Let’s take the data from Wish backend as an example.

Under the “Performance” menu is the most commonly used data analysis section in Wish operations.

Click the “Performance” menu and we can see the following content.

Among them, product overview, sales performance, user service performance, and sales charts are more important for products and sales. Rating performance, logistics performance, refund performance, user feedback, and counterfeit rate performance are very important data indicators for operations, and user service charts facilitate data comparison.

In addition to the “Performance” menu, there are also data we need to mine in the “Product” menu.

From the intuitive background data, the following three types of data will be helpful in mastering product information and shelf information.

We can see the SKU, last update time, and upload time of the current product in our store. They are three very intuitive data, among which the last update time and upload time can be sorted by time. At the same time, we cannot ignore the other two useful information: the number of Wishes and the number of sales. The number of product sales can also be sorted in order of sales volume.

Then we saw the data performance of a single product.

The above summarizes the display methods of background data. What we need to understand next is what thinking we can use to organize and analyze data.

1. Comparative thinking

This dimension is usually used to compare the sales of a single product or a single store in a unit of time. It is usually in the form of comparison and is displayed in the form of a bar chart.

This bar chart is easy to generate in Excel. Generally, for merchants that are more standardized in back-end operations and are managed by the system, the ERP used usually has a statistical function that can count the number of orders or the amount of orders.

The statistics of sales volume and sales revenue reflect an important and basic idea in data processing, which is the idea of comparison, commonly known as comparison. A single data is an absolute number and cannot represent anything. Only by comparing it with another data can its value be reflected.

2. Split thinking

Let’s take an example. For example, a salesperson A found that the sales of the store this week were not as good as the previous week. At this time, it is not very meaningful to compare the sales alone, and the sales data needs to be further split. Some common formulas in e-commerce can be used here.

Through further analysis, we can find out the reasons for the decline in sales. We can also further subdivide the conversion rate, traffic and other data to find out the deeper reasons until we find a solution.

3. Dimension reduction thinking

When there are too many dimensions of data, we cannot analyze every dimension. For some related data indicators, we can filter out the dimensions that can be merged.

When summarizing data dimensions, not every piece of data is worth analyzing. When there is a situation where one dimension can be calculated from other dimensions such as “total transaction amount = product unit price × order”, we can reduce the dimension. Usually we only care about the data that is useful to us. When the data of certain dimensions is irrelevant to our analysis, we can filter it out to achieve the purpose of dimensionality reduction.

4. Dimension-increasing thinking

Dimension-increasing and dimension-reducing are relative, and there must be an increase if there is a decrease. When our current data dimension cannot explain our problem clearly and comprehensively, we need to do some calculations on the data to add an additional indicator. When we analyze the data of a single product, if we cannot analyze the reasons for the increase or decrease in product sales simply from the number of views, sales, refund ratio, average rating data, etc., we need to add data dimensions such as the product’s average customer price and conversion rate to assist in the analysis.