot \left( {\sum\limits_{m = 1}^M {Q_{\omega zM}^\prime } } \right)} \right)} } - \sum\limits_{j = 1}^W {\left( {{P_{m\omega e}} \cdot {N_\omega } \cdot \left( {\sum\limits_{m = 1}^M {Q_{\omega zM}^\prime } } \right)} \right) - } \sum\limits_{j = 1}^J {\sum\limits_{\omega = 1}^W {\left( {{P_{m\omega j}} \cdot {N_\omega } \cdot \left( {\sum\limits_{m = 1}^M {Q_{\omega zM}^\prime } } \right)} \right)} } } \over {\sum\limits_{j = 1}^J {\sum\limits_{j = 1}^W {\left( {{N_{\omega z}} \cdot Q_{\omega zM}^\prime } \right)} } }}

The research hypothesis is based on the choice of the retailer by a certain group of customers, based on generalized costs that they will incur as a result of their choice. Each end-consumer choice leads to costs generation. The generalized costs of visit by residents of the ω-th neighborhood of the z zone who visit the j-retailer (Galkin, Dolia, and Davidich, 2017) are given as: (9) Θtotalwj=Θ1ωj+Θ2ωj+Θ3ωj+Θ4ωj+Θ5ωj→min \Theta _{total}^{wj} = \Theta _1^{\omega j} + \Theta _2^{\omega j} + \Theta _3^{\omega j} + \Theta _4^{\omega j} + \Theta _5^{\omega j} \to \min where Θ1ωj \Theta _1^{\omega j} is the monetary expression of shopping time spent on purchases, UAH.; Θ2ωj \Theta _2^{\omega j} the monetary costs on the purchase of goods in the j-retailer, UAH; Θ3ωj \Theta _3^{\omega j} the monetary expression of the energy (fatigue) spent on purchases, UAH; Θ4ωj \Theta _4^{\omega j} the monetary expression of other non-price factors, UAH; and Θ5ωj \Theta _5^{\omega j} is the cost related to delivery, UAH.

Increasing end-consumer losses will reduce the frequency of choice of logistics system and market share that belongs to it and vice versa.

For successful practical implementation of the research, we divided our method into three stages: 1)

Preliminary analysis, which includes the method of boundaries of the store's service area setting (Balandina,, et al., 2019) and socioeconomic data collection. This stage will provide information on end-consumers and their behavior (incomer, consumer basket, age, gender, etc.) and on competitor pool (stores, products, prices, services, etc.).

The total time taken to make a purchase via the Internet is calculated by the formula: (10) Tkpurch=Tkcomp+Tk′mov T_k^{purch} = T_k^{comp} + T_k^{\prime mov} where Tk′mov T_k^{\prime mov} is the walking time to receive the goods, hrs and Tkcomp T_k^{comp} is the working at the computer time, hrs.

The number of calories a person consumes when shopping online is calculated by the formula: (11) KCalovrI=Tk′mov⋅KCalhmov+Tkcmp⋅KCalhcomp=Tk′mov(BMR/24)⋅METwalk+Tkcomp⋅(BMR/24)⋅METcomp \matrix{ {KCal_{ovr}^I = T_k^{\prime mov} \cdot KCal_h^{mov} + T_k^{cmp} \cdot KCal_h^{comp} = } \hfill \cr {T_k^{\prime mov}\left( {BMR/24} \right) \cdot ME{T_{walk}} + T_k^{comp} \cdot \left( {BMR/24} \right) \cdot ME{T_{comp}}} \hfill \cr } where KCalovrI KCal_{ovr}^I is the number of calories a person consumes when buying goods online (working on PC, tablet, cell phone, etc.), kcal; MET_walk the metabolic equivalent of task (while walking), kcal; MET_comp the metabolic equivalent of task (while working at a computer), kcal.; and BMR is the basic metabolic rate in 24 hrs, kcal.

In BMR calculations, the Mifflin-St. Jeor formula was used for different sexes (Mifflin, 1990). When calculating MET, reference data from the following sources were used: Jetté, Sidney, and Blümchen (1990) and Ainsworth, et al. (2000). The cost of human energy was estimated using the formula presented in Halkin (2018) and Galkin, et al. (2020).

To study the patterns of consumption of goods via the Internet, six online stores were selected that specialized in delivering products to the consumer. All of them were characterized by different costs of the “consumer basket,” web page interfaces, range, and cost of delivery. Online stores such as “Rost”, “Tavria B”, “Produktoff”, “Metro”, “Vitok” and “Gastronom” were presented. An example of one of the online stores is presented in Fig. 2.

Figure 2

Information on the range and cost of delivery is presented in Table 1.

At the next stage, respondents were invited to make purchases online. The market of food products included in the consumer basket was considered as the labor market. Shopping mobility was considered as a leisure time action. The consumer basket is an approximate estimated set, the range of goods that characterizes the typical level and structure of monthly (annual) consumption of a person or family. This set was used to calculate the minimum consumer budget (subsistence level), based on the value of the consumer basket at current prices.

The consumer basket also serves as a basis for comparing estimated and real levels of consumption, as well as a basis for determining the purchasing power of currencies. The participants of the experiment had to go to the website of the online store and add to the “basket” the products included in the consumer basket, place an order for delivery, and after a while receive the order at the place of residence. Delivery time itself was not tracked. The main indicators were the time and number of calories burned.

Time costs were determined by timing. Walking time is defined as a constant for all online stores, as the walking time to get the delivery at the door or at the entrance will be the same in all cases.

The results of time spent are presented in Table 2.

The results of calorie expenditure according to the time spent on the computer and the Mifflin-St. Jeor formula (Mifflin, 1990) are presented in Table 3.

The cost of time of the buyer in monetary terms and energy costs of the consumer in monetary terms as a result of the purchase via the Internet are presented in Table 4.

The cost of 1 hr of the consumer was calculated for the level of the average salary in the country, that is, 8093.35 UAH/month ($ 300) in Ukraine (Kolodii, 2020).

The retailer's service area has been set up according the influence of end-consumers (Verroios, Efstathiou and Delis, 2011), trip-based approach (Russo and Comi, 2012), and the geographic characteristics of the territory (Comi, 2020). Considering the influence of these parameters will allow adjusting the limits of the retailer's service area, which will more accurately evaluate the demand. The presented zone is the basis for gathering socioeconomic information about the service area (Galkin, et al., 2019a). Taking into consideration the influence of these parameters will allow to adjust the boundaries of the service area. In order to estimate the boundaries, an algorithm is used, which consists of five stages (Halkin, 2020). The area of research was selected according to a previous research (Figs 2 and 3). The study area was chosen based on the density of the building and the presence of shopping facilities and an online store.

Figure 2

Figure 3

Presented trading zone analysis (Fig. 3) indicates a large cluster of retailers in it – 11 units. Twenty-five retailers are not included in the service area and waived after correction. According to this, it is possible to determine the area, quantity of inhabitants, end-consumers’ income and behavior, distances from buildings to each retailer, etc.

In Table 5 is presented distance from the building area to the retailer by air and along the road. The trip-based method was used for customer movement calculations (Galkin, et al., 2019b).

Table 5 shows the distances from neighborhoods to shopping facilities by air and along the road, as well as for the option from the online store – L. For the online store, the value of the nonlinearity coefficient was not calculated, so its value is 1. According to the approach given in Halkin (2020), calculations of other parameters and indicators to determine the probability of choosing a purchase option were performed. The results are presented in Table 6.

For an online store, travel time refers to the time a person spends on getting an order. Its value was equal to 0.02 hr, that is, 1 min. The time for sales service for the online store was equal to the time spent working on the computer and was 0.23 hr, that is, 15.2 min. The total time for online shopping was equal to 0.25 hr.

The results of the calculations showed the effect of the additional option on the probability of choosing existing options: 1)

as the number of options increases, the average probability of selecting an option will decrease;

Retailers D and G will have the highest probability of visiting and retailer B the lowest due to high generalized costs for them – without the Internet delivery channel (according to Halkin, 2020). With the addition of the Internet, retailers D and G remain the most attractive and the least attractive option is L – e-commerce. Less-attractive Internet options are associated with the poor development of this market in Ukraine. The study was conducted at the end of 2019 before the coronavirus pandemic. This also contributed to the result. Households traditionally walk to grocery stores (Zatonatska, et al., 2019)

The distance from the area of residence of end-consumers to the retailers is a significant factor that affects the costs when choosing a particular purchase option. Increasing the specified distance reduces the probability. It is advisable to determine the distance at which the end-consumer does not care which option to use – the Internet or traditional purchases. For a distance exceeding this value, the end user should choose e-commerce because its generalized costs will be lower than with traditional purchases. Thus, it is possible to conclude that the generated costs for comparable schemes of distribution (on and off-line) are equal for set group of parameters.

To assess the profitability of online purchases, the sum of two indicators was calculated (energy and time costs in monetary terms): (12) Θcall+timeij=Θ1ij+Θ3ij \Theta _{call + time}^{ij} = \Theta _1^{ij} + \Theta _3^{ij}

Fig. 4 presents a comparative graph of the total cost of purchase in a regular store and the cost of delivery when ordering products via the Internet (50 UAH), while the cost of the consumer basket was not taken into account (Table A1).

Figure 4

With a delivery cost of 50 UAH, the percentage of options below the delivery cost was 93.02%, that is, 80 points; 6 points were above this mark. It is possible to draw a conclusion that at the cost of delivery of 50 UAH to inhabitants of the area of research, it is more favorable to make purchases independently, without using delivery. As the affluence of families and the cost of time increase, the number of end users who choose e-commerce will increase, which is also confirmed in such studies.

Different areas (service areas) have different meanings of the factors that describe them. Thus, the population living in a particular area, the level of their income and the expenditure of residents, and the prices of goods from retailers determine the potential market share (Fatoki, 2020). Taking into account the parameters of cities (density of the transport network, the level of motorization and population density, the size of the service area) allows us to generate technologies for the promotion of freight flows for different levels of the country, region, or city

The level of organization of the competitive environment can facilitate or complicate the acquisition process and affect the fatigue and emotional state of buyers. The environment is characterized by economic, technological, political, demographic, and natural processes in it.