Amazon Product Positioning and SEO

June 11, 2023 by InsightLeap

You search a term your category should own and your ASIN comes back on page three, or it does not come back at all. The advice waiting for you after that search is the same ten-item checklist every time: rewrite the title, load up the backend search terms, shoot better images, chase reviews, watch your price. Plenty of that is worth doing eventually, but all of it assumes you already know what is wrong.

A low position covers two different failures, and they call for different work. Either Amazon is not putting your ASIN in front of shoppers for that query at all, or it is and other products are beating it. Work the checklist before you know which one you are looking at and a quarter can disappear into listing edits that never move the ranking. What follows separates the two problems, then walks the checks that tell them apart using data already sitting in your Vendor Central account.

Read the search page before you read your listing

Start with the page you are trying to rank on. An Amazon search results page mixes organic listings with sponsored placements, Amazon's Choice badging, and featured brand units, and most shoppers never scroll past the first page of it. That changes what a rank number is worth to you, because position 12 on a query whose top third is paid inventory in your category buys far less visibility than the same position on a quieter page.

You will also run into A9, and later A10, used as names for how Amazon ranks. Practitioners use those labels as shorthand, so they are worth recognizing when you read around the subject, but what you can work from directly is what you can observe in your own account. From there the outcome splits cleanly in two: either your ASIN comes back for the query or it does not.

Two reasons a listing ranks low

The listing never turns up for the query

Amazon matches a shopper's query against the text on your listing, the title and the backend search terms included. When the words a shopper types appear in neither, the listing does not come back for that search, and no amount of bullet-point polish changes that, because the polish is going onto a page that never turned up in the first place.

On the vendor side, the cause is often something other than the writing. More than one party writes to a vendor catalog, Amazon's own processes included, and those edits do not announce themselves, so the title and search terms live on the ASIN today are not necessarily the ones you filed. A listing change alert exists for exactly that: it tells you when your product listing content changes, so an unintended edit becomes an event with a date on it instead of something you find a month later while investigating a traffic drop.

The listing turns up and gets outranked

The second case is an ASIN that does come back for the query and sits below its competitors. What counts here is how shoppers behave around the listing: whether it gets clicked when it is shown, whether those clicks turn into orders, how the ASIN sells over time, and what its reviews and ratings look like, alongside price, image quality and count, in-stock availability, fulfillment method, and A+ content.

Vendors also carry a variable that the 3P-focused guides skip entirely: the featured offer, the position most people still call the Buy Box. When a third-party seller wins the featured offer on your ASIN, the numbers you would normally diagnose with stop describing you. Shoppers landing on that detail page may be buying from the other seller, and your glance views stop counting the traffic, because glance views (shown in Vendor Central as featured offer page views) count detail page views only while your own retail offer is the featured one.

How to tell which problem you have

Three checks, run in order, and none of them needs anything you do not already have.

1. Search the query the way a shopper would. Do it logged out or in a private window so your own account history does not skew the results, and page through far enough to be sure of the answer. If the ASIN does not appear, narrow the query: add your brand name, then search a distinctive phrase lifted straight out of the title. An ASIN that surfaces on the brand-qualified query but stays missing on the generic one is coming back for your brand and nothing else, which points at the listing text. An ASIN that surfaces at position 40 on the generic query is in the running and losing.

2. Line featured offer page views up against your catalog change history. In Vendor Central, Reports > Retail Analytics > Traffic carries glance views per ASIN. Pull a window long enough to contain whatever changed and mark the dates of every listing edit you know about. A sharp drop that lands on or within a day of a content change points at the catalog, while a gradual slide with the content untouched points at performance.

3. Confirm who holds the featured offer. Open the detail page and check whether the offer shown is yours. If a third party held it for part of the window you are looking at, the traffic and conversion figures for that period are not clean, and reading them as a ranking signal will send you the wrong way. Losing the featured offer is worth working as a problem in its own right, and it needs solving before the rest of the diagnosis means much.

The three answers together land you in one of four situations:

  • Missing on the generic query, present on the brand-qualified one: a relevance problem in the listing text.
  • Present, with glance views dropping on or near a known edit date: a catalog problem with a cause you can name.
  • Present, traffic steady, orders soft: a performance problem on the detail page.
  • Featured offer held by someone else: fix that first, then read the other three again.

What Amazon is measuring once you are in the running

It helps to keep two families of input separate. The relevance side is the listing text a query gets matched against, which is where your title and your backend search terms do their work. Getting those terms right is a keyword-research job, and we have covered the tactics in Amazon Keyword Research 101 and in our long-tail keyword guide instead of repeating them here, and the short version is that your fields have to cover the words shoppers actually type.

The performance side is everything that happens once the listing appears: click-through when it shows in results, conversion once clicked, sales over time, review count and rating, competitive pricing, image quality and count, in-stock availability, fulfillment method, and A+ content. The value of that list to you is triage. Every item on it is something you can measure from your own side, so you can find the weak one before you start rewriting anything.

Fixing each diagnosis

A catalog or relevance problem. Put back the fields that changed, starting with the backend search terms, since they are invisible on the page and nothing a customer sees will tell you they moved. Then confirm the correction stuck, because whatever overwrote your data once can do it again on its next run, and a fix that gets reverted on Thursday buys you nothing. Monitoring on the listing is what turns the next overwrite into something you catch in days.

A featured offer problem. Find the offers competing with yours on the ASIN and work out where that inventory came from, since unauthorized resale usually has a source you can address on the distribution side. While the featured offer is lost, treat that ASIN's traffic and conversion data as unusable for ranking diagnosis, and set an alert so you know the exact window later instead of reconstructing it from memory.

A performance problem. This is the case the standard checklist was written for, so this is where you work through it. Split the job by where shoppers drop off. Shown but not clicked points at what appears in the results row: the main image, the title, the price, the rating and review count. Clicked but not converting points at the detail page, so look at whether the content answers the objections buyers actually raise, whether the images show the product in use, what your A+ content covers, whether the item is in stock, and how your price reads against the alternatives beside you on the same page. Change one thing at a time and give it long enough to register, because sales history moves slowly.

Check the data before you touch the listing again

When a ranking question comes up, run the diagnosis before you commit a week of work to edits.

  1. Search the query logged out and record whether the ASIN appears at all.
  2. Pull Traffic in Retail Analytics and mark your listing edits on the glance view line.
  3. Confirm who held the featured offer during the period you are diagnosing.
  4. Fix the branch the data points at, then re-measure before starting the next round of edits.

Done by hand, that means opening Vendor Central and cross-referencing traffic against your own record of listing edits every time the question comes up, which is easy to let slide. InsightLeap carries those signals together: alerts when the featured offer is lost, and traffic reporting that puts glance views, page views, sessions, and conversion rate next to the sales numbers. Either way, keep the order. The diagnosis is what tells you which edit is worth making.