Amazon SEO: how a listing gets ranked, and what a consultant actually changes

Amazon search answers one question in two halves. Is this listing relevant to the words the shopper typed, which is decided by the text Amazon has indexed from your listing fields. And does this listing convert the shoppers who see it, which is decided by your click through rate, your unit session percentage, your sales velocity, your price, your in stock rate and your reviews. Relevance gets you into the result set. Performance decides where in it you sit. That is why keyword work alone moves a listing a little and keyword work plus conversion work moves it a lot, and why a consultant who only talks about keywords is doing half the job.

This page covers Amazon's own search surface and the work order behind it. For marketplace titles elsewhere, see eBay title SEO, and for the same product on your own site, ecommerce SEO.

The two halves, side by side

Relevance inputs	Performance inputs
Title	Click through rate from the search page
Bullet points	Unit session percentage, which is Amazon's conversion rate
Product description	Sales velocity over the recent window
Backend search terms	Price against comparable listings
Category and browse node	In stock rate and Buy Box ownership
Structured attributes such as size, colour, material	Review count, star rating and return rate

The attribute fields are the half most sellers skip. Shoppers filter by size, material, fit and colour in the left rail, and a listing with those fields empty is excluded from the filtered result entirely, however good the title is.

Filling the fields that carry keywords

• Title. Brand, then the product, then the two or three attributes a shopper searches with: material, fit, size, colour, pack count. No promotional words, no all caps, no symbols. Title length limits differ by category, so take the current figure for yours from Seller Central rather than a rule of thumb.
• Bullets. Five, each leading with the benefit in three or four words then the detail. Specifications, compatibility and sizing guidance go here because they answer the questions that otherwise become returns.
• Description or A+ content. A+ content earns its place by lifting conversion, so write it to answer objections and show scale, fit and use. Keep the keyword load in the fields you know are indexed.
• Backend search terms. Synonyms, spelling variants, regional words and use cases that have no natural home in the visible copy. No competitor brand names and no repetition of words already in the title.
• Attributes. Every applicable field completed. This is the cheapest ranking work available on Amazon.

Building the backend term string without waste

Repeating a word that is already in your title wastes a byte budget that is strictly limited. This script keeps the front end and back end from overlapping:

// backend-terms.js, run: node backend-terms.js
// Confirm the current byte limit for your category in Seller Central before relying on it.
const LIMIT_BYTES = 250;

const title = 'Northvale Leather Chelsea Boots Mens Waterproof Brown';
const bullets = [
  'Full grain leather upper',
  'Waterproof membrane',
  'Rubber commando sole',
  'Wide fit available'
];

const used = new Set(
  (title + ' ' + bullets.join(' '))
    .toLowerCase()
    .replace(/[^a-z0-9 ]/g, ' ')
    .split(/\s+/)
    .filter(Boolean)
);

const candidates = [
  'ankle', 'dealer', 'slip', 'on', 'pull', 'elastic', 'gusset',
  'work', 'smart', 'casual', 'formal', 'office', 'winter',
  'tan', 'black', 'suede', 'nubuck', 'goodyear', 'welted', 'resoleable'
];

const terms = [];
let bytes = 0;

for (const word of candidates) {
  if (used.has(word)) continue;
  const cost = Buffer.byteLength(word, 'utf8') + (terms.length ? 1 : 0);
  if (bytes + cost > LIMIT_BYTES) break;
  terms.push(word);
  used.add(word);
  bytes += cost;
}

console.log(terms.join(' '));
console.log(bytes + ' bytes of ' + LIMIT_BYTES);

Build the candidate list from four sources: Amazon's own search suggestions as you type, the search terms your Sponsored Products reports show converting, the words competitor listings use that yours does not, and Brand Analytics search query performance if you are brand registered. Amazon's seller learning resources cover where each report lives, and Brand Registry is the gate on the reporting a consultant will want.

What a consultant actually changes

Deliverable	How you verify it happened
Keyword set per ASIN, with source and search volume rank	A sheet you can read, not a tool screenshot
Rewritten title, bullets and backend terms	Before and after copy, with the change date
Completed attribute fields	Spot check the filtered results for your own product
Image set rebuilt: main image on white, scale shot, detail, in use, infographic	The listing itself
Browse node correction	Category breadcrumb on the detail page
Suppressed and stranded inventory cleared	Seller Central inventory reports
Review and rating programme inside Amazon's rules	Review count trend, no incentivised reviews
Advertising restructured around converting search terms	Search term report, with wasted spend negated

Measure it with Amazon's own numbers

1. Business Reports, by ASIN. Sessions, page views, unit session percentage. Unit session percentage is the number that moves rank.
2. Search Query Performance, for brand registered sellers. Impressions, clicks and purchases by query, which tells you which words you are actually winning.
3. Sponsored Products search term report. Paid data is the fastest honest read on which queries convert, and it feeds the organic keyword set.
4. Manage Your Experiments, where eligible, for A and B tests on titles, images and A+ content. One variable at a time.
5. Rank tracking on a fixed query list, run on the same day each week, so a change has a before and an after.

Work one listing at a time, change one thing, and leave it for two full weeks. Amazon's performance signals accumulate over a window, so a listing edited three times in a fortnight has no readable result.

The off Amazon half

Shoppers research on the open web and buy on Amazon, so the brand's own product pages still matter. Mark them up properly so Google can show price and availability:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Product",
  "name": "Northvale Leather Chelsea Boots",
  "sku": "NV-CHB-BRN",
  "brand": { "@type": "Brand", "name": "Northvale" },
  "image": ["https://example.com/img/chb-brown-1200.jpg"],
  "offers": {
    "@type": "Offer",
    "url": "https://example.com/boots/chelsea-brown",
    "priceCurrency": "GBP",
    "price": "149.00",
    "availability": "https://schema.org/InStock",
    "itemCondition": "https://schema.org/NewCondition"
  }
}
</script>

Keep the price in the markup identical to the price on the page. Google Merchant Center covers the feed side of the same data.

What gets asked about Amazon SEO

Is Amazon SEO the same as Google SEO?

The disciplines rhyme and the inputs differ. Amazon ranks on listing text plus commercial performance on Amazon, with no link signals at all. Conversion rate carries far more weight than it does on Google.

How long does a listing take to move?

Plan in two week blocks, because Amazon reads performance over a trailing window. Attribute and category fixes can show faster, since they change which result sets you appear in at all.

Do backend search terms still matter?

Yes, for synonyms, misspellings and regional words that would read badly in the visible copy. The field is byte limited, so spend it on words that appear nowhere else.

What does Amazon SEO consulting usually cost?

Engagements are priced per catalogue size and per ASIN count, so ask for the deliverable list in the table above and a per ASIN figure rather than a monthly retainer with no scope attached.

Does advertising help organic rank?

Ads drive sales and sessions, and both feed the performance half of the ranking question. The durable value is the search term data, which tells you which words to put in the listing.

What single change moves listings most often?

The main image and the first bullet, because they set click through rate and conversion rate together. Text work on a listing nobody clicks has nothing to lift.

Shoppers asking before they open the app

Amazon ranks listings by sales and relevance, and your consultant works the fields that feed it. A shopper who asks an assistant which brand of a product to buy sits outside that system. The reply names brands, often drawn from reviews, publications and brand sites, and Amazon rank counts for little in it.

Check your brand and your two top products against that question. AI Knows Us runs it across the major assistants and records which brands come back, with the sources used.