Research on the post-purchase behaviour tendency of the product quality to customers in fast-selling marketing

Fast-selling enterprises attach importance to repurchase and urgently retain high-quality repeat customers, so that such consumers can ‘come back’ to buy their own goods [20]. Because the concept of ‘fast-selling products’ is for merchants to deeply study the consumption regularity of such products, the concept of ‘fast-selling products’ is only within the research scope of merchants and will not have any negative impact on consumers themselves.

This paper combs the evolution of the consumer market from four aspects: the whole FMCG market, food and beverage, beauty cosmetics and brands. The general trend of FMCG consumption is slowly rising, and the consumption level of related categories that improve people's quality of life is rising rapidly [21, 22]. Although the domestic sales of FMCG in China dropped in the first three quarters compared with the same period of last year, the overall consumption level has steadily recovered due to the continuous development of the domestic middle class and the return of residents' consumption under the control of epidemic travel. After a steady growth of around 5% per year and a stagnation period of economic growth in 2020, in the first 9 months of 2021, the total sales volume and sales volume of global FMCG products will increase by 3.3% and 3.6%, respectively [23]. Although the growth rate is slow, the gap between price growth rate and China's inflation rate will still narrow compared with that in 2020.

Because of the differences in commodity characteristics and importance, people spend different amounts of time and spirit in making shopping decisions for various commodities, and the consumption process changes accordingly. However, when people choose a certain product, people usually have a decision-making process, which is different due to the differences in the types of commodities purchased and the types of consumers served, but there are still certain rules [24]. The typical purchasing decision-making steps are mainly divided into five aspects: determining demand, obtaining information, comparing and evaluating, choosing actions and post-purchase reaction (Figure 1).

Fig. 1

In fact, people do not stipulate that consumers have to go through these five stages when purchasing every product. The shopping process of some goods may be relatively simple, but consumers themselves will skip one stage or reverse another [25, 26]. As a relatively comprehensive shopping process, it tells about all kinds of considerations and activities made by consumers themselves when they are faced with brand-new markets or more complicated shopping situations [27].

Consumer's purchasing activity is a very complicated and changeable process. Consumer's own psychology, economic and social environment and various factors of enterprise marketing are the three main factors that affect consumer's choosing activities [30]. At the same time, the influence of various factors on consumer groups' choices is quite complicated. If the process of consumer groups' choices is examined again, the whole research structure will change dynamically, and it is difficult to capture the correlation between them. Although these research results are comprehensive, they contain a huge amount of research information, and the results of scientific research cannot explain the actual consumer groups' choices [31]. The understanding of in-depth research results will also be simpler and more credible, and the significance of this main factor in the commodity model will be obvious.

After-sales service reliability has a strong influence on the quality of goods, which is also reasonable. Therefore, after-sales service reliability also represents the most critical interest that consumers seek when purchasing in stores [32]. However, the quality of service facilities and the environmental conditions have a strong influence, and suitable working conditions and convenient service facilities are also the main influencing factors for the formation of consumer satisfaction. To some extent, this shows that Chinese consumers' satisfaction with supermarket services is deeply influenced by their attention to the experience of the supermarket hardware environment. The importance of ‘serving customers from the perspective of consumers’ is also very strong, which shows that they pay more attention to the personalised enjoyment they get from shopping in supermarkets [33].

After comparing various commodity theory models, it is found that Bell's model provides a comprehensive evaluation of commodities, which can basically cover all the information on commodities. According to the market environment of fast-selling products and their image characteristics, Bell's model is used to divide the product quality of fast-selling products into three aspects: enterprise image, product image and user image, thus forming a theoretical model suitable for this study, as shown in Figure 2.

Fig. 2

Based on the above theoretical considerations, Bell put forward that commodity quality can be characterised by three seed images: commodity quality, product or service image and user image, and the author also intends to take these three dimensional images as the starting point and provide the basic assumptions of this survey [34] (Figure 3).

Fig. 3

In the study of the influence of the quality of fast-selling products on the shopping behaviour of consumer groups, the author first used Bell's measurement model, including the product quality, product and service representation and user representation, and then added some related background variables, such as the characteristics of consumer groups. In the process of consumers' purchase, it is generally stipulated that seven hypotheses should be provided to study the influence of these factors on the shopping index when making purchase decisions. The research model structure is as follows (Table 1).

Partial least squares (PLS) modelling theory is applied to the analysis and modelling of customer post-purchase behaviour (PPB) intention. Without losing generality, we define independent variable data table X and dependent variable data table Y as follows: (1) X=(X1,X2,⋯,Xp)=(x11x21⋯xpx12x22⋯xp⋮⋮⋮⋮x1nx2n⋯xpn) X = \left({{X_1},{X_2}, \cdots,{X_p}} \right) = \left({\matrix{{{x_{11}}} & {{x_{21}}} & \cdots & {{x_p}} \cr {{x_{12}}} & {{x_{22}}} & \cdots & {{x_p}} \cr \vdots & \vdots & \vdots & \vdots \cr {{x_{1n}}} & {{x_{2n}}} & \cdots & {{x_{pn}}} \cr}} \right) (2) Y=(Y1,Y2,⋯,Yq)=(y11y21⋯yq1y12y22⋯yq2⋮⋮⋮⋮yny2n⋯yqn) Y = \left({{Y_1},{Y_2}, \cdots,{Y_q}} \right) = \left({\matrix{{{y_{11}}} & {{y_{21}}} & \cdots & {{y_{q1}}} \cr {{y_{12}}} & {{y_{22}}} & \cdots & {{y_{q2}}} \cr \vdots & \vdots & \vdots & \vdots \cr {{y_n}} & {{y_{2n}}} & \cdots & {{y_{qn}}} \cr}} \right)

PLS combines the characteristics of non-model cognitive methods such as principal component analysis, canonical correlation analysis and multiple regression modelling methods. The basic principle is as follows. First, the first PLS components t₁ and u₁ are extracted from data tables X and Y, respectively, and t₁ and u₁ are required to represent the data variation information in data tables X and Y, respectively. According to the principle of principal component analysis, the variance of t₁ and u₁ is reached.

Maximum extraction conditions: (3) Var(t1)→max \mathop {{\rm{Var}}}\nolimits_ \left({{t_1}} \right) \to \max (4) Var(u1)→max \mathop {{\rm{Var}}}\nolimits_ \left({{u_1}} \right) \to \max

At the same time, due to the need for regression modelling, t₁ needs to have the maximum explanatory power to u₁. According to the principle of typical correlation analysis, when the correlation coefficient between t₁ and u₁ reaches the maximum, t₁'s explanatory power to u₁ reaches the maximum, that is: (5) r(t1,u1)→max r\left({{t_1},{u_1}} \right) \to \max

At the same time, considering the above two factors, the PLS regression analysis, i.e. the maximum covariance of t₁ and u₁ is the extraction condition: (6) Cov(t1,u1)=Var(t1)⋅Var(u1)⋅r(t1,u1)→max \mathop {{\rm{Cov}}}\nolimits_ \left({{t_1},{u_1}} \right) = \sqrt {\mathop {{\rm{Var}}}\nolimits_ \left({{t_1}} \right) \cdot \mathop {{\rm{Var}}}\nolimits_ \left({{u_1}} \right)} \cdot r\left({{t_1},{u_1}} \right) \to \max

The above comprehensive analysis can summarise the PLS modelling algorithm as follows:

Standardised independent variable data table and dependent variable data table are obtained from the original data tables X and Y. (7) E0=(E01E02⋯E0p) {E_0} = \left({\matrix{{{E_{01}}} & {{E_{02}}} & \cdots & {{E_{0p}}} \cr}} \right) (8) F0=(F01F02⋯F0q) {F_0} = \left({\matrix{{{F_{01}}} & {{F_{02}}} & \cdots & {{F_{0q}}} \cr}} \right) (9) (1) wi=Ei−1T⋅ui {\rm{(1)}}\;{w_i} = E_{i - 1}^T \cdot {u_i} (10) (2) And standardise ‖wi‖=1 {\rm{(2)}}\;{\rm{And}}\;{\rm{standardise}}\;\left\| {{w_i}} \right\| = 1 (11) (3) ti=Ei−1⋅wi {\rm{(3)}}\;{t_i} = {E_{i - 1}} \cdot {w_i} (12) (4) bi=uiT⋅ti/‖ti‖and standardise ‖bi=1‖ {\rm{(4)}}\;{b_i} = u_i^T \cdot {t_i}/\left\| {{t_i}} \right\|{\kern 1pt} {\rm{and}}\;{\rm{standardise}}\;\left\| {{b_i} = 1} \right\| (13) (5) Fresidual=Fi−1−bi⋅ti⋅uiT {\rm{(5)}}\;{F_{{\rm{residual}}}} = {F_{i - 1}} - {b_i} \cdot {t_i} \cdot u_i^T (14) (6) Eresidual=Ei−1−ti⋅piT {\rm{(6)}}\;{E_{{\rm{residual}}}} = {E_{i - 1}} - {t_i} \cdot p_i^T (15) (7) Another i=i+1, Ei−1=Eresidual, Fi−1=Fresidual {\rm{(7)}}\;{\rm{Another}}\;i = i + 1,\;{E_{i - 1}} = {E_{{\rm{residual}}}},\;{F_{i - 1}} = {F_{{\rm{residual}}}}

(8) Turning back to Step (1), after all PLS components are extracted, there are: (16) E0=t1⋅pT+t2⋅p2T+…+tm⋅pmT {E_0} = {t_1} \cdot {p^T} + {t_2} \cdot p_2^T + \ldots + {t_m} \cdot p_m^T (17) F0=t1⋅b1T+t2⋅b2T+…+tm⋅bmT {F_0} = {t_1} \cdot b_1^T + {t_2} \cdot b_2^T + \ldots + {t_m} \cdot b_m^T

Because t₁ is a linear combination of E_0j and F_0k is a linear combination of t_i, F_0k is written in linear combination form about E_0j (18) (i=1,2⋯m; j=1∼p, k=1∼q) (i = 1,2 \cdots m;\quad j = 1 \sim p,\quad k = 1 \sim q)

According to the reverse process of standardisation, reducing E₀F₀ to x y includes the following steps: (19) F0k=α1k⋅E01+α2k⋅E02+⋯αpk⋅E0P (k=1∼q) {F_{{0_k}}} = {\alpha _{1k}} \cdot {E_{01}} + {\alpha _{2k}} \cdot {E_{02}} + \cdots {\alpha _{pk}} \cdot {E_{0P}}\quad (k = 1 \sim q) (20) Yk=α1k*⋅X1+α2k*⋅X2+…αpk*⋅XP+C (k=1∼q) {Y_k} = \alpha _{1k}^* \cdot {X_1} + \alpha _{2k}^* \cdot {X_2} + \ldots \alpha _{pk}^* \cdot {X_P} + C\quad (k = 1 \sim q)

According to the customer value (CV) theory, when the perceived utility for customers is higher than the perceived cost, the calculation formulas of perceived value for risk-averse customers and risk-oriented customers are as follows: (21) Ua=U−U(edSCSC−1)−U(edMCMC−1)−U(edPCPC−1)−U(edRCRC−1)−U(edECEC−1) {U^a} = U - U\left({{e^{{d_{SC}}SC}} - 1} \right) - U\left({{e^{{d_{MC}}MC}} - 1} \right) - U\left({{e^{{d_{PC}}PC}} - 1} \right) - U\left({{e^{{d_{RC}}RC}} - 1} \right) - U\left({{e^{{d_{EC}}EC}} - 1} \right) (22) Un=U−US¯C¯SC−UM¯C¯MC−UP¯C¯PC−UR¯C¯RC−UE¯C¯EC {U^n} = U - {U \over {\bar S\bar C}}SC - {U \over {\bar M\bar C}}MC - {U \over {\bar P\bar C}}PC - {U \over {\bar R\bar C}}RC - {U \over {\bar E\bar C}}EC

First of all, the value of online search can be obtained by calculating the difference between online search utility and search cost.

Yes. Based on the above literature analysis and the AHP method, this paper obtains that the search utility perceived by customers online is (23) Us=UQs+USs+UCs=0.4PQ+0.3SQ+0.3CN {U^s} = U_Q^s + U_S^s + U_C^s = 0.4PQ + 0.3SQ + 0.3CN

Through the extreme value test, the determination of coefficient is reasonable and feasible. Therefore, the online search cost perceived by customers can be obtained as shown in the following formula. (24) Cs=U(e12PQ2PQ−1)+U(e14CN2CN−1)+U(e13SQ2SQ−1) {C^s} = U\left({{e^{{1 \over {2P{Q^2}}}PQ}} - 1} \right) + U\left({{e^{{1 \over {4C{N^2}}}CN}} - 1} \right) + U\left({{e^{{1 \over {3S{Q^2}}}SQ}} - 1} \right)

Because the customer-perceived value can be obtained by calculating the difference between perceived utility and perceived cost, the customer-perceived online search value can be obtained: (25) Vs=Us−Cs=0.4PQ+0.3SQ+0.3CN−Us(e12PQ−1)−Us(e14CN−1)−Us(e13SQ−1) {V^s} = {U^s} - {C^s} = 0.4PQ + 0.3SQ + 0.3CN - {U^s}\left({{e^{{1 \over {2PQ}}}} - 1} \right) - {U^s}\left({{e^{{1 \over {4CN}}}} - 1} \right) - {U^s}\left({{e^{{1 \over {3SQ}}}} - 1} \right)

With the same method as above, we can get the utility function and cost function of customers in the purchasing process. (26) Up=UQp+UCp+USp=0.54PQ+0.16CN+0.3SQ {U^p} = U_Q^p + U_C^p + U_S^p = 0.54PQ + 0.16CN + 0.3SQ (27) Cp=U(e23PQ2PQ−1)+U(e13CN2CN−1)+U(e12SQ2QQ−1) {C^p} = U\left({{e^{{2 \over {3P{Q^2}}}PQ}} - 1} \right) + U\left({{e^{{1 \over {3C{N^2}}}CN}} - 1} \right) + U\left({{e^{{1 \over {2S{Q^2}}}QQ}} - 1} \right)

Customer's perceived online shopping value: (28) Vp=Up−Cp=0.54PQ+0.3SQ+0.16CN−Up(e23PQ−1)−Up(e13CN−1)−Up(e12SQ−1) {V^p} = {U^p} - {C^p} = 0.54PQ + 0.3SQ + 0.16CN - {U^p}\left({{e^{{2 \over {3PQ}}}} - 1} \right) - {U^p}\left({{e^{{1 \over {3CN}}}} - 1} \right) - {U^p}\left({{e^{{1 \over {2SQ}}}} - 1} \right)

The perceived utility and cost for customers after purchase are as follows: (29) UA=UQA+UCA+USA=0.65PQ+0.1CN+0.25SQ {U^A} = U_Q^A + U_C^A + U_S^A = 0.65PQ + 0.1CN + 0.25SQ (30) CA=U(e34PQ2PQ−1)+U(e12CN2CN−1)+U(e12SQ2QQ−1) {C^A} = U\left({{e^{{3 \over {4P{Q^2}}}PQ}} - 1} \right) + U\left({{e^{{1 \over {2C{N^2}}}CN}} - 1} \right) + U\left({{e^{{1 \over {2S{Q^2}}}QQ}} - 1} \right)

Customer's perceived online post-purchase value: (31) VA=UA−CA=0.65PQ+0.1CN+0.25SQ−UA(e34PQ−1)−UA(e12CN−1)−UA(e12SQ−1) {V^A} = {U^A} - {C^A} = 0.65{\rm{PQ}} + 0.1CN + 0.25SQ - {U^A}\left({{e^{{3 \over {4PQ}}}} - 1} \right) - {U^A}\left({{e^{{1 \over {2CN}}}} - 1} \right) - {U^A}\left({{e^{{1 \over {2SQ}}}} - 1} \right)

Based on the purpose of data analysis and the requirements of the analysis process, this study used EXCEL software and SPSS data management software to carry out resource management and data analysis. The methods adopted include the following five.

In the long run, for brand development, fast-selling products should pay attention to the material of products, improve the quality from the source of the industrial chain and promote the overall improvement of product quality. Especially in the process of brand development, only by grasping the balance between speed and quality, truly establishing a good brand image and fashion with quality assurance can we stand the test of the market and be recognised by consumers.

Product competition and brand competition are the main aspects of enterprises' marketing activities. Product competition is mainly reflected in the degree of functional satisfaction, while brand competition lies in satisfying consumers' emotional and psychological demands. Brand culture plays an important role in enhancing a brand's cultural competitiveness.

I hope that the products used by Bai Ji can be branded with Bai Ji, which can reflect Bai Ji's unique personality, interests and mood, and the brand personality can realise this desire of consumers. The concept and personality of the brand impress consumers' hearts by satisfying their psychological needs, and they have a desire to buy and become loyal customers.

From the perspective of consumers, based on Bell's model, this study divides the image of FMCG into three dimensions, namely company image, product/service image and user image, and explores the relationship between them and their influence on consumers' purchasing behaviour, so as to provide a reference for FMCG enterprises in shaping and maintaining their brand image. Through the correlation analysis of survey data and empirical research of questionnaire survey and data analysis, the following conclusions are drawn: the attributes of FMCG products are positively correlated with PPB and CV. In other words, the image of products and services, as well as the image of companies and users, has a significant impact on consumers' purchasing behaviour. Consumers evaluate the attractiveness of the market offerings according to the characteristics and quality of products, the combination and quality of services and the appropriate price of the offerings. When purchasing products, consumers are most concerned about product quality and price, which serve as the basis for decision-making.

Based on Bell's model, according to the characteristics of fast-selling products and the reality of China's consumer market, this study designed a questionnaire on the influence of fast-selling products' image on consumers' purchasing behaviour, in order to explore the role and weight of each dimension of fast-selling product image on consumers' purchasing behaviour, and drew some conclusions through empirical analysis. Throughout the whole research process, due to the limitation of the author's ability and time, there are some shortcomings in the paper.

The research design is insufficient: On the basis of synthesis of the research of relevant scholars, the author, according to his own understanding and considering the applicability of the research, will inevitably miss something.