An Ecological, Power Lean, Comprehensive Marketing Evaluation System Based on DEMATEL–CRITIC and VIKOR: A Case Study of Power Users in Northeast China

Abstract

The reduction of carbon emissions in the power industry will play a vital role in global decarbonization. The power industry has three main strategies to achieve this reduction in emissions: to implement lean marketing strategies that effectively target users of power and encourage them to adopt decarbonizing technologies and services; to optimize the efficiency of these users of power; and to improve the efficiency of renewable energy sources. This paper establishes a comprehensive evaluation system of indexed data from power industry customers for the development of lean marketing strategies. This system evaluates indexes derived from customer data on renewable energy sources, carbon emissions, energy efficiency, and customer credit. It adopts the DEMATEL–CRITIC combination weight assignment and VIKOR method for system evaluation and conducts simulation experiments on customer data in a region of Northeastern China to give an example of how this method could be applied in practice to lean marketing. The results show that the evaluation system proposed in this paper can govern the lean marketing decision-making of power sales enterprises.

1. Introduction

Given the current context of anthropogenic global warming, many countries have prioritized energy strategies that reduce or even eliminate carbon. The electric power industry, as a part of every nation’s core infrastructure, must play a leading role in decarbonization, especially for countries with nationalized energy industries such as China. [1,2] The drive for decarbonization has affected and will inevitably affect both residential and commercial Chinese customers and electricity companies thus need to make changes in their marketing to promote a low-carbon future. While industrial users consume the most electricity and thus “consume” the most carbon in China, the majority of people nationwide are still residential users of electricity and thus also contribute to carbon consumption rates. In turn, electricity companies have a vital role to play in marketing low-carbon electricity generation to both residential and industrial users to reduce nationwide carbon consumption. Importantly, electricity marketing needs to address the negative financial effects of decarbonization for energy users and to promote renewable energy sources. After all, industrial users have been negatively affected by the abolition of the time-of-use tariff, which will result in a significant increase in operating costs. Meanwhile, data on the use of renewable energy sources by industrial users in China is not widely available, so power companies could implement lean marketing to promote the use of renewable energy sources in the industry [3].

Recent research has created methods to evaluate various indexes of power use in the power industry. The relevant research mainly focuses on how to establish an index system for electricity users and how to assign weights to the individual index systems that can be used in an evaluation of the significance of these indexes of data. Ge et al. [4] proposed three types of power user assessment (PUE) indexes that factor in data on energy efficiency, security monitoring, and demand response. Considering the uncertainty of user data and expert scores, Ge et al. contended using an interval analytic hierarchy process (IAHP) combined with interval entropy to evaluate the importance of indexed data of electricity users. Using indexed data on customers’ economic losses caused by power outages, the economic losses of power supply enterprises, and the impact of power supply enterprises’ supply reliability assessment indexes, Lin et al. [5] used the entropy value method to realize the quantitative calculation of complex indexes. They employed the hierarchical analysis method to set different index weights for different assessment subjects, which achieved a multi-faceted and comprehensive assessment of indexed data.

To assign weights to a given index system, Bai et al. [6]. proposed a scientific and a systematic method based on a combination of hierarchical analysis, entropy power method, and ensemble pair analysis. This was used to evaluate the significance of index systems of user-side integrated energy system planning schemes in which indexes were based on energy efficiency data, cost-effectiveness, energy supply quality data, and environmental protection factors. Zeng et al. [7] designed and refined a software-defined networking (SDN) environmental profit assessment index system, that indexed data on energy consumption, pollution emissions, resource utilization, and ecological coordination. They based their design on the whole life cycle theory, and they analyzed the formation mechanism of SDN cleanliness. Finally, they proposed the equilibrium-principal component two-stage combined evaluation model for the comprehensive evaluation of the positive environmental impacts of the program according to set environmental evaluation characteristics.

The above research continued long-standing work on index construction and analysis, aiming to index various data from electricity users, mainly focusing on customer credit data and low-carbon and energy efficiency data. In addition to these traditional indexes of data, renewable energy consumption factors ought to be a concern for indexes of power use in the context of global decarbonization. This paper carried out an evaluation of power users’ indexed data to implement pro-decarbonization lean marketing. To the traditional indexes of data on carbon emissions, energy consumption, and customer credit, the renewable energy index was added given the importance of meeting the global carbon emissions target. These indexes will create the first-level index layer of the comprehensive evaluation system. By applying the comprehensive evaluation system to these indexes, marketers can identify optimal lean marketing strategies to implement for decarbonization. In the assignment of index weights, the combination of subjective and objective methods was selected for weight assignment to overcome the limitation of traditional single assignment, which ensures the use of both subjective and objective measures in the evaluation of power user data indexes.

After a comparison of various combinations of assignment methods, the Decision-Making Trial and Evaluation Laboratory and Criteria Importance Though Intercriteria Correlation (DEMATEL-CRITIC) method was finally selected for weight assignment. Instead of the traditional TOPSIS evaluation method, the VIKOR method was chosen to evaluate the index evaluation system to reduce the influence of mutual compromise among the indexes. As an example, the comprehensive evaluation system was applied to indexes of power users’ data in a region of Northeast China to demonstrate how lean marketing could be implemented in commercial scenarios. By developing this new evaluation system and giving an example of a practical application, this paper provides new ideas for power consumer marketing programs to achieve global decarbonization goals.

2. Selection of Key Indexes for the Comprehensive Evaluation System

The selection of key indexes of user data to which the comprehensive evaluation system can be applied is based on the major strategic goal of power industry marketing services: global decarbonization. Key indexes are thus selected that collate renewable energy data, carbon emission reduction data, energy efficiency data, and credit factors.

2.1. Selection of Renewable Energy Consumption Indexes

In addition to purchasing electricity from electricity sales companies, various users of power also have their ways to generate electricity, such as solar energy, wind energy, and other equipment. This represents only a small number of users; most still purchase electricity generated by traditional thermal power generation for life and for operation. From a decarbonization perspective, however, it is important to include the user’s renewable energy production and usage in the lean marketing system for electricity users as the capability to produce renewable energy will inevitably reduce energy use from the grid and reduce carbon emissions.

Whether electricity users can make use of renewable energy sources depends on equipment and rates of usage. As renewable energy is not widely available in China, residential users generally use small household appliances such as solar water heaters, while businesses and industrial users use some wind power and photovoltaic equipment. All users are prone to be influenced by weather in the use of renewable energy devices, so the usage of renewable energy by users is the key index to judge whether users are helping decarbonization (Figure 1) [8,9].

2.2. Selection of Carbon Emission Index

Given the need for global decarbonization, the carbon emissions of electricity consumers have become an important evaluation index for monitoring the likelihood of meeting decarbonization goals. The most important factor to evaluate when trying to ascertain this likelihood is carbon emissions resulting from industrial production.

Residential users will use firewood to heat hot water in their lives [10], which will generate more carbon emissions than electric water heaters. The focus of industrial and business users is on production. Coal consumed in this process and the usage of alternative low-carbon technologies are important indexes to consider when evaluating the consumption behavior of industrial electricity users (Figure 2) [11,12].

2.3. Selection of Energy Consumption Index

The efficiency of energy consumption is dependent on the user equipment. Energy-saving technologies are widely popular among residential users, mainly in the form of various iterations of energy-efficient appliances. In industrial production, it is difficult to find a balance between energy efficiency and production capacity. Therefore, in lean marketing for electricity customers, the focus will be on addressing the energy consumption of industrial customers.

Energy efficiency indicators have been selected in the knowledge that residential, industrial, and business users are mainly concerned with energy conservation. Residential users are concerned with the purchase of energy-saving commodities, while industrial and business users are more concerned with the application of energy-saving technologies in the production of products, and whether they can reduce the energy consumption required to produce a unit of product. Some users, for example, have implemented energy storage equipment and other energy reuse facilities for this purpose. For example, in the mining industry, the use of integrated methods for automated mining also requires certain planning for the energy consumed [13]. The popularity of energy consumption is also one of the criteria to measure whether users reduce their energy efficiency (Figure 3) [14,15].

2.4. Selection of User Credit Index

In marketing services, the user’s credit is one of the key data points of the credit index in Figure 4. Ensuring prudent financial management across the industry isn’t the only important part of good marketing services, however, as is the case for other parts of the power industry, reducing users’ electricity consumption should be a key part of a decarbonizing marketing strategy.

Credit evaluation is mainly for industrial and business customers, and residential customers generally rarely default. Business and industrial customers with long payment cycles and large financial obligations have a much higher chance of default. In addition to defaulting on payments, electricity users can commit electricity theft. Theft of electricity, a violation of the law, is an important factor in the credit index as it affects the user’s credit, and thus theft significantly affects the lean marketing strategies used on electricity consumers [16,17].

The integration of the four constructed sub-indexes forms a comprehensive basis for an evaluation system of user data for lean marketing in

Table 1. Comprehensive evaluation index system of user lean marketing.

Criterion Layer	Index Layer	Explanation of Indexes	Index Number	The Direction of an Index Value
Renewable energy index (A)	Percentage of renewable energy access	Share of renewable energy use out of all energy use by customers	A1	+
	Frequency of use of renewable energy devices	Percentage of renewable energy equipment use out of all equipment use	A2	+
	Renewable Energy Insufficiency	Whether renewable energy generation is sufficient to meet energy demand	A3	−
	Renewable Energy Installed (Penetration Rate)	The proportion of renewable energy installations out of the total installed capacity	A4	+
	Percentage of renewable energy power	The proportion of renewable energy generation out of the overall electricity consumption of customers	A5	+
Carbon emission index (B)	Carbon emissions per unit of production	The amount of carbon emitted during the production of a unit of product	B1	−
	Low-carbon technology input rate	The total share of input of low-carbon technologies in production technologies	B2	+
	Coal savings	The total amount of coal saved in the production process	B3	+
	CO2 emission reduction rate	Percentage of CO2 emissions of users in the month compared to the previous month	B4	+
	Share of coal consumption in total energy consumption	Share of coal use in all energy consumption	B5	−
Energy efficiency index (C)	Energy consumption per unit of product produced	The amount of energy consumed to produce a unit of product	C1	−
	Energy-saving technology input rate	The proportion of energy-saving technologies in the production process	C2	+
	Decrease rate of energy consumption per unit product	The proportion of energy consumed per unit product compared with the previous production	C3	+
	Energy reuse rate	The proportion of energy used for production and reuse in total energy consumption	C4	+
Credit index (D)	Amount of liquidated damages	The financial cost of untimely payment of electricity bills annually	D1	−
	The number of payments that give rise to a breach of contract	Number of times a customer fails to pay the annual electricity bill on time	D2	−
	Amount of electricity theft from customers	The amount payable by the user for annual electricity theft	D3	−
	Number of electricity thefts by customers	Number of electricity thefts by users in a year	D4	−
	Cumulative number of unpaid power outages	Number of annual power outages due to non-payment by customers	D5	−

. On this basis, it is necessary to come to a qualitative judgment on the indexes for the next weight calculation and evaluation.

3. Construction of Evaluation Model of Power Users for Lean Marketing

3.1. Flow Chart of Index Weights

The Decision-Making Trial and Evaluation Laboratory and Criteria Importance Though Intercriteria Correlation method (DEMATEL-CRITIC) is used for weighting the evaluation of indexes of user data, and the flow chart is shown in Figure 5.

The index weight is obtained from subjective and objective aspects. The DEMATEL method was used for subjective weights and the CRITIC method was used for objective weights, and then the subjective and the objective weights of the indexes were combined by the least-squares method. Finally, the profit ratio values of each scenario were calculated and ranked based on the VIKOR method.

3.2. Subjective Weight Assignment of Evaluation Indexes Based on the DEMATEL Method

Decision-Making Trial and Evaluation Laboratory (DEMATEL) is a method for complex system factor analysis. This method is a thinking tool based on graph theory and matrix theory to analyze the intrinsic correlations between factors of complex systems. With its universality and simplicity of mechanism, it has received a lot of scholarly attention in recent years, especially in the study of the interaction between factors of complex systems and the determination of index weights. The steps for determining the subjective weights of indexes based on the DEMATEL method are as follows [18,19]:

(1) The first calculation is done by arithmetic means of expert scoring (see

Table 2. Expert commentary on semantic scales.

Semantic Variables	No Influence	Low Influence	Moderate Influence	High Influence
Scale	0	1	2	3

for specific scoring criteria). The influence degree between each index is calculated to construct the influence matrix X, and the expression is

$$X=\left[\begin{array}{cccc}{x}_{11}& x_{12}& \dots & x_{1n}\\ x_{21}& x_{22}& \dots & x_{2n}\\ \dots & \dots & \ddots & ⋮\\ x_{n1}& x_{n2}& \dots & x_{nn}\end{array}\right]$$

(1)

where x_ij indicates the direct influence of the index X_i on the index X_j.

(2) The influence matrix X is then normalized, and the direct influence matrix after normalization is noted as Y, and the expression is

$$Y=\frac{X}{\underset{1<i<n}{\max }{\displaystyle {\sum }_{j=1}^n{x}_{ij}}}$$

(2)

(3) After obtaining the direct influence matrix Y, the next step is to calculate the integrated influence matrix G, with the expression

$$G={(g_{ij})}_{n\times n}=\underset{h\to \infty }{\lim }(Y^1+Y^2+\dots +Y^h)$$

(3)

(4) After obtaining the integrated influence matrix, the influence degree M_i and the influenced degree F_j among indexes are calculated. The expressions are

$$M_i={\left[m_i\right]}_{n\times 1}={\left[{\displaystyle {\sum }_{i=1}^n{g}_{ij}}\right]}_{n\times 1}$$

(4)

$$F_j={[ f_j ]}_{n\times 1}={[ f_j ]}_{1\times n}^{\prime }={\left[ {\displaystyle {\sum }_{j=1}^n{g}_{ij}} \right]}_{1\times n}$$

(5)

(5) Let i = j, the influence degree and being influenced degree of the index X_i can be added to obtain the centrality degree R_i, and the reason degree N_i can be obtained by subtraction of them, the expressions are

$$R_i={\left[r_i\right]}_{n\times 1}={\left[m_i+f_i\right]}_{n\times 1}$$

(6)

$$N_i={\left[n_i\right]}_{n\times 1}={\left[m_i-f_i\right]}_{n\times 1}$$

(7)

(6) After obtaining the centrality degree and reason degree, the index is calculated based on the subjective weight of the DEMATEL method, and the formula is

$${\overline{\alpha }}_i=\frac{\sqrt{m_i{}^2+n_i{}^2}}{{\displaystyle {\sum }_{i=1}^n\sqrt{m_i{}^2+n_i{}^2}}}$$

(8)

3.3. Objective Weighting of Evaluation Indexes Based on the CRITIC Method

Criteria Importance Though Intercriteria Correlation (CRITIC) is an objective weighting method based on data characteristics, which essentially determines objective weights based on the amount of information in the data. The method is mainly measured in two aspects: one is to express the contrast intensity by the standard deviation; the other is to express the conflict between indexes by the correlation coefficient. Finally, the multiplication of the contrast intensity and conflict is used to express the information of the indexes. By this method, weights are assigned to the lean marketing evaluation indexes of electricity users, accounting for both the variability of the same index among different users and the coupling relationship between the indexes [20].

Suppose an evaluation matrix B is formed by m indexes of n electricity users, and the expression is

$$B=\left[\begin{array}{cccc}{b}_{11}& b_{12}& \dots & b_{1m}\\ b_{21}& b_{22}& \dots & b_{2m}\\ \dots & \dots & \ddots & ⋮\\ b_{n1}& b_{n2}& \dots & b_{nm}\end{array}\right]$$

(9)

The objective weighting of indexes based on the criteria CRITIC method is determined as follows:

(1) Because of its direction, the index is co-directed before the objective weight is obtained. Co-direction of indexes is the conversion of negative indexes into positive indexes. The conversion formula is

$$b_{ij}^{\prime }=\frac{1}{\sigma +\max (B_i)+b_{ij}}$$

(10)

where max(B_i) represents the maximum value of the index i among all electricity users. $\sigma$ represents the coordination coefficient, generally taking a value of 0.1. $b_{ij}^{\prime }$ represents the element $B^{\prime }$ in a forward transformation.

(2) After the indexes are processed in the same direction, the index data are normalized due to the sizes, implications, and dimensions of the elements in the evaluation matrix $B^{\prime }$ being affected by the index ranking. The normalized formula for the indexes is

$$b_{ij}^{″}=\frac{b_{ij}^{\prime }}{\sqrt{{\displaystyle \sum _{j=1}^m{(b_{ij}^{\prime })}^2}}},i=1,2,\dots ,n$$

(11)

The standard matrix $B^{″}={\left[b_{ij}^{″}\right]}_{n\times m}$ can be obtained after the normalization of indexes.

(3) After obtaining the standard matrix, the standard deviation, $S_i$ as well as the correlation coefficient ${\rho }_{qr}$, is calculated and the expressions are shown as

$$s_i=\sqrt{\frac{1}{m}{\displaystyle {\sum }_{j=1}^m{(b_{ij}^{″}-{\overline{b}}_{ij}^{″})}^2}},i=1,2,\dots ,n$$

(12)

$${\rho }_{qr}=\mathrm{cov}(B_q^{″},B_r^{″})/(s_q{s}_r),q,r=1,2,\dots ,n$$

(13)

where ${\overline{b}}_{ij}^{″}$ is the mean of row i in the matrix $B^{″}$, specifically the mean of the index i; $\mathrm{cov}(B_q^{″},B_r^{″})$ represents the covariance of the rows q and r in the standard matrix $B^{″}$.

(4) After the standard deviation and the correlation coefficient are obtained, the information quantity of each index is calculated, and the information quantity calculation formula is as follows.

$$G_i=s_i{\displaystyle {\sum }_{r=1}^n(1-{\rho }_{qr})},r=1,2,\dots ,n$$

(14)

(5) The objective weight of each index is calculated by the quantity of information. The formula for calculating the objective weight is as follows

$${\beta }_i=\frac{G_i}{{\displaystyle {\sum }_{r=1}^n{G}_r}},r=1,2,\dots ,n$$

(15)

where ${\displaystyle {\sum }_{r=1}^n{G}_r}$ indicates the total quantity of information of n indexes.

3.4. Combined Weight Calculation Based on the Least-Squares Method

The DEMATEL method is somewhat subjective, and the CRITIC method is more objective and cannot be subjectively biased according to the actual situation. To ensure the integrity of the index weight assignment, the combination of subjective weight and objective weight of the index by the least-squares method is selected.

If the comprehensive weight is ${\omega }_j$, the evaluation value of i power users evaluated is as follows

$$f_i(\omega )={\displaystyle {\sum }_{j=1}^m{\omega }_j\cdot b_{ij}^{″}},i=1,2,\dots ,n,j=1,2,\dots ,m$$

(16)

The smaller the deviation between the evaluation value (corresponding to the combined weight) and the evaluation value (corresponding to the subjective weight and the objective weight), the better. Therefore, the formula for calculating combination weights ${\omega }_j$ by the least-squares method is

$$\left\{\begin{array}{c}\min F={\displaystyle \sum _{i=1}^n{\displaystyle \sum _{j=1}^m{\left[({\omega }_j-{\alpha }_j)b_{ij}^{″}\right]}^2+{\left[({\omega }_j-{\beta }_j)b_{ij}^{″}\right]}^2}}\\ s.t.{\displaystyle {\sum }_{i=1}^n{\omega }_j=1{\omega }_j\ge 0}\end{array}\right.$$

(17)

3.5. Calculation of Project Profit Ratio Based on VIKOR Method

The VIKOR method is a compromise multi-criteria decision-making method based on the ideal point solution proposed by Opricovic et al. The basic principle of this method is to determine the positive and the negative ideal solutions under different criteria and then rank the solutions according to the closeness of the evaluation value of each solution to the positive ideal solution under the condition of acceptable advantage and the stability of the decision process [21,22]. The steps for assembling evaluation information based on the VIKOR method are as follows:

(1) Calculate the positive ideal solution $L^{+}$ and the negative ideal solution $L^{-}$ with the expressions

$$\left\{\begin{array}{c}{L}^{+}=(\max (B_1^{″}),\max (B_2^{″}),\dots ,\max (B_j^{″}),\dots ,\max (B_m^{″}))\\ L^{-}=(\min (B_1^{″}),\min (B_2^{″}),\dots ,\min (B_j^{″}),\dots ,\min (B_m^{″}))\end{array}\right.$$

(18)

In the formula, $\max (B_j^{″})$ represents the maximum of value index j; $\min (B_j^{″})$ represents the minimum value of index j; m is the number of total indexes.

(2) Calculate the group profit value $C_i$ and the individual regret value $D_i$ of the evaluated users, and the expressions show as follows

$$\left\{\begin{array}{c}{C}_i={\displaystyle \sum _{j=1}^m\frac{{\omega }_j\left(\max (B_j^{″})-b_{ij}^{″}\right)}{(\max (B_j^{″})-\min (B_j^{″}))}}\\ D_i=\underset{j}{\max }\left(\frac{{\omega }_j\left(\max (B_j^{″})-b_{ij}^{″}\right)}{(\max (B_j^{″})-\min (B_j^{″}))}\right)\end{array}\right.$$

(19)

where $b_{ij}^{″}$ represents the value of the index j of the evaluated power user i.

(3) Calculate the value of the profit ratio $Q_i$ of the evaluated electricity consumers with the expression

$$Q_i=v\frac{\left(C_i-\underset{i}{\min }{C}_i\right)}{\left(\underset{i}{\max }{C}_i-\underset{i}{\min }{C}_i\right)}+(1-v)\frac{\left(D_i-\underset{i}{\min }{D}_i\right)}{\left(\underset{i}{\max }{D}_i-\underset{i}{\min }{D}_i\right)}$$

(20)

where $v$ represents the coefficient of the decision mechanism, here taken as 0.5, which is taken to maximize group utility as well as minimize negative effects.

(4) Rank the evaluated power users by the values of $Q_i$, $C_i$ and $D_i$. When the following two conditions are qualified, the ranking can be done according to the value $Q_i$. If the value $Q_i$ becomes smaller, the evaluated power user i is ranked higher and higher. The evaluation conditions qualified are as follows

(1) $Q(l^2)-Q(l^1)\ge \frac{1}{(n-1)}$

Where $l^1,l^2$ represents the best power users in the ranking as well as the second-best power users, respectively, and n represents all the evaluated power users.

(2) $l^1$ is the front-ranked evaluation object of C or D.

If condition 1 is not qualified, then the compromise solution is $l^1,l^2,\dots ,l^n$, where $l^n$ fits the condition $Q(l^n)-Q(l^1)\ge \frac{1}{(n-1)}$.

If condition 2 is not qualified, then the compromise solution is $l^1,l^2$.

4. Example and Application Analysis

4.1. Construction of a Comprehensive Evaluation System for Lean Marketing of Electricity

For example, application of the constructed lean marketing evaluation system for indexes of electricity users’ data, the data of electricity users in a certain area of Northeast China during 2018–2020 were selected according to the index system outlined above.

Evaluation of customers in the region was carried out through the comprehensive evaluation system given above. This will direct lean marketing strategies.

4.1.1. The Value of the Subjective Weight of the Index

The subjective weights of indexes are calculated by the DEMATEL method. Ten experts in the fields of electric power, electric energy, electricity marketing, and electric power credit were invited to score according to the scale given in Table 2, through e-mail, interviews, and questionnaires. After the average value was calculated, the initial direct influence matrix was obtained in Appendix A (

Table A1. Initial direct influence matrix.

	X1	X2	X3	X4	X5	X6	X7	X8	X9	X10	X11	X12	X13	X14	X15	X16	X17	X18	X19
X1	0.0	1.2	2.6	2.8	3.0	1.5	0.0	0.5	0.8	1.2	1.2	0.2	1.2	1.6	2.8	3.0	2.4	1.4	0.4
X2	1.4	0.0	2.3	2.2	2.8	0.6	0.0	0.2	0.2	0.2	0.2	0.0	0.8	1.4	1.0	1.0	1.3	1.5	0.0
X3	3.0	2.6	0.0	2.8	3.0	2.4	0.4	0.4	0.4	1.1	0.0	0.0	2.8	2.6	2.2	2.3	0.5	1.4	1.5
X4	3.0	2.4	3.0	0.0	3.0	2.4	0.6	1.7	1.4	1.4	1.4	0.0	2.4	2.9	2.4	2.2	1.5	0.5	2.2
X5	3.0	2.8	3.0	3.0	0.0	2.5	0.4	0.6	0.6	1.6	1.8	0.0	2.8	2.8	2.7	2.4	0.5	1.5	2.0
X6	1.0	1.2	3.0	2.5	3.0	0.0	0.0	1.0	0.0	0.5	0.8	0.0	0.0	1.2	0.4	0.8	1.5	2.0	2.5
X7	0.3	0.0	0.4	0.4	0.8	0.4	0.0	0.2	0.0	0.6	0.8	0.0	0.0	0.2	0.0	0.2	2.7	1.0	1.7
X8	0.0	0.2	2.0	2.3	2.3	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.1	0.0	2.0	2.0	2.2	2.0
X9	1.8	0.3	0.6	1.0	1.0	0.0	0.0	0.0	0.0	0.2	0.2	3.0	0.0	0.8	0.0	0.8	1.0	1.5	0.0
X10	1.7	0.8	2.0	2.2	2.4	1.2	0.0	0.2	0.0	0.0	0.0	0.0	0.0	1.1	1.4	1.7	0.0	2.0	3.0
X11	1.8	0.9	2.2	2.2	2.7	1.7	0.0	0.2	0.0	0.0	0.0	0.0	0.0	1.2	1.0	0.3	1.0	3.0	1.5
X12	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	2.8	0.0	0.0	0.0	0.0	1.4	0.1	0.1	2.0	0.5	1.0
X13	2.2	1.0	2.2	2.2	2.2	1.2	0.2	0.0	0.0	0.0	0.0	0.0	0.0	0.8	0.0	0.2	1.0	0.0	2.0
X14	2.7	1.9	2.6	2.8	3.0	2.0	0.4	0.6	0.8	2.4	2.5	0.0	1.9	0.0	1.7	1.7	3.0	2.0	0.5
X15	2.0	1.6	3.0	3.0	3.0	1.0	0.2	0.4	0.0	2.2	2.2	0.0	1.8	1.4	0.0	1.6	2.0	3.0	0.5
X16	2.2	1.4	2.6	3.0	3.0	1.4	0.6	0.8	0.8	2.0	2.0	0.5	1.8	1.2	2.0	0.0	1.2	2.0	1.0
X17	1.5	0.0	0.5	2.0	0.0	2.0	0.0	0.0	1.0	2.0	0.0	1.5	2.0	0.0	0.5	2.5	0.0	2.0	3.0
X18	2.0	1.0	2.0	1.0	0.0	1.0	2.0	0.5	2.0	0.0	2.0	0.0	3.0	1.6	2.0	1.0	1.8	0.0	1.0
X19	1.0	2.0	0.8	0.0	0.0	2.0	1.5	3.0	2.0	1.6	1.6	1.0	2.2	1.0	2.0	0.0	1.6	2.0	0.0

). Then, the comprehensive influence matrix was calculated, and the influence degree, being influenced degree, centrality degree, and reason degree of the index were obtained in

Table 3. Influence degree, being influenced degree, centrality degree, and reason degree.

Index	Influence Degree	Being Influenced Degree	The Centrality Degree	The Reason Degree
X1	2.908	3.173	6.081	−0.265
X2	1.898	2.398	4.296	−0.501
X3	3.257	3.663	6.919	−0.406
X4	3.462	3.718	7.181	−0.256
X5	3.505	3.981	7.486	−0.476
X6	2.011	2.357	4.368	−0.347
X7	0.522	0.381	0.904	0.141
X8	1.260	0.884	2.144	0.376
X9	1.038	0.816	1.854	0.222
X10	1.950	1.869	3.819	0.081
X11	2.070	1.782	3.852	0.288
X12	0.408	0.254	0.662	0.154
X13	1.610	2.130	3.740	−0.520
X14	3.171	2.576	5.747	0.595
X15	2.863	2.426	5.289	0.437
X16	2.954	2.474	5.428	0.480
X17	1.578	1.526	3.104	0.052
X18	3.426	3.256	6.682	0.170
X19	3.976	3.692	7.668	0.284

.

The subjective weight of the index is further calculated from the data in Table 3. The subjective weights of the indexes are obtained based on the DEMATEL method in

Table 4. Influence degree, being influenced degree, centrality degree, and reason degree.

Index	Subjective Weights	Objective Weights	Combined Weights
X1	0.0696	0.0314	0.0503
X2	0.0494	0.0121	0.0335
X3	0.0792	0.0399	0.0585
X4	0.0821	0.0940	0.0823
X5	0.0857	0.0494	0.0652
X6	0.0501	0.0831	0.0639
X7	0.0105	0.1132	0.0749
X8	0.0249	0.0438	0.0332
X9	0.0213	0.0379	0.0287
X10	0.0437	0.0086	0.0293
X11	0.0441	0.0054	0.0293
X12	0.0078	0.1016	0.0672
X13	0.0432	0.0173	0.0306
X14	0.0660	0.0857	0.0713
X15	0.0607	0.0627	0.0575
X16	0.0623	0.0524	0.0537
X17	0.0355	0.0417	0.0361
X18	0.0764	0.0667	0.0668
X19	0.0877	0.0533	0.0677

.

4.1.2. Objective Weighting of Indexes

The objective weights of the sample are calculated by referring to Section 2.3 for calculating the objective weights of indexes through the CRITIC method. The objective weight of the index is based on the criteria method in Table 4.

4.1.3. Weighting of Index Combinations

According to Section 2.4, the combined weights of indexes are calculated by the least-squares method in Table 4.

From the results of the combined weights, the top five indicators in the entire index system are the renewable energy installed (penetration rate), low-carbon technology input rate, energy reuse rate, cumulative number of unpaid power outages, and energy-saving technology input rate. The entire index system is based on both subjective and objective aspects, which means that both the expert’s score and the actual electricity consumption of users indicate that there is insufficient popularization of low-carbon equipment and energy-saving equipment among electricity users in Northeast China. Energy efficiency is also a problem for electricity users. Regarding the issue of reducing carbon emissions, electricity sales companies can provide certain guidance to electricity users in the above-mentioned aspects during lean marketing. This goes a long way toward reducing overall carbon emissions.

4.1.4. Results and Analysis

The combined weight value of the index can be calculated based on the VIKOR method. Because condition 1 and condition 2 in Section 3.5 are qualified, the ranking is just sorted by the profit ratio values in Figure 6.

Data in Figure 6 show the ranking of the profit ratio values for electricity users. The ranking is based on the analysis of five representative electricity users among residents, businesses, and industries. From the ranking of the profit ratio values, the ratings of residential customers are generally higher in the comprehensive evaluation index system of electricity users constructed for the purpose of targeting carbon emission reductions. An initial reason may be that these residents are generally using few high energy-consuming appliances, and they have embraced renewable energy technologies (such as the use of solar water heaters, etc.). The second reason is that the businesses users selected have not been using or generating renewable energy due to the influence of business hours, the long usage time, high seasonal variations, and a need for extensive summer cooling and winter heating. All of these factors lead to high energy consumption. Industrial users likewise have high energy-consumption rates; and, while they were previously on a preferential time-of-use tariff so that operating costs remained stable, the abolition of this tariff means that there is no discount fee for industrial users. Restricted by scale, capital, and other issues, these industrial users are clearly not making the transition to renewable or low-carbon energy sources, making rates of default, electricity theft, and carbon consumption much higher.

Electricity sales companies can customize different electricity sales packages for different electricity users according to the ranking of the benefit ratio value.

For residential users, the problem of carbon emissions in life is not serious. At the same time, because the cost of renewable energy equipment is not affordable for ordinary residents, when electricity sales companies formulate electricity sales packages, the primary consideration is energy conservation and utilization. According to the resident user who does not work at home on weekdays but still operates electrical appliances and other equipment, it can be considered to formulate different electricity prices according to the overall load value of the user.

For business users and industrial users, a certain amount of carbon emissions will be generated in both production and operation. Industrial users are limited by equipment and other problems, which will lead to low energy utilization. The lack of low-carbon equipment will also lead to a large number of carbon emissions. Business users will use air conditioners for heating or cooling in winter and summer to ensure customers’ feelings, which makes the operation mode of business users unhelpful in reducing carbon emissions. When electricity sales companies customize electricity sales packages for two types of users, they can encourage business users and industrial users to use more renewable energy equipment and energy storage equipment in exchange for lower electricity prices.

4.2. Application of the Evaluation System in Business Scenarios

As an example of the application of the evaluation system of indexes, here we introduce the customized packages of electricity sales companies for electricity users and the cut-off of electricity users in the case of high grid load. The electric power company ranked customers according to the comprehensive evaluation system and formulated possible scenarios on which it could base analysis of actual business models. The scenarios feature possible customized sales packages for users and the possibility of load shedding processing when the power system is in a high load state, in this way, the stable operation of the power grid can be ensured and users with good evaluations can have power supply guaranteed.

Based on the user ranking obtained from the comprehensive evaluation system, power sales companies could customize different sales packages for different users. When deciding a customer’s access to a sales package, indexes of users’ power consumption could be considered, such as overall power consumption, power consumption habits, and power consumption trends in the population. Since complete customization would be financially and practically unfeasible for power supply companies, the sales package cannot be customized completely according to the user’s situation. Nevertheless, the comprehensive evaluation system of power use suggests that as a lean marketing practice, offering customized sales packages to customers is good practice for reducing carbon use.

The main reference in the analysis performed by the electric power company was representative power consumption habits. The habit of using electricity cannot be obtained directly from smart meters and other devices but from historical data processed and analyzed by Apriori Association analysis.

In the Apriori Association analysis—select 2019–2020 electricity user load data in Northeast China—the load data of various users are selected to create an item set, and the item set is collected 48 times a day for records. An example of an association rule is the set of items for two consecutive periods. The support and the confidence of each association rule are calculated to represent the strength of the association.

For an association rule, its support is expressed as

$$\begin{array}{l}support(A\to B)=P(A\cup B)\\ =count(A\cup B)/count(D)\end{array}$$

(21)

where D represents the data set of all items.

The confidence is expressed as

$$\begin{array}{l}confidence(A\to B)=P(B|A)\\ =support(A\cup B)/support(A)\\ =count(A\cup B)/count(A)\end{array}$$

(22)

Finally, the strong correlation data are expressed in the form of a curve, and the daily power consumption habit curve of residents, businesses, and industries is obtained in Figure 7.

Taking residential users as a first example, the main power consumption weekday peaks for residential customers are from 6 a.m. to 8 a.m., 11 a.m. to 1 p.m., and 6 p.m. to 8 p.m. According to the results of the comprehensive evaluation of lean power marketing, residential customers are generally ranked at the top, and electricity sales packages can be made for such customers in favor of residential customers based on their electricity consumption habits.

The industrial user’s main electricity consumption peaks are concentrated between 6 a.m. and 7 p.m., and because industrial users consume at high rates, they should have been given customized exclusive sales packages. But given the comprehensive evaluation results of indexes for lean marketing shown in Figure 6, the overall ranking of the listed industrial users is low, and their ratings are in the lower middle class, so it is necessary to offer packages that both conform with these usage statistics but that also encourage industrial users to seek low-carbon and low-energy consumption patterns.

In addition to the implementation of biased customization of power sales packages, the power system sheds load during peak consumption periods to maintain the balance and the stability of the overall system.

As shown in Figure 7, the high load times for all three types of users were concentrated from 11 a.m. to 1 p.m. and 6 p.m. to 7 p.m. During these two time periods, the power system is under high load. To ensure the stability of the entire power system, power sales companies must perform load shedding operations for power users. When performing load shedding, it is possible to load shed according to customer ranking as well as users’ electricity consumption habits. During periods of high load, industrial users ranked significantly lower in consumption, therefore, priority should be given to industrial users when cutting the load.

Considering the stability of the entire power system, load shedding is a method, but this method needs to consider the social necessity of users at a low ranking. Therefore, electricity sales companies can average the load of the entire power system by formulating electricity sales packages, such as encouraging industrial users to use wind energy for production at night and using solar energy to power batteries during the day. Users who adopt this method can obtain lower electricity prices.

5. Conclusions

This paper proposes to implement lean marketing with the goal of reducing carbon emissions, targeting all users of power. By including a renewable energy index among indexes of electric power users’ data such as carbon emissions, energy efficiency, and customer credit, and by evaluating these indexes through the comprehensive evaluation system outlined above, new strategies for lean marketing emerge.

The paper has shown that the DEMATEL-CRITIC method offers a more mature and comprehensive way to assign subjective and objective weights to power industry indexes. For the DEMATEL-CRITIC method, the DEMATEL method reduces the composition of system elements, and it simplifies the relationship between system elements. The CRITIC method considers the correlation of indicators when calculating the weights accounting for both the variability of the same index among different users and the coupling relationship between the indexes. Finally, the method of combining subjective and objective weights is adopted to ensure that the weights contain both subjectivity and objectivity. The disadvantage of the DEMATEL–CRITIC method is that when calculating the subjective weight, experts need to score the relationship between the indexes, so it needs to consume a certain number of human resources; and, at the same time, it must ensure the accuracy of the scoring method and the expertise of the experts in the field. For user evaluation, the VIKOR method is optimal for user evaluation decisions, because methods such as TOPSIS do not account for concessions between indexes. Compared with the TOPSIS method, the VIKOR method has one more decision-making mechanism coefficient, which can enable decision-makers to make more aggressive or conservative decisions.

Finally, the paper offered a scenario-based simulation analysis of the marketing service business model according to the evaluation results of the suggested indexes. The results show that the indexes can direct the development and the implementation of lean marketing strategies for electricity sales companies based on the goal of carbon emission reduction, especially in the customized electricity sales package for various power users—and, at the same time, a more comprehensive customized electricity sales package can be made in combination with the electricity consumption habit curve of power users. The results also show that in maintaining the stability of the power system, it is still possible to use the experimental results combined with power marketing to ensure the stable operation of the power system from another aspect.

The innovation of the paper is regional. Some European countries and America have done a very good job in dealing with carbon emissions, but the energy usage of these countries is different from that of some developing countries such as China so it is not very useful to learn from such experiences. Taking developing countries with a large population like China and relying too much on traditional fossil energy as an example, this paper proposes a lean marketing index system suitable for developing countries such as China. This index system can help to reduce carbon emissions and help the electricity sales companies to provide more reasonable advice on electricity sales.

Combined with the above representations, the limitations of the paper are also revealed. There are two aspects in total. First, the national conditions are different, and the experiment is based on China, so it is not helpful to developed countries with national conditions that are different from China, but it is of great help to developing countries with similar national conditions to China. Second, the management system is different as the China power grid is based on unified management by the government, and it is a state-owned enterprise so it will give priority to social responsibility rather than the profit of electricity sales companies when considering problems, so the experimental results may not help countries with private power companies.

For developing countries like China, the road to carbon peaking and carbon neutrality has just begun. This paper only studies a certain area in northeastern China, so the results of the research are limited by region, and there are also seasonal factors, social factors, and other factors that are not considered in the index system in the experiment. In the following experiments, more influencing factors will be considered, and the application of the indicator system will be deployed to other regions for research. It is hoped that this paper can provide help and suggestions on how developing countries can promote carbon emission reduction in electricity marketing.