The following section offers a summary of the correlations we found between each ranking factor and Amazon search engine results as well as a glossary of terms used to describe each ranking factor here and in the insights section of this report.
In the chart below, each ranking factor is assigned an R-value. A positive value indicates the relationship is positive. In other words, the higher the R-value, the more likely that products or pages fulfilling that ranking factor will rank better in search results.
A “0” R-value indicates that our data shows there is no relationship between the ranking factor and Amazon search results, and a negative R-value indicates that the relationship is negative. For more information on our data collection and analysis, check out the methods section of this report.
NOTE: Our data reveals and summarizes correlations between ranking factors and Amazon search results, but it’s important to acknowledge that correlations don’t always indicate causations. That being said, see why SEO expert Rand Fishkin believes correlations are as important as causation in SEO.
Glossary of Terms
Sales Rank – Sales rank is an ordinal ranking that Amazon displays to the public. This is, in ascending order, the best to worst sellers of each category.
A product’s sales rank is only relevant to the category (or search node) in question. Since each correlation was calculated by search query, the calculation was always contained to a single category or search node.
In other words, the top selling item in Electronics (sales rank 1) cannot be compared to the top selling product in Home & Kitchen (sales rank 1) in any meaningful way.
For products without a sales rank, we assumed a 999,999 sales rank, which meant the highest possible value was assigned to it by the Spearman correlation formula.
For more information on our Spearman calculations, check out the methods section of our report
Sold by Amazon.com – To find items sold by Amazon, we looked for when the merchant was listed as “Amazon.com.” We then assigned a binary value to calculate its Spearman correlation.
NOTE: It is important to consider that a product may be sold by Amazon one moment and a seller the next.
The “Sold by Amazon.com” value was collected as we tracked the rank of the product in the summer and fall of 2015, so the subsequent comparison is for the exact point in time that the product ranked for the search term.
Product Title – Quality of Match – The next ranking factor was a modified broad match, quality of match score. We assigned a value of one for each unique keyword matched in the product title from the search query.
For example, for the search query “tank top blue” a product containing only “tank” and “top” would receive a score of two. A product containing all three words would receive a score of three.
Prime eligibility – Based on the Prime eligibility of the item (at the time we ran the data) we assigned a value. Again, we used a binary value in our Spearman calculations.
Discount amount – To find each product’s discount amount or “percentage off,” we subtracted the selling price from the list price and then divided by the list price provided by Amazon.
In many cases a list price was not provided, so we assumed a discount amount of zero. We also found items were being sold above list price roughly 5.2% of the time. These values were assigned a negative discount.
MBM Product Title – For this match type, we assigned a binary value to a modified broad match. If the product title included any of the words in the search query, it was a match. If it included none, it was not a match.
For instance, a search for “red sweater” only required a mention of “red” or “sweater” to be considered a match.
Total BM occurrences – For this metric, we examined the product title, features, and description as one collection. We then counted the total number of mentions of any word in the search query.
In the same example for a search for “red sweater,” we found “red” was mentioned 12 times and “sweater” was mentioned five. We considered this a total of 17 BM occurrences.
Description BM occurrences – This is the same measurement as the Total BM occurrences; however, it is only applied to the description field.
EM Product Title – This is an exact match (not case sensitive) of the search query in the product title. Close variations or plurals were not considered; these matches are exact in every regard, except case.
Feature BM occurrences – This is the same measurement as the Total BM occurrences; however, it is only applied to the features field.
Total EM occurrences – This is a running total of every case insensitive exact match in the product title, description, or features fields.
Description EM occurrences - This is a running total of every case insensitive exact match in the description field.
Feature EM occurrences - This is a running total of every case insensitive exact match in the features fields.
Third Party FBA – To calculate this metric, we looked at two fields: merchant and Prime eligibility. When the merchant was not Amazon.com and the item was Prime eligible, we considered the product to be “Third Party FBA.”
It’s important to note that during our research, Amazon introduced a new “Seller Prime Fulfillment” program. This program allows a select number of qualified sellers to fulfill their orders and still be considered Prime eligible.
If a seller has not qualified for this program, the only other way for products to become Prime eligible is for merchants to sign up for Amazon’s Fulfillment by Amazon program.
Third Party Merchant Fulfilled – Similar to how we calculated the value for Third Party FBA, we looked at the merchant field and the item’s Prime eligibility for this metric.
In this case, the item was considered Third Party Merchant Fulfilled if the merchant was not “Amazon.com” and the item was not Prime eligible.
NOTE: there may be some circumstances where a seller uses FBA but his or her product(s) aren’t Prime eligible, but this is very rare.
For additional data points and observations, check out the insights section of this report.