There are many products in the shop, it includes, but is not limited to, the following metrics: views, number of visitors, average length of stay, detail page bounce rate, next conversion rate, order to pay conversion rate, payment conversion rate, order Amount, order number of items, order number of buyers, amount paid, number of items paid, number of purchases, average visitor value, number of collectors , guest unit price, search payment conversion rate, search guide number of visitors, pay the number of buyers, so how to rank their products, draw a group top out:
For small White The best is to use the comprehensive evaluation method, for each column of indicators are:
For example, the 22 indicators mentioned above, how do we combine the normalized values of these 22 indicators now?
----------------------I'm a split line-----------------------
AHP Analysis method: Also called analytic hierarchy process. The so-called analytic Hierarchy Process (AHP) refers to a complex multi-objective decision-making problem as a system, which decomposes the target into multiple targets or criteria, then decomposes it into several levels of multi-index (or criterion, constraint), and calculates the hierarchical single sort (weighting) and the total sort by the qualitative index fuzzy quantization method, as the target A systematic approach to multi-scheme optimization decisions.
In order to optimize the weights of the comprehensive evaluation method in the final calculation, this paper uses the AHP analysis method to evaluate the weight of the product, first the indicator is listed as matrix form:
|
A1 |
A2 |
A3 |
A4 |
... |
A1 |
A11 |
A12 |
A13 |
A14 |
... |
A2 |
A21 |
A22 |
A23 |
A24 |
... |
A3 |
A31 |
A32 |
A33 |
A34 |
... |
A4 |
A41 |
A42 |
A43 |
A44 |
... |
... |
... |
... |
... |
... |
... |
Assuming that A21 represents the extent to which A2 affects A1, set the following specifications for the degree of impact:
Evaluation description |
Score |
The influence is stronger than its absolute |
9 |
Influence than its apparent strong |
7 |
Influence than its strong |
5 |
The effect is slightly stronger than |
3 |
Affect the same |
1 |
By using the method root method, 22 judgment matrix is used to find out their maximal characteristic root and eigenvector, and the method is as follows:
(1) Judge the result of each line element of the Matrix, wherein:
(2) The second root of the calculation:
(3) Normalization of vectors, that is, the characteristic vectors to be asked;
(4) The most characteristic root of the computation judgment, the first element expressed in the formula;
(5) defined as the consistency index of matrix A, in order to determine the extent of the degree of inconsistency of a, we need to find the criteria to measure the consistency of a, and introduce the consistency indicator, the values are as follows:
N |
1 |
2 |
3 |
4 |
5 |
6 |
7 |
8 |
9 |
10 |
RI |
0 |
0 |
0.58 |
0.9 |
1.12 |
1.24 |
1.32 |
1.41 |
1.49 |
1.51 |
(6) Order, called the consistency ratio, at that time, this paper believes that the judgment matrix has a high degree of satisfaction with the consistency, otherwise it needs to adjust the matrix, so that it has a satisfactory consistency.
---------------------------I'm a split line--------------------------
By using the AHP method to get the weights multiplied by the comprehensive evaluation method to get the standardized value is the final comprehensive score, can be ranked, I get the highest error 3.9% duly completed
---------------------------I'm a split line--------------------------
In fact, in a lot of standardized values can be used to do radar chart, the effect is as follows:
Obviously see the short board is the dwell time, the payment conversion rate and so on
Ranking of comprehensive evaluation of commodities