The parent product keyword is the primary keyword. Amazon’s backend Search Terms originally had 5 fields. Since 2018, 5 rows have been changed to 1 row, and the key length limit of each attribute has been reduced to less than 250 bytes. (not letters), special characters take up more than 1 byte. This requires sellers to select and edit keywords to a certain extent, and cannot simply pile up phrases. Generally, the reference basis for selecting parent keywords is “accuracy + traffic size”. The estimated traffic size can be viewed through Google Trends to view search popularity. Generally, the criteria for judging “good” keywords are as follows:
(1) The number of search pages is not large;
(2) The number of high-quality listings with high reviews on the home page of the search page is large (note that There are more listings with high reviews than just reviews. Please see Chapter 7 for the specific logic);
(3) The keywords themselves conform to the characteristics of the product;
(4) The brands of high-quality listings with high reviews are diversified, rather than monopolized by a single brand;
(5) The products exposed on the search results page have many of the same attributes as their own products and belong to the same category of goods.
Since (3) can be judged subjectively, a mathematical simulation can be done on the loss of traffic. The number of search pages is set as p (page), and the home page traffic is set as n (number). If the loss rate of page traffic is set to x (percent), then the estimated traffic that the product can obtain is calculated as follows:
y=n/(l+x)^p.
In the formula, x can be substituted based on experience, or precise values can be substituted through data reports or programming. Here, the traffic loss rate of 0.1 or 10% is used to calculate (in a real online shopping environment, single page traffic The churn rate is as high as 60%~70%).
If there are 10 pages of search results, assuming that the traffic on the homepage is 5,000, then the traffic on the last page will be approximately 1,900; if there are 50 pages of search results, assuming the traffic on the homepage is 10,000, then the traffic on the last page will be approximately 1,900. 85. It can be seen that the competitiveness of keywords set based on the search results of the former will be much higher than that of the latter.
From the simple simulation results using Matlab software, you can see that when the home page traffic is 5000 and the number of search results pages is 50 pages, the traffic is lost very quickly, so when setting keywords, do not set them all For broad words, a certain proportion of precise words or high-scoring words are required.