Discussion on the method of brand influence evaluation

Recently, the poly cost-effective in various types of media put a large number of ads, the rapid spread of the brand image of a cost-effective, in the non-Taobao users to cultivate a brand awareness, but also in Taobao users to consolidate the brand understanding, so as to improve the http://www.aliyun.com/zixun/ Aggregation/12972.html "> Brand value plays a positive role."

Before putting on the advertisement, the research team Q3 has carried on the research to the cost-effective brand influence, has focused on six main indicators:

--Familiarity: the extent to which users know and understand the brand, reflecting the performance and status of the market

-Popularity: Popular/popular brand, reflecting market performance and status

Relevance: Brand and products are suitable for the target group, is the basis of brand equity

-Uniqueness: The brand distinguishes itself from other brands, reflecting the brand's ability to outperform rivals and the future of the brand

-High quality: the quality of user perception, is the most basic element of brand longevity

--Trust degree: Whether the user trusts, reflects the brand to bring the user the benefit and the interest ability

These six indicators in the market research industry, widely used in brand research, and each indicator in one sentence, form a scale, users fill in the cost is not large. This transplant to the Internet innovation product research, basically can achieve the anticipated research goal. This paper does not make a thorough analysis of the brand influence of the poly-cost, but focuses on the analysis method.

Index dimensionality reduction

In this example, we evaluate the impact of a poly-value brand through six indicators, but we can see that there is a strong correlation between these six indicators (the correlation coefficient is more than 0.5, and significantly positive correlation), and you need to consider simplifying the metrics when conducting a comprehensive assessment.

In general, principal component analysis or factor analysis method is used. In most cases, the principal component analysis is not easy to explain, and the factor rotation can reduce the ambiguity in the main component, so that the structure of the factor load is simpler, so it is more advantageous to the explanation of the factor. In addition, the commonly used software can directly output the value of the factor score, while the value of the principal component needs to be manually recalculated.

First, the factor analysis of six indexes is made, and the KMO value of 0.891,bartlett is 2747210.409 (df=15), which is very suitable for factor analysis. The final extraction of 2 common factors, the cumulative variance contribution rate is 84.6% (see table below), the interpretation effect is very strong; the number of residual >0.05 between the original matrix and the reconstructed matrix is 26%, and the fitting effect is good.

The factor load matrix can be found that the first common factor is more focused on the use of the psychological cognition after the cost, and the second factor is more to reflect the perception of facts and phenomena, so the first common factor is named as the deep feature, the second male factor is named as the surface characteristics.

In addition, the correlation index has a large load on two common factors, in general, when considering the construction validity, it will delete the indexes with larger load on different common factors, but this example does not pursue the construction validity, but hopes to keep the index in the index system as far as possible, and to achieve the goal of reducing dimension for further analysis, so there is no deletion of correlation index.

A comprehensive score for calculating brand influence

With two common factor scores, there is still no comprehensive assessment. In order to simplify, the final use of a data to reflect the overall brand impact, needs to be further calculated.

The comprehensive score of factor analysis needs to use the characteristic root and factor score of each factor, this example calculates the equation as follows:

Comprehensive score of single sample = (characteristic root 1)/(characteristic root 1+ 2) * Common Factor 1 score + (characteristic root 2)/(characteristic root 1+ characteristic root 2) * Common Factor 2 score

Because the factor score is the standardized score, the comprehensive score of the calculation is also the standardized score, which needs to be converted into 0-100 to reflect the brand's comprehensive influence. The following should be done:

Single sample brand comprehensive impact score = (Comprehensive score of a single sample-total score minimum for all samples)/(Maximum of all sample scores-minimum) *100

In this way, we can get the comprehensive impact score of each sample on the poly-cost, and make a comparative analysis between different groups.

Comparing the difference of brand influence among groups

First, the male and female groups do a differential test. T-Test analysis of independent samples shows that there are significant differences between men and women on the brand influence of the poly-cost, and men's comprehensive evaluation of the cost is higher;

Secondly, the different city level of the group to do the difference test. Through the card-side test (22 comparison), we can know that different city level users have very significant differences in the impact of the brand, the third-tier city users of the comprehensive evaluation of the brand impact of the highest, while the first-tier city is a low overall evaluation.

After the adjustment of the business and the spread of brand advertising, I believe that the brand impact score will be significantly improved, pending research and verification.

Summary

Brand impact assessment methods, the basic operating procedures are as follows:

1. To determine the partial indicators of the assessment;

2, through factor analysis (or principal component analysis), to reduce the dimension of the sub-item index, and calculate the common factor score;

3. Calculate the synthetic score of the common factor and transform it into 0-100 intuitive score;

4, the comparison of different groups of brand impact comprehensive score.

When comparing the influence of different brands, or the different stages of the same brand, the data need to be integrated together for the above calculation process. Only in this way can the contrast be maintained at the same level.

Source Address: http://ued.taobao.com/blo......uence_method/