2.1. General Concept of Model
The concept of research relies on the developed model to choose the best product considering customer expectations and then predict the direction of improvement for current products to adjust them to the future requirements of customers. According to the literature review and earlier research, the model was developed so as to include in a combined way the following:
Individual customer requirements and expert opinion [
6,
31];
Actual parameters of the product criteria [
32,
33,
34];
Criterion of the impact on the natural environment [
33,
35,
36,
37];
The model is preferred for use in the case of choice of product for a customer and also to determine design and improvement actions for production enterprises. In addition, the environmental impact is included in this method. In turn, the prediction of improvement of the product will be realized according to customers’ opinions about the importance of the product criteria.
The model can be used to select each RES due to customer expectations. The idea is to support an expert (e.g., an entity offering a product) in predicting which of the products available on the market will be the most satisfying for the customer. It is possible to select each RES due to the fact that the analysis is based on current data on RES, which relate to the following: current customer expectations towards RES, current RES criteria (attributes) and their environmental impact, or the current purchase price. Furthermore, the analysis is supported by expert knowledge (the entity that uses the proposed model). This means that analyses can be carried out mainly on the basis of the customer’s expectations regarding the current parameters of the criteria of any RES, which do not affect the limitations in the selection of the product. The general concept of the model is shown in
Figure 2.
In the model, and integrated in a sequential way, the selected instruments support the decision making and assessment of product quality. These techniques are as follows: brainstorming (BM) [
41], SMARTER method [
42], rule 7 ± 2 [
31,
33,
43], a questionnaire with the Likert scale [
44], AHP method [
45,
46,
47], and PROMETHEE II method [
48,
49]. The analysis was also supported by the actual purchase price of the product in a matrix data analysis dedicated to this model [
6,
37].
Initially, the product for verification is selected, and the choice is made by an expert after consultation with the customer. Herein, it is possible to use brainstorming (BM). Then, for the selected subject of study, the purpose is determined. In the proposed approach, the purpose should refer to the choice of the product the most preferred by customers and then predict the direction of product improvement to meet customer expectations. The purpose can be determined by using the SMARTER method. The choice of this method, conditioned by applying this method to precision, determines the purpose of the research’s dependence on the verified subject of study, as shown in, e.g., [
42].
Then, all products should be characterized according to the criteria, where the number of products and criteria should be the result of the rule of decision methods and be an equal maximum of 7 ± 2 [
43]. This principle indicates that the customer is able to compare a maximum of five to nine criteria at the same time. Hence, this rule was assumed to be effective regarding the customers’ processing of the criteria and their assessment of the criteria’s importance. This characteristic should include the current parameters of the product criteria.
The customer expectations are obtained for these criteria, which should refer to the importance of the criteria for the customers and their satisfaction with the current quality of the product. In this case, the Likert scale is used, in which choice results from a product’s popularity in the evaluation by customers [
44].
Then, according to customer assessments, the weights of the criteria are estimated, and then, the product’s quality level is calculated. The AHP method is used to calculate the weight of the criteria. This method allows one to determine the weight (importance) of criteria according to comparison in pairs. It is an easy and effective method of comparison criteria, in which the number is consistent with that previously mentioned 7 ± 2 [
43].
Moreover, the PROMETHEE II method is used to develop a ranking of choice. Its choice is conditioned by the possibility of calculation by real data and also eventually including data obtained in a qualitative (description) way. Additionally, in this method, the analysis is performed considering the importance of criteria [
45,
46,
49]. As a result, a product-selection ranking is created, which concerns the quality of products according to the fulfilment of customer expectations. The qualitative price analysis is realized by using matrix data analysis [
37], which shows which product is the most favorable for quality, impact on the environment, and price of purchase.
The last stage of the model is dedicated to the entity (expert) that, according to weights of the criteria, can plan future improvement actions. The choice of improvement actions results from the preferences of customers and, for example, the company’s capabilities (resources and others).
2.3. Algorithm of Model and Its Description
As part of the development of the model, the procedure algorithm was developed. The algorithm was developed in seven main stages. These are shown in
Figure 3.
The characteristics of the model are shown in the next part of the study in seven main stages.
- Stage 1.
Selecting products and determining the purpose of research
The selection of products for verification is made by an entity (expert, broker, or bidder). The choice is also dependent on the individual preference of the customer. Therefore, it is recommended for the entity to initially consult with the customer. Then, the entity selects products of the same kind (type), e.g., with current availability or on sale.
The most preferred product will be selected from these products. Then, within these products, the direction of their improvement will be predicted. According to the authors of [
43], the number of products of the same type should be equal to 7 ± 2.
Then, the purpose of the research is determined according to the subject of the study. The purpose should be determined by an expert. In the proposed approach, the purpose should refer to selecting the product most preferred by customers and then predicting the direction of improvement of the product to meet customer expectations. To develop the purpose of the investigation, the SMARTER method is used [
42].
- Stage 2.
Characteristic of products according to criteria
It is necessary to characterize all products according to their criteria (attributes). Therefore, initially, for all products, it is necessary to select criteria that characterize them. These criteria should be mainly product criteria, i.e., impacting to a significant (importance) degree the quality of the product. These criteria should be qualitative (utility) and technical, e.g., measurable or unmeasurable. In the proposed approach, it is necessary to include the additional impact of the product on the natural environment.
Therefore, it is important that this criterion is considered among the others. After a review of the literature, e.g., according to authors of studies [
33,
35,
36,
37], it is crucial to include the criterion of impact on the natural environment (which refers to the negative impact on the natural environment, e.g., due to the materials used for its production, packaging, or the method of degradation/lifetime). This is due to the global need to care for the natural environment, which is weakened by climate change and further exacerbated by waste and a competitive production environment.
According to the literature review [
43], the number of product criteria should be equal to a maximum to 7 ± 2. The choice of criteria can be made according to the catalogue (specification) of the products. It is necessary to realize brainstorming (BM), as shown in studies, e.g., [
41]. After determining product criteria, it is necessary to describe (characterize) them, e.g., according to product criteria. It consists of assigning, e.g., measure values (parameters) or parameter ranges to the criteria. Characterization can be carried out in the summary table of products and their criteria.
- Stage 3.
Obtaining customer expectations
At this stage, it is necessary to acquire customer expectations in order to achieve the following: (a) choose a product satisfying the customer and (b) predict the direction of current product improvement for adjustment to future customer requirements. Therefore, it is necessary to obtain customer requirements, e.g., using a questionnaire or interview with the customer.
In order to choose products for customers and predict the direction of improving current products, it is necessary to obtain from them assessments determining the weights of product criteria, i.e., assessments of the importance of criteria for customers during use of the product. The popular scale of assessments is the Likert scale (1–5); hence, we choose this scale, where 1 means a practically irrelevant criterion, and 5 means a very important criterion [
44]. In order to obtain customers’ expectations, a questionnaire assessing all criteria using a Likert scale must be prepared. These assessments are processed in the next stages of the model.
- Stage 4.
Estimating product criteria weights
According to the assessments of customers referring to the weights (importance) of criteria, it is necessary to estimate the weights of criteria. The weights of criteria are used to calculate the quality of products in the next stage of the model but also to predict the direction for improving the current product. It was assumed that for this purpose, the AHP method will be used, the idea of which is to compare the evaluation of criteria in pairs. Therefore, initially, the weights of the evaluation of the criteria should be written in a pairwise comparison matrix (Saaty’s matrix) (1) [
45]:
where i, j = 1, 2, …, k; the proportion of the weight of i-th and j-th criterion from a square matrix of order
, where n—number of criteria, is determined as a domination matrix (2) [
37,
46]:
The diagonal of matrix S always contains values equal to 1. This means that the criteria compared with each other are equivalent. Therefore, above the diagonal are the values of pairwise comparisons, while below the diagonal are the inverse values.
Then, the calculation is performed based on Saaty’s matrix. The calculation of criteria weights is realized by using the geometric mean of S matrix rows, which are then normalized (3) [
37,
45,
46]:
The sum of values for the criterion should be equal to 1. Then, the correctness of the calculations is achieved, and it is assumed that values (
) are product criteria weights, i.e., the importance of criteria for the customer. Ultimately, it is necessary to check that the results obtained do not violate the principle of constancy of preferences. It is verified by calculating of inconsistency factor
, compliance factor CI, and inconsistency relationship CR, as shown in Formula (4) [
37,
45]:
where
—inconsistency factor; n—number of criteria; w—weight of criterion; i, j = 1, 2, …, n; r—the average value of the random index for n according to Saaty (
Table 2).
The compatibility of results is obtained when λ max = n, CI = 0, and CR = 0. The compatibility of results is acceptable for
close to n and for CI < 0.1 and CR < 0.1 [
46]. The lack of full or acceptable agreement of the obtained results proves that the grades were awarded inconsistently [
37]. In this case, the process of estimating weight should be repeated from the beginning. After obtaining the consistency of the results, the next stage of the model should be implemented.
- Stage 5.
Developing ranking of selecting products according to customer expectations
At this stage, it is necessary to create a ranking of products selection, according to which there will be the possibility to select a product that satisfies the customer, i.e., meet his expectations in terms of quality, price of purchase, and impact on the natural environment. When creating the ranking, it is necessary to rely on the criteria (from stage 2) for which the values of current parameters, purchase price, and environmental impact are determined. Unmeasurable (qualitative) criteria are assessed by an expert on a Likert scale, where 1 means the quality of the criterion practically does not meet customer satisfaction, and 5 means the quality of the criterion fully meets customer satisfaction [
44]. At the same time, this method should take into account the weight of the criteria, which was specified by the customer (from step 4). In order to create a ranking for selecting the products, the PROMETHEE II method is used (a method of the family of PROMETHEE methods) [
50].
In the PROMETHEE methods for each pair of variants, calculations are made to determine the aggregate index of preference and the positive and negative outflow flow. Positive outflow is a degree determining the outflow of one variant over the others. In turn, the negative outflow determines to what extent this variant outperforms the others. Decision ranking is created based on determined values of outflows [
48,
49,
50,
51].
The PROMETHEE II method is one of the variants of PROMETHEE methods as part of the creation of alternative variants [
48,
49]. In this approach, the purpose is to determine the synthetic ranking of alternatives (in the case of different products). It is realized using the techniques of comparison in pairs and outranking relationships [
49]. Firstly, the normalization of parameters and the evaluation of criteria is realized according to Formula (5):
where
– assessments (or parameters) of i-th and j-th product relative c-th criterion; i, j = 1, …, n.
Then, a comparison of products relative criteria is performed, as in formula [
49] (6):
where
– assessments (or parameters) i-th and j-th product relative c-th criterion; i, j = 1, …, n.
Then, the individual index of preference is determined. This index is determined by each pair of products relative to each criterion. In this aim, the preference function is adjusted to each criterion, for example, in the form of Formula (7): [
49,
50,
51]:
where a—product with respect to the c-th criterion; r—indicators of comparison in pair; i, j = 1,…, n; c = 1,…, m.
The next stage is determining multi-criteria indexes of preference. These indexes are determined for each pair of products [
48,
51] (8):
where
—weight of criterion;
.
Then, positive flow
, negative flow
, and net flow
are determined. The flows are determined for each product according to Formulas (9)–(11) [
50,
51]:
where a—product with respect to the i-th and j-th criteria; i, j = 1,…, n.
The ranking of products is created on net flows. In the classification, we assumed that the higher the net value (
, the more favorable the product is for customer. Furthermore, the positive value of net flow for the i-th product determines this product as superior to the others in terms of the analyzed criteria (and vice versa). Therefore, the higher the net flow, the better it is [
48,
49,
50,
51]. In the proposed approach, a ranking of the selection of the current product for the customer is created. On its basis, the product expected by the customer can be selected according to the product quality, environmental impact, and purchase price of the product.
- Stage 6.
Qualitative price analysis
In the proposed approach, it was assumed that the customer selected the product in view of the quality of product and its impact on the natural environment and also the price of purchase. This is determined by combining the value from product ranking (including quality and impact on the natural environment) with the real price of purchasing the product. Hence, it is necessary to obtain the unit price of purchasing the product, e.g., a single piece of product in the appropriate currency (e.g., euro). Combining ranking values of the product with the price of purchase is possible by using a quality management tool belonging to the group of new tools, i.e., table (matrix) data analysis. This tool is used to compare (graphic) dependencies between two variables. The procedure for creating the table data analysis in the proposed approach is the following:
- (1)
Calculate values from ranking products (considering the quality of the product and its impact on the natural environment);
- (2)
Estimate the current price of purchase for analyzed products;
- (3)
Draw coordinate axes for parameters ( values and price of purchase);
- (4
Scale or mark the axis;
- (5)
Enter the parameter values ( and price of purchase);
- (6)
Analyze the distribution of the values obtained.
As a result, it is necessary to choose a product that will meet customer quality expectations (product criteria and its impact on the natural environment) but will also satisfy the customer in view of the price of purchase.
- Stage 7.
Predicting direction of product improvement
At this stage, it is possible to predict the direction of product improvement considering customer expectations. It was assumed that prediction is realized based on the weight of product criteria, i.e., the importance of the criteria for the customer. These weights are calculated during the assessment of each customer in the fourth stage of the model using the AHP method. The idea of this stage is to determine which criteria should be used to start improving the quality of products to achieve customer satisfaction. This means identifying a group of product criteria that are worth improving in the first place, where changes in these criteria will determine a significant increase in customer satisfaction.
In the proposed approach, it is necessary to acquire and process the weights of the criteria specified by the customers when using the developed model in a given time range. As mentioned in the third stage of the model, weight evaluations should be obtained through the Likert scale (1–5), where 1 is a practically irrelevant criterion, and 5 is a very important criterion. To determine the ranking of improvement criteria, the value of the arithmetic mean of the obtained weights must be calculated, according to Formula (12) [
31]:
where w—weight of product criterion; n—number of customers; i = 1, 2, …, m.
This refers to the calculation of the arithmetic mean from weight assessments for each criterion. It is necessary to calculate the average weight for each criterion selected for verification. As a result, the ranking of product criteria according to their weights is obtained. The first position in the ranking is the most important criterion, and the last position in the ranking is the least important criterion. Based on the obtained ranking, the production enterprises can select strategies for improving products, e.g., improving selected criteria according to this ranking.