Calculation method of user Satisfaction Index weight

User Satisfaction Survey is an important activity in user experience work. In order to understand the overall satisfaction degree, the first level index satisfaction, the two level index satisfaction, we also need to understand the weight of the next level index to the above index, help to determine the work priority of each aspect, and provide the decision basis for the http://www.aliyun.com/zixun/aggregation/11260.html "> Product Optimization improvement direction." The following is a brief introduction to several methods of calculating the weight of satisfaction Index.

Part 1. The function of weight calculation

Index weight can be more reasonable score user satisfaction, guide user experience Optimization direction.

The main functions of the satisfaction survey are:

understand the current product user satisfaction (how about the customer satisfaction of the product?) Find product satisfaction Short board (which aspect of satisfaction is most dissatisfied with the user?) To determine the direction of improvement (what are the priority improvements?) ）

The impact of different evaluation indicators on overall satisfaction is not the same, but previously we basically default different indicators of influence is the same.

In the satisfaction evaluation of the multilevel index structure, the arithmetic average method is used to calculate the first level index and the overall satisfaction is calculated using the first level index. This method of calculation is not reasonable.

When calculating user satisfaction improvement priorities:

in the absence of weight, the improved priority index of different indicators = (maximum-score)/(maximum-minimum); In the case of power weight, the improved optimization index of different indices = weight * (maximum-Score)/(maximum-minimum).

When determining the improvement priority of satisfaction Index, we not only consider the elevation space of the satisfaction Index, but also consider the index weight (i.e. influence), so the satisfaction index improves the priority more reasonable.

Part2. Weight calculation Method

Weight calculation methods are divided into 2 categories: direct weighting and indirect inference;

Direct weighting: The weight is calculated by subjective judgment of the importance of each index and factor. Grand: To calculate the weights of each index and factor by the user satisfaction survey score.

2.1 Direct Weighting 2.1.1 subjective weighting method

The most commonly used method is to use the list of indicators to evaluate the importance of each indicator, the importance of the score called the claim of importance, as a weight calculation of the data.

Use the score mean as the original relative influence factor.

2.1.2 Objective method of weighting

Direct comparison method

The index of the same set is set to "1" according to the index of the least importance, and the other indexes are compared with each other to make the judgment of the important degree of how many times, then the weights of each index are obtained.

Use the score mean as the original relative influence factor.

Sorting method

The metrics of the same set are sorted by importance.

The mean value of the sorted score was used as the original relative influence coefficient.

2.1.3 Delphi Method

With back-to-back communications, the Expert Group's projections were consulted and, after several rounds of consultation, the Expert Group's projections tended to be concentrated, finally making predictions that were in line with the future trend of the market.

The average value of the expert score was used as the original relative influence factor.

2.1.4 Analytic Hierarchy Process

Analytic Hierarchy Process (AHP) is a method of qualitative and quantitative analysis based on the analysis of the elements which are always related to decision-making, such as objectives, criteria and schemes. The Analytic Hierarchy Process (AHP) will compare the different indexes under the same set in 22.

Use analytic hierarchy process software to calculate the original relative influence coefficient.

2.2 Grand 2.2.1 Linear regression method

Linear regression is one of statistical analysis methods, which uses regression analysis in mathematical statistics to determine the quantitative relationship between two or more variables, and is widely used.

The regression coefficient is used as the original relative influence coefficient.

2.2.2 Factor Analysis

The basic goal of factor analysis is to use a few factors to describe the relationship between many indicators or factors. The next few related variables are grouped in the same class, and each class of variable becomes a factor (it is called a factor because it is not observable, that is, not a specific variable), Most of the information of the original data is reflected by a few factors.

The regression coefficient is used as the original relative influence coefficient.

2.2.3 Structural Equation

The structural equation model assumes a causal relationship in a set of potential variables, which can be represented by a set of observable variables, and is a method of establishing, estimating and verifying the causal relationship model. The model contains both observable and visible variables and potential variables that cannot be observed directly.

The factor score coefficient was used as the original relative influence coefficient.

Part3. 3.1 Weight calculation process

The weight calculation process is divided into 3 steps:

Step 1: According to the project situation, choose the appropriate method to obtain the original research value.

Step 2: Calculate the index influence coefficient according to the original research obtained.

Step 3: Return to 1 processing. The weights of different indexes are 1 under the same index set (different level index is one index set, the other two level index is one index set).

3.2 Influence System 1 Processing method

W (i) =x (i)/(X (1) +x (2) +x (3) +......+x (n))

Note: X (i) the original influence coefficient, W (i) to 1 after the weight of processing.

Part4. Concluding remarks

Due to the limited space, here is a brief introduction to the various methods and processes. This article just throw quoted Jade, need to see more information in order to understand.

For the 4 methods of analytic hierarchy process, linear regression, factor analysis and structural equation, the original satisfaction score value is collected, and the corresponding statistic software is used to calculate it. Analytic hierarchy process uses AHP related software to calculate, linear regression, factor analysis using SPSS software to calculate, structural equation need to use relevant software (recommend Amos) to calculate.

User Satisfaction Survey is an important activity in user experience work. In order to understand the overall satisfaction, the first level index satisfaction, and the two level index satisfaction, we also need to understand the weight of the next level index to the above index, help to determine the work priority of each aspect, and provide the decision basis for the product optimization and improvement direction.

Source: http://www.aliued.cn