Amazon's new search algorithm changes! Sellers need to pay attention to these

Recently, when I was researching products at the front desk, I unexpectedly discovered that after entering a search term in the search box, not only would associated terms pop up in the search box, but now something new appeared - peripheral product associations.

After typing "blanket", several other product types with similar attributes will be automatically displayed at the top, including "heated", "weighted", etc. Electric blankets and weighted blankets are different variants of blankets. Now it is not only possible to associate phrases with words, but also to further associate with similar products.

At first it looked like Amazon had given a new ad location (this display position was simply incredible), but after actually clicking on this icon, it simply changed the customer’s search term into the target product.

For example, if I search for blanket and then click on weighted, I will be redirected to the search results page for weighted blanket instead of a product related to the image.

After I tried a few main words, I found that most search words did not have this function yet. Only some words in certain categories would have this kind of image association. However, during the trial, I found that this new function also has another association mode, peripheral product association.

Look at the example in this picture. When searching for "knife", it does not associate different knives. Instead, the picture shows knife-related products, such as whetstones and sharpening rods.

This means that the association logic of this new set of functions has at least two modes. One is to associate some similar products, such as searching for blankets to associate blankets with various functions; the other is to associate peripheral products, such as searching for earphones to associate silicone plugs, earphone box covers and other earphone accessories.

This change in search has far-reaching implications. Originally, sellers in different categories on Amazon basically played their own games, and some operations that could not be put on the table were also internal struggles. Now this association algorithm breaks down the walls between categories. Who would have thought that the seller of gravity blankets would compete with them for traffic one day? Electric blankets.

However, category integration is also a new opportunity for sellers, which means that there are more ways to obtain traffic. If you spend time on search terms (add me to receive the basic keyword strategy skills package: zyqmda), you may even be able to snatch traffic from the neighboring category!

Combined with the fact that Amazon revealed a general trend to foreign media some time ago that it will transform the search results page into a "shopping page" within three years, I estimate that the above two association models are the cornerstones of the new shopping page.

Recently, foreign media broke the news that Amazon’s internal documents show that it is improving the search bar and search association logic, adding more personalization and associations between previous and subsequent search behaviors, and using machine learning to make the entire search process smarter.

For example, Amazon has previously added associations with famous brand words in the search box, which can directly jump to the brand flagship stores of these big brands on Amazon. This is part of the overall intelligent search association.

Now Amazon not only adds peripheral products and similar products to search terms (the two methods we mentioned above), but also adds some understanding of natural sentences. For example, if a customer searches for "things that connect bicycles", the search results will recommend bicycle racks.

In addition, the document also revealed another very important association, which is to match the customer's previous and subsequent search behaviors. For example, a customer searched for "camera" in the last search, and in his next search, as long as he types the letter "N", the algorithm will automatically recommend the camera brand "Nikon".

Srikanth Thirumalai, Amazon's vice president of search, said, "Our vision is to transform the search module from a page that simply displays search results to a page for shopping and discovering more products, making shopping easier and more enjoyable."

In other words, cross-category matching and association will be the general trend of Amazon's search algorithm in the future. For sellers, mastering the keyword matching mechanism is an important means to open up traffic entrances in the future.