Others wait to be hired.We already started.
This is a real diagnostic we ran on a live account before that brand hired us. The brand name, product names, ASINs and competitor names are removed. The scoring, the method and the findings are exactly what the scan returned. Share a product or storefront link and this is the document that comes back on your catalog.
Run this on my brandRedacted brand · 5 generations, still hand sewn in the original mill building
A family manufacturer, 5 generations in, still cutting and sewing by hand in the building where the company started. The line began in industrial textiles, moved through military contracts, and settled on foam seat cushions after the postwar housing boom. The current generation built the consumer brand in the 2010s and sold direct for the first time, on the brand site and then on marketplaces. Every cushion is made to order in more than 100 fabrics with US sourced foam.
Your top 5 parents, the money and the condition, in one view.
One row per parent listing, not per variation.
| Product | % of top 5 rev | Price | Est. 30 day rev | Est. annualized | Ratings | LQS | Alexa | COSMO | Listing signals |
|---|---|---|---|---|---|---|---|---|---|
| Dining Chair Cushions, 4 pack, flagshipB0•••••••• | 84.4% | $124.99 | $103,447 | $1,241,364 | 71 4.5 | 8.4 |
Queued | Queued | A+VideoS&SDeal |
| Porch Rocker Cushion Set, extra largeB0•••••••• | 7.0% | $79.99 | $8,630 | $103,560 | 745 4.5 | 8.4 |
Queued | Queued | A+VideoS&SDeal |
| Dining Chair Cushions, smallB0•••••••• | 4.5% | $36.99 | $5,477 | $65,724 | 41 4.7 | 8.5 |
Queued | Queued | A+VideoS&SDeal |
| Dining Chair Pad with Ties, mediumB0•••••••• | 2.1% | $52.99 | $2,552 | $30,624 | 138 4.6 | 8.1 |
Queued | Queued | A+VideoS&SDeal |
| Bench Cushion, 3 seatB0•••••••• | 2.0% | $44.99 | $2,410 | $28,920 | 29 4.4 | 7.9 |
Queued | Queued | A+VideoS&SDeal |
LQS is listing quality out of 10. Alexa is Alexa shopping (formerly Rufus) answer readiness. COSMO is Amazon's semantic relevance read. Alexa and COSMO read Queued here because the agent queue had not completed these 5 parents when we pulled the file. Those scores are live for other listings in the catalog and appear in Finding 04. Revenue figures are public third party estimates, not your reported sales.
7 findings, every one of them from your live listings.
These are observations, not prescriptions. Each one names the mechanism first, then what we saw. What we would do about them is the conversation, not the document.
Your top parent is indexed almost everywhere. It ranks well almost nowhere.
Your top parent is indexed on 930 organic terms, which is a mature footprint, so the opportunity is in placement rather than in coverage. The other 4 parents sit between 340 and 630 organic terms, which is a working footprint with room left in it.
583 terms carry paid visibility with no organic ownership behind them. Paid is holding shelf space that organic is not yet earning, so the spend is renting a position rather than building one.
0 terms rank organically with no paid support underneath them. Where organic is already strong on real volume, sponsored coverage appears to be running alongside it, so there is no exposed position sitting open. That is disciplined, and it is the reason the opportunity here is placement rather than budget.
65 terms sit at organic 20 to 50, which is one push from page one. That is the band a rotation program works, and it is large enough to run as a standing motion rather than a one time fix.
The filter is straightforward. We look for terms where paid impressions exist and organic rank does not, then rank what is left by search volume against the effort to move it. On the call we would walk you through where we would set those thresholds for your catalog, since that judgment is where the list gets short enough to act on. That same question drives what an assistant can say about you, which is the next finding.
The 5 scanned parents each index between 340 and 930 organic terms. The footprint is already built, so the work is placement rather than discovery.
You win the comparison. You lose the question that comes before it.
Alexa shopping (formerly Rufus), AI Overviews and chat assistants read a short structured answer off your detail page, then decide whether to name you. A missing attribute is not a soft copy issue at that point. It is a silent disqualification, because the assistant will not guess a number on your behalf.
We scored a bank of spoken buyer queries against each scanned parent, 120 in total. 85 answered, 23 lost. Comparison lands, which tracks with 5 generations in the category and 5,940 reviews at 4.86 stars on your own site. Product fit and use case are where it thins out, at 7.0 and 4.2 out of a possible 18.
The single most repeated miss is reversible, absent as an extractable value on 3 of the scanned parents. You build a genuinely reversible cushion and say so in your own marketing, so this is a listing structure problem rather than a product one, and an assistant asked for a two sided cushion has nothing to read.
AI knows who you are.It does not appear to recommend the product.
When a shopper asks an assistant for the best in a category, the answer is assembled from the pages that already rank and already get cited. Being excellent is not the qualifier. Being present in those sources is.
The brand score of 74 and the product score of 69 tell the story in the gap between them. Asked about you directly, the answer is complete and favourable. Asked to recommend cushions without naming you, the answer is assembled from marketplace bestseller pages and affiliate roundups, and you surface in 1 of the 6 category authority prompts we ran.
The exception is instructive. On the made in USA and machine washable framing you surface readily, and one rocking chair guide names you outright as the buy that outlasts the chair. That is the position the category has left open, and almost nobody is holding it.
Your listings are written for humans. Machines are increasingly the ones adding to cart.
Alexa shopping (formerly Rufus) and COSMO decide whether an agent can confirm a product is right before it adds it to a cart. Compliance is the part you pass. The part you lose is everything that requires the listing to answer a question rather than describe a product.
The 6 listings our agent queue had completed at the time of this pull all land in the same band, 38 to 54 out of 100, and none of them passes. A tight band across every completed listing reads as a catalog condition rather than one weak page, which is the more useful finding, because it means one structural fix travels across the catalog instead of a rewrite per listing.
Compliance averages 77%. Assistant answers average 25%, with 13 of 90 buyer questions fully answered. The listings are clean. They are just not answering.
The catalog is well built. 3 mechanical things are holding it back.
A+ content and video are present on every scanned parent, images run 5 to 9 per listing, and the review component earns effectively everything available to it. The expensive part of the build is already paid for, so the money here is in the mechanical items below rather than in a rebuild.
The 3 mechanical items are the copy structure, the media weighting, and the titles. Listing copy earns 27% of what is available and media earns 41%, while reviews earn 99%. Your reviews are carrying pages that your copy is not.
All 4 scanned parents run titles longer than 75 characters, between 148 and 199. Amazon's non media title cap took effect on July 27, 2026 and moves the overflow into Item Highlights. Left alone, that rewrite happens to you rather than by you, and the words that survive are chosen by a truncation rule rather than by a merchandiser.
You own the buy box on every parent and use almost none of what that unlocks.
The offer is the last thing a shopper compares. It is also the only lever on a detail page that can move conversion without touching the copy, the images, or the price.
At the time of our scan no coupon, promotion, virtual bundle or merchandising badge surfaced on any scanned parent. You hold the buy box on all of them and you are merchant fulfilled across the board, which means you control the offer completely and are currently running it flat.
Your own site runs instant coupons and a try before you buy option. Neither appears on the marketplace listings, so the shelf where you have the least pricing pressure is the one running your least developed offer.
| Lever | Status | Coverage |
|---|---|---|
| Coupon | None found | 0% |
| Promotion | None found | 0% |
| Subscribe and Save | Not applicable | — |
| Virtual bundles | None found | 0% |
| Merchandising and badges | None found | 0% |
Every figure above reflects what was publicly visible at the time and location of our scan. Marketplace results vary by shopper, device and region, so treat direction as reliable and any single value as an estimate. Revenue figures are third party estimates rather than your reported sales.
That was 7 findings from public data alone. Your own scan runs on your catalog and comes back in one business day.
Run this on my brandThe same brand, scored on 2 shelves.
Most diagnostics stop at Amazon. Your Walmart listings answer to a different set of signals, and the comparison is where the sequencing decision actually gets made.
Walmart scores higher on the machine read and lower on the human one.
Walmart's semantic coverage reads more than twice your Amazon average, so the content written to Walmart's taxonomy is doing work for a semantic engine that the Amazon copy is not. Assistant answers are weak on both shelves, which is one gap in 2 places and one fix that pays twice.
One gap does not transfer. Amazon rewards the structured attribute fields an agent reads. Walmart rewards category taxonomy and content completeness. A brand that fixes only the shelf it sells most on tends to widen the gap on the other.
| Signal | Amazon6 completed scans | Walmart1 item |
|---|---|---|
| Agent Add to Cart Readiness | 44 of 100 | 56 of 100 |
| Assistant answer readiness | 25% | 17% |
| Semantic coverage | 31% | 74% |
| Marketplace compliance | 77% | 81% |
The findings above are the part we can read from outside.
Below is the shape of what the full analysis produces. The forecast is a transparent worked example so you can see the method. The other 2 stay locked, because they are the work rather than the teaser.
Axis starts at $120k so the slope is readable. The dashed line is your own trend continuing, held to a gently rising path because a public scan gives us no basis to model a decline. The shaded area is the lift above that line, and each figure is measured against the dashed line at the same month rather than against month 0. Directional sample built on your top 5 parents, not a projection of your account and not a guarantee.
The segmented funnel separates branded, generic and conquest demand so you can see which of the 3 your spend is actually buying, where conversion is losing against the category, and how much of the shelf is reachable rather than theoretical.
| Play | Pairing | Read |
|---|---|---|
| Bundle | A + B | Bought together, sold separately |
| Conquest target | A + C | Basket shared with a rival |
| Cross sell sequence | B then D | Second purchase window |
| Catalog expansion | Gap | Demand you do not make yet |
The value is the play, not the basket list. For a catalog that spans dining, rocker, bar stool and table linens, the cart gaps are also a product development brief, because they show demand you could make rather than only demand you could bundle.
We are an operator group, not an Amazon agency.
Amazon is one shelf. For a manufacturer that owns its own production, the margin conversation runs wider than any single marketplace.
Questions we answer for you, using numbers we already have.
We are not going to ask you for your revenue or your margin. We have done the reading. The call is where we hand back what it means.
We already started.Let us do this on your catalog.
Everything above came from public data on 5 listings of someone else's account. Share a product or storefront link below and we run the same scan on yours, then send it back inside one business day.