One of the most annoying things about working inside Amazon’s walled garden is the limited amount of information they release on how the platform works.
The lack of information has agencies, brands, and sellers on the platform sharing individual learnings about what’s working (and what’s not) for getting your products to rank.
There was, however, one time when Amazon actually explained how their algorithm worked… 😱
This little known video from a machine learning conference in 2016, where Daria Sorokina, a Machine learning researcher that worked on the Search Relevance team at Amazon’s secretive “A9” research arm.
Daria explains the problem of a new Harry Potter book hitting the market.
“How do we program the algorithm to ensure that when people search ‘Harry Potter,’ it gives the customer the newest book they want, rather than hundreds of Harry Potter related products that already have relevance to the keyword?
How was it solved? By assigning a relevance score to a product, based on how many customers bought that product on a given day, divided by the total number of customers that bought anything on Amazon in that category, on that day.
So when J.K Rowling released a new book, the purchases would spike relative to other products on Amazon in the book category, and that new book would be given a high relevance score to keep it there.
How can we use the relevance score to our advantage?
At Pilothouse, we’ve seen excellent results from what we call “The New Product Effect.”
Basically, the idea is to make your product look totally new to Amazon in your category, and then you launch your listing, your keywords and your ads all at once.
By going from zero to a coordinated push, you send a signal for Amazon to increase your relevance score, which means higher organic ranking for your keywords, and lower CPCs in the auction.
Start by removing all relevant keywords, so you don’t index on any until you’re ready to push.
If you’re releasing a new running shoe, and your main keyword is “running shoe,” then ensure the word “running shoe” is nowhere to be found in your title, bullets, backend search terms, or PPC campaign.
*This includes all combinations, so you need to fully omit the word ‘running.’
Then slowly add in each keyword as you’re fully ready to push into it with ad spend.
Don’t ever set your product live until you’re ready to push hard to every keyword you’re going to index for.
Finally, if you really want to be absolutely sure you’re taking advantage of this effect, don’t even build the product listing at all until the latest possible time.
By blitzing Amazon with a NEW, highly targeted, and ad spend backed listing and keywords, you’ll be more likely to get a higher relevance score, see increased scale and decreased costs because of it.
If you’re looking for an extra push on your Amazon listings, reach out to the Pilothouse Amazon team! 🚀
