The Amazon product ranking factors that move position are, in order of weight: conversion rate, keyword relevance, sales velocity, price competitiveness, availability, review quality, and delivery speed. Everything else that gets discussed is either a component of one of those or has so little effect it is not worth your attention.
I’d reduce all of it to one sentence: Amazon ranks the listing most likely to result in a sale for that specific search. Not the best product. Not the best-written listing. The one most likely to convert.
This page is the reference. If your ranking has already fallen and you need to work out why, that is a diagnostic rather than a reference question and it lives on Amazon product ranking dropped.
The Amazon product ranking factors, weighted
| Factor | Weight | What it actually responds to |
|---|---|---|
| Conversion rate | Highest | Images, price, reviews, delivery promise, listing clarity |
| Keyword relevance | High | Title, bullets, backend terms, and what buyers search before buying |
| Sales velocity | High | Units per period, weighted toward recent |
| Price competitiveness | Medium-high | Your price against the page-one median, not against cost |
| Availability | Medium-high, binary | In stock or not. There is no partial credit |
| Review count and rating | Medium | Volume builds slowly, rating moves conversion fast |
| Delivery speed | Medium | Prime eligibility, fulfillment method, regional stock placement |
| Session count | Low on its own | Only matters as the denominator of conversion |
I’d read that table as one idea rather than eight. The top three are the same thing viewed from different angles, because conversion times traffic equals velocity, and relevance decides which traffic arrives.
Conversion rate is the whole game
If you change one thing after reading this, make it this. I’ve found nothing else comes close.
Amazon can see, for every search term, how often each listing gets clicked and how often that click becomes an order. A listing that converts better than its neighbors on a term gets shown more for that term, which produces more sales, which reinforces the position. The loop compounds in both directions.
This is why a listing with fewer reviews sometimes outranks one with thousands, which in my experience is the fact sellers find hardest to believe. It converts better for that specific phrase, usually because it matches the search intent more exactly.
What actually moves conversion, in the order I’d work on it: the main image, price relative to page one, the star rating, whether the first bullet answers the buyer’s actual question, and the delivery date shown. Copy quality matters least of those, which surprises people who have spent money on copywriting.
Relevance is about the search term, not the product
Amazon needs to know which searches your listing should appear for, and it learns that from two sources: what you tell it, and what buyers do.
What you tell it is the title, the bullets, the backend search terms, and the structured attributes. The title carries the most weight. Backend terms are genuinely useful and genuinely limited, and I’ve found sellers either ignore them entirely or stuff them to no effect.
What buyers do matters more over time. If people search a phrase, click your listing, and buy, Amazon learns the association whether or not the phrase appears in your copy. That is why advertising on a keyword can eventually help organic rank for it: not because ads directly boost organic, but because the resulting purchases teach the association.
The practical rule: put your main phrases in the title and bullets, fill the backend field with the variants that do not fit naturally, and then let purchase behavior do the rest.
Velocity, and why it is recency-weighted
Sales velocity is units sold per period, and recent sales count for far more than old ones.
This is what makes ranking perishable. A listing that sold well six months ago and has been quiet since does not hold its position on history. It is also why launches need concentrated sales rather than steady trickle: the signal is the rate, not the total.
And it is why a stockout is so expensive. Velocity does not pause when you run out, it goes to zero, and the ranking follows.
Price, availability and delivery
Three factors that behave differently from the rest, and that I’ve found sellers consistently misjudge.
Price is judged against page one, not against your costs. Sitting meaningfully above the page-one median suppresses conversion, which suppresses ranking, regardless of whether your product justifies the premium. I’d check the median before setting price rather than working up from margin.
Availability is binary and unforgiving. There is no partial credit for being nearly in stock.
Delivery speed is increasingly weighted, and it is largely decided by fulfillment method and where your inventory physically sits. Prime eligibility matters. In our experience regional stock placement is the most underused lever here, because two sellers with identical listings can show different delivery dates to the same buyer.
Reviews: count and rating do different jobs
Worth separating, because they are usually discussed as one thing and in our experience they behave nothing alike.
Rating moves conversion immediately. A drop from 4.4 to 4.1 changes buyer behavior right away, and the ranking follows within days.
Count builds credibility slowly and is the hardest advantage for a competitor to copy. It is the reason review accumulation is worth pursuing even when it does nothing this month.
You cannot buy either legitimately. What you can do is use Amazon’s own review request mechanism consistently, and fix whatever the negative reviews name, which raises the rating by removing the cause rather than by burying it.
What does not matter as much as people claim
Being direct about this saves effort, so here is where I’d stop spending it.
Backend keyword stuffing. The field has a length limit and repeated or irrelevant terms do nothing. Fill it once, properly.
A+ content as a ranking factor. It improves conversion, which helps ranking indirectly, but it is not itself a ranking input. Worth doing for the conversion, not for the rank.
Posting frequency or account age. Neither is a ranking factor in the way it is for web search.
Off-Amazon traffic, usually. External traffic can help, but the effect is smaller and less reliable than it is sold as, and the traffic has to actually convert to count for anything. I’d not build a strategy on it.
Exact keyword density. This is web SEO thinking applied to a marketplace where it does not hold. Say the phrase clearly once or twice and move on.
How I use these Amazon product ranking factors
Work top down and stop when you find something broken.
Check availability first because it is binary and takes a minute. Then price against the page-one median. Then the star rating. Then conversion rate in your Business Report. Only then the listing copy and keywords, which is where most sellers start and which is usually the least broken thing.
For products in gated categories, none of this applies until you can list at all, which our guide to getting ungated on Amazon covers. And if you are still choosing what to sell, the competitive screen in finding low competition products is the upstream version of this same logic.
What I check first on a client account
When we take over an account, I go through the Amazon product ranking factors in a fixed order, and it is not the order they appear in the table.
Availability first, because it is binary and takes a minute. I pull the stockout history for the last ninety days rather than the current level, since a gap two months ago can still be suppressing a listing today.
Price second, against the page-one median rather than against cost. I’ve found this is where the largest number of quick wins sit, because sellers price from their margin requirement and never look at what the page actually costs. A product sitting 30% above the median is being punished for it whether or not the premium is deserved.
Star rating third, then the negative reviews behind it. If a rating has slipped, the reviews name the reason, and fixing the named thing raises conversion far more reliably than any listing change.
Only after those three do I open the listing itself. In our experience the copy is the least broken thing on most accounts, and it is where almost every seller starts. That mismatch is the single most common waste of effort I see in this work.
The last thing I check is advertising, and specifically whether it is buying sales on terms the listing should be earning organically. That is not a ranking factor as such. It is a symptom that one of the factors above is broken and being papered over with spend.
FAQ
What are the main Amazon product ranking factors?
In order of weight: conversion rate, keyword relevance, sales velocity, price competitiveness, availability, review count and rating, and delivery speed. All of them serve one objective, which is that Amazon shows the listing most likely to result in a sale for that particular search.
What is the single most important Amazon ranking factor?
Conversion rate. Amazon tracks how often each listing converts for each search term and shows the better converter more often, which produces more sales and reinforces the position. This is why a listing with fewer reviews can outrank one with thousands.
Do backend keywords still affect Amazon ranking?
Yes, but modestly and only when used properly. Fill the field once with genuine variants that do not fit naturally in the title or bullets. Repetition, irrelevant terms and stuffing do nothing, since the field has a length limit and Amazon ignores duplicates.
Does A+ content improve Amazon rankings?
Not directly. A+ content is not a ranking input, but it improves conversion rate, which is the strongest ranking factor, so the effect arrives indirectly. Build it for the conversion gain rather than expecting a ranking change on its own.
How much does price affect Amazon ranking?
More than most sellers expect, because price is judged against the page-one median rather than against your costs. Sitting well above that median suppresses conversion, and suppressed conversion suppresses ranking regardless of how well the premium is justified.
Does external traffic help Amazon ranking?
It can, but less reliably than it is usually sold. The traffic has to convert to count for anything, and poorly converting external traffic can hurt by diluting your conversion rate. It is not a foundation to build a ranking strategy on.
Last updated: September 12, 2026. Amazon does not publish its ranking algorithm and its behavior changes, so treat this as an informed model built from observed behavior rather than documentation. Verify report locations and current features in Seller Central. ZonHack is an Amazon Ads verified partner and an Amazon SPN Verified Partner.
