The method and practice of character classification based on user scale evaluation

Personas can be divided into a number of dimensions, such as user goals, user scenarios, user behavior, experience cycles, user value 、......, and according to the actual situation, personas may be divided by one dimension or by combining multiple dimensions.

The user goal is the most important and essential dimension of the character division in order to design the service personas.

For the Information/Information analysis object, the information/information attention can well reflect the user's use target.

Based on the questionnaire of 1688 website users on supplier information, this paper introduces the character classification method based on user scale evaluation, which is mainly divided into the following processes:

1 obtained scale evaluation result 1.1 research target

Provide design reference and basis for the organization and presentation of the supplier information of the Product Details page by understanding the customer's concern about the different information of the supplier.

1.2 Research Content

According to the present content of the supplier information, as well as the relevant information in the company files, combined with the previous relevant research results, a total of 17 supplier information of the focus of research (see the Factor Analysis results list, the factor analysis results list listed only 16, according to factor analysis theory, "safeguard" Not included in the final factor analysis).

1.3 Evaluation Scale

The user's evaluation scale of supplier information is divided into 5 levels, respectively:

1-No attention, 2-indefinable, 3-a little attention, 4-more attention, 5-very concerned.

1.4 Research Sample

This survey in 1688 site product Details page, after screening to obtain a valid sample of 3,032.

2. Get evaluation Scale Factor 2.1 project analysis

17 supplier information was analyzed separately, the results were significant, which indicated that 17 supplier information was suitable for user's attention evaluation.

(Project analysis: Each sample of 17 suppliers of the total focus on the information, the lowest total attention of 27% samples as a low group, the focus of the highest score of 27% samples as a high group.) The low group and the high group in the 17 supplier information of the attention rating of the Independent sample T test, to determine whether there is a very significant difference in the evaluation of high-low groups. If the Independent sample T test is significant, the supplier information is suitable for user evaluation. ）

2.2 Reliability Test

Reliability tests were performed using Cronbach ' s alpha, with an alpha coefficient of 0.906. When the "location" is deleted, the alpha factor becomes 0.909, which can be considered for deletion.

However, it is considered that "location" is still necessary, and the "location" is retained in factor analysis according to the actual meaning of the result of the final factor analysis.

2.3 Factor analysis of 2.3.1 significance test

The appropriate KMO value of the final factor analysis results 0.905,bartlett spherical test approximation of the 26152.146 (df=120,p value =.000), which is very suitable for factor analysis (initial KMO value, Bartlett spherical test are significant).

2.3.2 Factor Analysis Results

As the "protection" of the supplier information on the main factor of the Load = 0.482, at the same time, customer satisfaction, supply of the full load of 0.452, 0.326, "protection" This supplier information from the factor analysis of the elimination.

Considering index cluster analysis, factor eigenvalue and the practical meaning of factor, the user evaluation scale is divided into 4 factors, the interpretation rate is 64.356%, and the effect is good.

The following 4 evaluation scale factors were basic information, customer satisfaction, supply capacity, and transaction history, which means that users understand suppliers from these 4 dimensions when they view supplier information.

3. Get evaluation scale factor score

The factor analysis can directly obtain the standardized factor score, but our daily contact is the more intuitive non-standard user scale evaluation score, in the sample cluster analysis result interpretation, the result is also more intuitive. So we should use the non-standard factor score based on the user scale evaluation to carry on the sample cluster analysis.

Non-standardized factor score calculation process is as follows:

Step1: The Factor score factor (Factor Score coefficient) is based on the normalized original influence factor X (i), X (i) represents the score factor of the first supplier information on the principal factor. STEP2: Non-normalized reduction based on standardized original influence coefficients, X (i) *d (i), d (i) represents the standard deviation of the supplier information. STEP3:1 treatment. 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). The influence coefficient of 1 processing formula is as follows:

W (i) =x (i) * d (i)/(X (1) * d (1) +x (2) * d (2) +......+x (n) * d (N))

W (i) represents the weight after 1 processing, X (i) represents the first supplier information to the main factor score factor, D (i) represents the first supplier information standard deviation.

STEP4: Calculates the non-standard factor score, the formula is as follows:

F (j) = W (J1) *f (J1) + W (j2) *f (J2) + ... + w (JN) *f (JN)

F (j) represents the non-standardized factor score for the J Factor, and F (JN) represents the rating of nth supplier information under the J factor.

4. Get character Category 4.1 sample cluster

Using the Twostep cluster method of cluster analysis, cluster analysis was carried out according to 4 evaluation scale factors. 3,032 research samples were divided into 3 categories, representing 42.5%, 41.9% and 15.7% respectively.

Calculation of these 3 types of samples in the evaluation scale factor score can be found:

category 1th Users: To basic information, customer satisfaction, supply capacity, transaction history are very concerned about the 2nd category of users: The user satisfaction is very concerned about the basic information, supply capacity, transaction history of more attention to the 3rd category of users: The basic information, customer satisfaction, supply capacity, transaction history are very little attention.

5 Summary and application of user evaluation scale factor

For supplier information, users focus on 4 dimensions, namely basic information, customer satisfaction, supply capability, and transaction history.

When the page content is laid out, similar content (strong correlation) should be put together. At present, 1688 Web site product details on the page of the layout of information quotient and users concerned about the dimensions there are more inconsistencies. With the relational vendor information together, the user first looks at the vendor information of one dimension, then looks at the vendor information of another dimension, enhances the user browsing order, and improves the readability of the supplier's information content.

Character Classification

For vendor information concerns, there are 3 types of typical roles, namely:

Category 1th Users (42.5%): The basic information, customer satisfaction, supply capacity, transaction history are very concerned about

Category 2nd Users (41.9%): The user satisfaction is very concerned about the basic information, supply capacity, transaction history more attention

Category 3rd Users (accounting for 15.7%): Little attention is paid to basic information, customer satisfaction, supply capacity, and transaction history.

With such a character division, in the design, you can know how the core users, Non-core users are concerned about the supplier information and its importance, but also can conduct qualitative research to explore the different user groups on the supplier information of the root cause of the difference is what, for the design to provide decision-making basis.

the user evaluation scale factor, the character role classification provides some valuable basis for the supplier Information Block content optimization, in the actual design optimization, but also considers the different information itself the importance, the other research result, as well as the commercial angle consideration. Personas based on user-scale evaluation is just a key dimension of role division, and to create complete personas, you need to add other richer attributes and content. To the user evaluation scale, this research uses the attention degree, according to the research object's difference, the user appraisal scale may also be the satisfaction, the importance, the conformity degree and so on.

Source: http://www.aliued.cn