The logic of optimizing keywords from the buyer’s perspective is “whatever the buyer searches for, the keyword will be set accordingly.” The reason why this strategy can be adopted is that Amazon is an e-commerce platform, and its most original function is to facilitate transactions. Buyer searches represent demand, while seller products represent supply, and the A9 algorithm serves as an intermediate bridge to connect the two. However, there are millions of listings on the site, and even the A9 algorithm cannot guarantee the most accurate traffic for each link every day. At this time, sellers need to proactively discover customer needs and optimize listings. There are various keyword traffic analysis tools on various third-party charging platforms. Their specific function is to provide search data for each keyword on the Amazon platform, including the number of searches, number of clicks, purchase rate, click-through rate, etc.
1. Go to large-traffic orders in the category involved.
Take women’s dresses as an example. First find the most intuitive word “dress” for dress, and then use the drop-down menu to see its related matching words.
You can quickly get related words such as “pocket”, “work”, “casual”, “party”, “wedding”, etc.; then search for “dress” in the search engine, and then search for any words that appear twice or more on the homepage. The vocabulary is summarized and sorted, so that the second set of matching words “mididress” and “maxidress” can be obtained.
2. Arrange and combine the above words into related phrases.
The combination method adopts AB combination, that is, “single adjective + single noun” such as “casualmididressmaxidress for party” and so on. Note: Singular and plural changes and word morphological changes need to be taken into account when combining.
3. Analyze and compare the traffic volume between each phrase.
Because of the consistency of the purchasing habits of American buyers, you can directly use Google Trends to compare the popularity of keywords (the consistency of purchasing habits here refers to the customer’s search habits on Amazon’s search engine and their daily Search habits are basically the same).
Note: In the region column, you need to select “United States” instead of “Global”. It can be clearly seen that the search popularity of “maxidress” is much higher than that of “mididress”.
Because “maxidress” and “mididress” belong to different categories in the “dress” category and it is difficult to compare, we can also use this method to analyze the traffic of “detail words + qualitative words” Compare, for example, “sexymaxidress” to “casualmaxidress.”
It can be clearly seen from the figure that although the search popularity of “casualmaxidress” and “sexymaxidress” partially overlap, the search popularity of the former is still much higher than that of the latter, so in When a product can be described as “sexy” and “casual” at the same time, from the perspective of pure traffic address, “casual should be given priority. (Note, this refers to purely from the perspective of traffic, not comprehensive considerations, whether to choose casual in the end” is still requires a series of breaks).
Now everyone should be able to use Google Trends, a simple tool, to compare the search popularity of various keywords, right? However, judging from the popular operating experience on the market, there should be many details. If you have any questions, please use Q&A to answer them below.
Q: Through GoogleTrends, you can only compare the search popularity of keywords, but you cannot know the specific traffic of each keyword on the Amazon platform. Do you need to find some third-party software to obtain specific traffic?
p>
A: Not required. Let’s explain this problem from several points: First, from the perspective of operational experience of large enterprises, even having data sources is still far from “win every battle”. As long as a sufficient amount of money is paid, it is completely possible for large companies to find third-party data companies to update a search data of Amazon’s entire platform every week or even every day, but you see so many large cross-border e-commerce companies do not have the data even if they have it. Whatever it sells, it explodes, which shows that the value of the data itself is not so exaggerated. Second, from the perspective of data analysis, on the premise that the correlation of search popularity of each keyword is known, there is no need to discuss its specific numbers. Because from a mathematical point of view, it is completely possible to establish a standard value “t-shirt” is 1.5x, “blouse” is 1.2x, etc.). Third, the traffic size of a single keyword is far less important than we think. What is more important is its “effective traffic”.
Q: The search data that can be queried on Google Trends is the search data of Google browser. Can it be universally used on Amazon?
A: It can be universally used, and the error is completely acceptable. Because as emphasized in the first Q&A, what we want is a relative value, a reference quantity, rather than a precise absolute value, and the customer’s search habits will not change due to changes in the platform.
4. Analyze the effective traffic of each phrase.
The following conclusions can be drawn:
(1) There are two ways to reduce traffic loss, reduce the traffic loss rate/reduce the number of search result pages.
(2) It only makes sense to increase the selection of keywords with high traffic while keeping the traffic loss rate/number of search pages basically unchanged.
As for how to capture the overall data of a single day on Amazon, many third-party sales assistance software will provide part of it. You can also deduce the traffic changes of a single day by obtaining the ranking and sales changes on Amazon.
For example, assuming that our listing click-through rate is 80% of the effective exposure, the search page has 48 products, and the products are ranked in the lower half of the exposure page, then parameter b can be set to 24 (the Parameter settings can be set based on experience. If the listing is in the upper half, b can be set to 48; if it is in the lower half, 6 can be set to 24), that is, the effective exposure probability is: 24/48=50%. At this time, the click-through rate multiplied by the effective exposure rate is 40%.
In the remaining parameters, it is assumed that a certain keyword is now known, the number of searches in a single day is 200,000 (can be understood as the number of effective exposures), and through this keyword in a single day, searches are The total number of products is 500, then nqin) is equal to 500 at this time. Assume that our product has just been put on the shelves and is ranked on the last page of the exposure pages among these 500 products, then our product will be on page 11 .
Finally, multiply the calculated value by the originally obtained click-through rate value of 40%, then you can calculate that the effective traffic of the listing under this keyword is approximately 3.36.
However, we don’t need to know how much traffic there is every day, because the comparison of traffic for each keyword can be completely solved by free tools like GoogleTrends. Only when it is impossible to compare the popularity of very similar keywords Accurate traffic data is required.