The Search Query Performance Report Data provides a wealth of information for Amazon sellers over a 2 year historical time frame. This information helps to identify the top terms associated with a seller’s account which can be used to identify potential issues in the conversion funnel where a seller has an opportunity to grow their market share for specific search terms.
When a seller has a lot of products in their catalog what is important is whether those products are very similar to each other, i.e. they sell through the same set of customer search terms or very different from each other, i.e. they sell through different types of search terms. The more “niches” a seller is selling within the more context shifting an account manager will have in dealing with their account as it takes time to become a market expert within a specific niche.
We can use the data from the search query performance report to group ASINs into clusters based on how closely similar they sell through the same customer search terms. Here is an example of what that looks like for a large account that is selling primarily in 2 different categories.
We see that their catalog of products primary falls within these two categories and therefore we could tag these groups of products by the category they should be in to get a better understanding of the category performance segmented from the overall account performance. For example if the account decreased 10% in sales, looking at this at the top level doesn’t provide us much useful information we need to dive down deeper to the level of categories. If we do this we might see that category 1 increased in sales by 10% whereas category 2 decreased in sales by 30%. This would help to direct our focus on where to dive in deeper.
The other context is how the category performed as compared with it’s peers in the category. So for example if sales increased by 10% but market share decreased by 40% that is not a good sign. The other thing to look at is how BSR has evolved for the products in the category as compared with competitor BSRs to get an idea as to how well a category is performing.
If your overall ASIN BSRs have improved while your competitors BSRs have declined this would be a sign that you are doing better. If given this your sales increased by 0% this is actually a good thing given most of your peers have seen a decline in their sales because of worsening BSRs for their ASINs so you are doing better than the market.
How well you are doing is always in relation to how others are doing you can’t assess performance in a vacuum and need to have comparables in order to have the context that lets you know how well you’re doing overall.
Finally this clustering algorithm is an easy way to get a quick understanding of a new clients catalog and if they haven’t already organize the products into categories it can help facilitate that quickly.
I am also pretty sure the clustering algorithm is how Amazon determines how to organize products on their platform and determine relevancy of products for specific search queries - it uses a combination of CTR, conversion rate, return rate and other factors to determine a relevancy score to see how close a product is to a specific keyword versus other products. I imagine it uses a vector space set up where the query is a point in 3D space and the products and placed in 3D space between a universe of queries that are positioned by how close they are to each other.
There used to be a software that visualized this called Yasiv.com but they no longer work because of an update Amazon made to the data 3rd party companies could get.