5 hidden Amazon algorithm tips finally revealed

Amazon rank algorithm

We all know there are many ranking factors

Amazon product search results are influenced by 100+ factors. Even though only a core set have the heaviest weight on the ranking, knowing all of them gives brands an opportunity to stay ahead of the competition and innovate with their marketing. An Amazon team launching a new “feature” such as Alexa search, Amazon Augmented Reality View, Clip Coupon, etc., each becomes part of the equation of what creates a better customer experience, increases website engagement, and ultimately boosts long-term purchase conversion rates. 

Amazon algorithm similar to Google search?

What’s clear is that over the years the Amazon algorithm has been evolving. For those familiar with Google SEO, it’s not surprising the direction that Amazon has been changing. Search results are accounting for more and more factors, personalization is becoming more important, and relevancy of the search factored by the quality of content and interests of audience are paramount. The way Google is built is a function of its employees. 

As Inc Magazine discovered from a Talentful survey (image), the number of employees poached from Google by Amazon is among the highest across all of the tech giants (Microsoft, Facebook, etc.).

With that in mind, let’s take a look at some of the factors that are important for Google SEO. These insights may help us predict where the Amazon algorithm may be heading (or is already at). The analyses and surveys below from top Google experts show 2 main findings, as our own interpretations: 1) There are many dominant Google ranking factors that are already similar to the known Amazon ranking factors that all have to do something with quality of the page (content, keywords, related links and offerings) and multiple trust signals (authority, social proof, reviews). These are some of the main ones across the categories of Google’s off-site and on-site components, similar to Amazon’s account history and metrics related to brand growth. 2) The 2 red arrows below point to something else that is obvious to all of us for Google and websites overall. However, we know it makes sense for Amazon too but it has been much more ambiguous. These arrows highlight the relevancy of content and engagement. Both categories are broad but they have metrics associated with them including conversion rates, friction, usage, site behavior, hover time and obvious ones such as click-through-rates and bounce rates. 

What do all of these mentioned above have in common? Let’s put on our marketing hat and behavioral psychology glasses to answer the question… They are all symptoms of what’s going on in the life of the Amazon website visitor and if their needs are met by what they are browsing. Their condition is the level of buying intent that they have of researching or purchasing, the level of connection or disconnect to the brand/product and problem/solution they are experiencing, the depth of customer journey that brought them to the page, and overall the interests and demographics of the individual.

Ok, so what? Google has mastered this one next thing. Amazon is possibly (probably?) getting there. Amazon is either making big engineering progress of rolling it out at any point in the near future OR Amazon sellers are behind in figuring this out while the Amazon algorithm has been alive and kicking with this ranking factor (or historically we lacked a tool to get at this, hint hint).


Location. Region. States. Geo...

Amazon product rank is affected by location of where the Amazon customer is searching from. Not just the delivery address or product location. The customer geowhereabouts. Actually, it’s not the location. Remember the way-too-long-paragraph above about Google and relevancy + engagement. The interest and life of the user is what will determine if they will buy something or not. What is correlated with the traits of an individual? The place they live at and the state they live in. It’s a generalization with many exceptions but it’s still true. 

Let’s illustrate with an example. This scenario would make sense in the context of your brand too. The degree to how much it happens and how strongly affects the Amazon algorithm is a whole different question that finally we can start testing and addressing. Here’s a map of 30-day sales for one Amazon brand segmented by each U.S. state and the % of orders from each state relative to the U.S. as a whole. The obvious notes are that for any seller the most populated states will usually be the most selling states too, even though that becomes key later on it’s actually irrelevant for this part of the discussion.

If we look at 2 similar size U.S. states, in terms of population size and total number of orders coming from the state, Washington and Arizona are perfect examples for this. Maybe I’m slightly biased picking these 2 for the example because I live in WA not far from Amazon HQ and I grew up in AZ getting my Ph.D. there in the process. Now that population and sales are fairly similar for the two, we can make stronger interpretations on predicting Amazon search results.

If this were a gardening brand selling backyard and garden furniture accessories, how would the search results differ depending on where you are searching from? We have some extra insights specifically in the gardening topic from conversations happening inside one of the largest garden groups on Facebook that we own (I Love Gardening 130k+ members).

What are some of the main differences between Arizonians and Washingtonians in their interests, demographics, behaviors, lives, etc? In AZ, you probably want something that creates more shade to block out the scorching sun and materials that don’t heat up from the sun as much. In WA, you want materials that can dry quicker and the sunlight to come through but prevent the rain. WA has significantly higher wages and lower cost of living (can spend more money) than AZ shoppers. With this realistic example, you can bet yourself that Amazon searchers from AZ and WA seeing an identical search result page with 40 products will have different purchase behaviors for what products will be bought from which states. Are we on the same page about that? Google has mastered this. If you search for “carpet cleaning”, you’re more likely to see carpet cleaning companies in your city than nationwide or popular related companies that might not be in your actual city. 

Relevance is key.

So does Amazon do this? There may have been anecdotal evidence of this over the years or maybe you made an observation once or twice for your own brand. That is the question and finally we can try to address this scientifically and systematically. Well, maybe we do know the answer (actually we know now through our AMZJet software) but it might not be a simple generalization of patterns across all brands, products, audiences, and regions, so we hope we caught you attention long enough to find out yourself. If Amazon does not do it yet, at some point they have to increase their own profitability and growth even further. One way of doing that is personalization of search results even further. 


Amazon website already uses marketing platforms powered by location data.


There is software that allows geo-personalization and location-based testing that shows you website content dependent on where the website visitor is coming from. You can even do this on your own ecommerce website. I was doing this for Heroclip using VWO software when I was their Director of Ecommerce. Adobe offers this at enterprise level (check this link or if link breaks just Google for Adobe Target personalization). Well what do we have here, it looks like Amazon.com actually has the Adobe software installed on their site (or Omniture Sitecatalyst) and I’m sure the NinthDecimal tech (also on the BuiltWith list) does some geo-insight magic. It’s not surprising if there are Amazon teams thinking about some sort of geo-segmentation or have even built their own internal tech that syncs into all of the multi-variate testing and customer-centric targeted delivery of content, and more importantly directly into the Amazon A9 algorithm for search result ranking. Okay Doctor Ph.D., enough of connecting the dots and what does this mean for all of us Amazon sellers?  


Introducing: AMZJet Georanker by Zontracker

Here are the 5 tips and strategies why georank is important and how to leverage georank tracking for your product.

Have you ever ranked #1 for a keyword? Sorry to tell you but that was likely incomplete information. Do you know if you were ranked #1 in all U.S. states, or maybe it was #1 in majority of the states but #2 in the rest. This means that all #1 ranks are not created equal.

– All other tools show you just one total rank (country level or one random U.S. state). Some tools have an hourly rank tracker like Helium10 (first to launch hourly) and Viral Launch. But now with our georanker you can have extra peace of mind with more knowledge trying to diagnose what is happening to your products or competitors when rank suddenly drops, skyrockets, or fluctuates.

– Confirm if rank changes are happening in every state, multiple states, high-sales states, low-sales states, states where you have extra marketing, etc. Have more data and confidence in your decisions.

– The overall aggregate rank you’re used to seeing doesn’t mean that every single state follows the same ranking pattern. This 1st tip isn’t really a tip, it’s an opening to a whole new wild-west that you can now start exploring and testing such as with the next sections below.


Did you know that as your rank was increasing overall in your last launch, it was likely that it was actually decreasing in some states? That’s right. The rank you always saw before is just the aggregate, or the sum rank. Your dominant states overshadowed some states, but you have low-hanging fruit states that can be targeted to boost sales.

– Now you can have earlier prediction power. If your rank isn’t sticking because of many possible factors, you usually find out that ranking campaign was unsuccessful as your entire aggregate rank drops. With this new data, you might notice some states start declining sooner (red arrow in the image) than what you can normally tell from just the aggregate rank. Be more proactive to modify your strategy such as giving an extra push to your declining states.

– Have new insights to test ads for state-targeted marketing to get incremental sales amplification.



Ok, this is about the only one on the list that has not been hidden. Before AMZJet launched Georanker, this section was probably just about the only factor that has been discussed in any detail in the Amazon seller community. Make decisions where and how much inventory to send to be available for Amazon to sell.

– Everyone knows that rank is based on many factors (reviews, sales, price, conversion rate, etc.). However, regional rank is affected by factors that are rarely discussed. The two factors that are occasionally mentioned are shipping speed and inventory availability, yes these are logical factors. Geo-rank is also affected by many others such as local product market fit and regional audience fit, similar to how Google displays the search results based on many factors of behavior, history, and demographics.

– Peace-of-mind. Ever have your ranking tank or other crash happens? Not sure if it’s competitors doing something or what’s going on… Now you can have more info to diagnose problems if they are happening everywhere or only certain states.

– Want a quick win? Send way more inventory to Amazon. It will be almost guaranteed that you will get more sales. Why? Because your product may be present in more than just one warehouse so you will be closer to the shipping address location. These fluctuations in Amazon search results are frequently observed such as the changes in rank due to “Get it as soon as x” line next to products in the search results.


What if sales history and behavior from one state actually spills over to affect ranking in other states? Pinpoint U.S. states as low-hanging fruit to add consistent sales because rank was lower (easier competition), even though your total aggregate rank was seemingly increasing or is higher for some states than others.

– If you’re struggling to increase rank in hard-to-rank states, such as going from #2 to #1 (which may be driven by California or other top states), then why not allocate a much smaller part of your budget to easier-to-rank states and see incremental sales increase as you boost the remainder states into the top 10. This whole time you thought you were already ranked “#2” but that may have been averaged and rounded up because of the top states.

– Fill-in-the-gaps. Not sure how to get more sales? Try georanking in specific regions that are more likely to show movement.

– Algorithmic georanking. Just by focusing on increasing your rank in lower states (dotted purple and red lines if you left them untouched, solid lines by getting their rank up), the aggregate rank (gray line) and higher states (such as blue for California) may become easier to rank in from indirect rank increase in other states.


At the last moment we decided not to show this one yet, and because we’re good at open loops and marketing funnels 😉 It’s an extra feature that we haven’t seen anyone track data in this way or even discuss it in Amazon communities. Since the entire AMZJet is so new, all of the above should be plenty for you to start leveraging anyway. This particular section for now you’ll only be able to see it on the inside on any paid plan.


We engineered a method to check individual keyword rank but being searched from many different regional locations.


All Amazon sellers. Beginners and Advanced brands all need it for different goals. Stay ahead of the curve.


Get extra insight about your changes in rank. Leverage it for ultra-targeted advertising to incrementally boost state ranking.

If you enjoyed this article, don’t miss our upcoming content. We’ll be diving in actual data and strategies from early users.

Plus, ways you can use this for beginner or advanced brands using the FREE FOREVER or premium plans.