Skip to content
Build log

Published 8 min read

Operational streamlining

How we give a brand’s Amazon listings one owner, starting with a report we check first

A brand that sells through independent sellers needs someone to run its listings. We design that job with one owner, limited access, brand approval for every change and a report checked before anyone acts on it.

Contents
  1. A discount that was not there
  2. The job between brand and sellers
  3. Who runs the listings
  4. A key cut to the job
  5. A few products, then the catalogue
  6. Where the money comes from
  7. What listing work cannot fix
  8. For a brand in the same position
  9. Where we would look first

An early build of our price report said that in one marketplace, the brand’s products were offered at a median of about half their list price. List price here means the brand’s recommended retail price. For a brand that sells through independent sellers, that reads like a price war, and like sellers who might soon stop carrying the range.

Our re-audit, run before the report went out, traced the figure to a matching step. Cheap accessories, sold under titles that name the main product they fit, had been matched to that main product. Their prices counted as deep discounts on the real thing.

One rule removed the false discount. When an offer costs less than a quarter of the matched product’s list price, the row loses its match, and the reason is written next to it. With no new data, the deep discount left that market’s figure.

That episode shaped the rest of the setup. On Amazon, someone has to run the listings for such a brand: the content, the duplicates and the links between markets. Whoever holds that job acts on reports like this one, and acts wrongly when the report is wrong. What follows is the design and the working tooling behind it.

A price report we check comes first. Then one owner prepares catalogue changes, and the brand approves them.

  • Price report, one marketplace
    Found in our re-audit. A person checks the extremes and a sample of the rest.
    A false discount removed before the report went out
    Before
    Accessories matched to main products, read as deep discounts
    After
    Offers under a quarter of list price lose their match, with the reason
  • Effort per further market
    The first market carries the full setup.
    Each further market should need less effort
    Before
    Full setup in every market
    After
    Records, content, rules and report reused
  • Who runs the listings
    Sellers keep stock and prices, and the brand approves each catalogue change.
    One accountable owner for every change
    Before
    Spread across the brand and its sellers
    After
    One owner, limited rights, brand approval
The first row describes a correction our re-audit made to the price report on public marketplace information.
Where one price-position number comes from Lineage of a market's median price versus list price in our price report. Five steps: offer taken from public marketplace information, offer matched by name to a catalogue product, list price of the matched product, a price rule as gate, gap and market median. The rule removes the match when an offer costs under a quarter of the matched product's list price and writes the reason into the row. Before the rule, cheap accessories matched to main products pulled one market's median to about half below list. After the rule, with no new data, the deep discount was gone. Accessories above the floor can still pass, so a person checks the extremes and a sample. Where one price-position number comes from Lineage of our price report, from public marketplace information to the market median Offer Price and title, public market data Name match Catalogue product matched by name List price Of the matched product Price rule Under a quarter of list price? Gap, median Per offer, then per market 01 02 03 04 05 Match removed Reason written into the row. A person checks the extremes and a sample of the rest. yes Without the rule Cheap accessories matched to the main products they fit. Their low prices count as discounts on the real thing. Market median: about half below list. With the rule The same offers lose their match, with a reason. Accessories above the floor can still pass. No new data. The deep discount is gone. Values rounded. No market, product or seller named. Where one price-position number comes from Lineage of a market's median price versus list price in our price report. Five steps: offer taken from public marketplace information, offer matched by name to a catalogue product, list price of the matched product, a price rule as gate, gap and market median. The rule removes the match when an offer costs under a quarter of the matched product's list price and writes the reason into the row. Before the rule, cheap accessories matched to main products pulled one market's median to about half below list. After the rule, with no new data, the deep discount was gone. Accessories above the floor can still pass, so a person checks the extremes and a sample. Where one price-positionnumber comes from Lineage of our price report, from publicmarketplace information to the market median 01 Offer Price and title, public market data 02 Name match Catalogue product matched by name 03 List price Of the matched product 04 Price rule Under a quarter of list price? yes Match removed Reason written into the row. A person checks the extremes anda sample of the rest. 05 Gap, median Per offer, then per market Without the rule Cheap accessories matched to the mainproducts they fit. Their low prices count as discounts on thereal thing. Market median: about half below list. With the rule The same offers lose their match, with areason. Accessories above the floor can still pass. No new data. The deep discount is gone. Values rounded. No market, product or seller named.
Where one price-position number comes from, and where the rule sits. Lineage of our price report.

A discount that was not there

A price position is a short chain. Each offer comes from anonymised public marketplace information that names no seller. We match it by name to a product in the brand’s catalogue. Then we take that product’s list price and compute the gap. The market median is the middle of those gaps.

Matching by name was the step that needed a second signal. Accessories carry the name of the product they fit. Name matching had no way to tell that a very low price meant a different item.

Price was the missing signal. A genuine product rarely sells for a fraction of its list price, while an accessory usually does. So the rule sits between the list price and the gap, and it removes a match instead of correcting a price. The quarter is a deliberate setting, low enough that genuine discounts keep their match.

No rule catches everything. An accessory priced above a quarter of the list price passes it. So does a lookalike from another brand that carries the product’s name in its title. Before a price statement goes into a report, a person checks the rows at its extremes and a sample of the rest.

A false discount would have sent the brand after a problem that did not exist. It could also have misled the brand’s own decisions on its list prices and its pages.

What sellers charge stays their decision. Under EU competition rules, a brand may recommend a resale price but may not fix one or set a minimum, including through pressure. The report serves the brand’s own decisions and is never a lever on a seller’s price.

The job between brand and sellers

One level up sits the commercial question. A brand that sells through independent sellers wants to show up well in every marketplace without becoming a seller there. Its sellers run everything from stock to the customer. The brand owns the products, the name and the list prices.

Between the two sits a job that decides how the brand looks in each market. It covers one listing per product, one content standard, duplicates removed, pages linked across markets and a price picture the brand can rely on. We describe this layer in The Amazon Shelf Layer, a free framework in our library. In many brands, nobody is assigned to it.

Who does what on the listings Operating model in three lanes. The brand owns products, name and list prices, sets standards and approves changes. The listings owner, one accountable party, keeps one listing per product, the content standard, cross-market links and an honest price report, and prepares and files changes. Sellers hold stock, set prices, ship, take returns and serve customers. Approvals flow from the brand to the listings owner, reports flow back, and the listings owner gives sellers clean pages while public offer data feeds the report. Who does what on the listings Three layers, one owner for the middle one. Brand Products, name, list prices Sets the content standard Approves each change Decides on new products Grants inventory rights Listings One accountable owner One listing per product Content to one standard Links pages across markets Keeps the price report honest Sellers Independent, many per market Hold stock and set prices Ship and take returns Serve customers Win the order on the page approvals, standards clean pages reports, proposals public offer data The brand approves every change. Sellers keep prices and stock. Operating model as we propose it. Who does what on the listings Operating model in three lanes. The brand owns products, name and list prices, sets standards and approves changes. The listings owner, one accountable party, keeps one listing per product, the content standard, cross-market links and an honest price report, and prepares and files changes. Sellers hold stock, set prices, ship, take returns and serve customers. Approvals flow from the brand to the listings owner, reports flow back, and the listings owner gives sellers clean pages while public offer data feeds the report. Who does what on thelistings Three layers, one owner for the middle one. Brand Products, name, list prices Sets the contentstandard Approves eachchange Decides on newproducts Grants inventoryrights approvals, standards reports, proposals Listings One accountable owner One listing perproduct Content to onestandard Links pagesacross markets Keeps the pricereport honest clean pages public offer data Sellers Independent, many per market Hold stock and setprices Ship and takereturns Serve customers Win the order onthe page The brand approves every change. Sellerskeep prices and stock. Operating model as we propose it.
Who does what on the listings. Operating model as we propose it.

In our model the split of work is plain. The brand sets the standards and approves each change. Whoever runs the listings prepares and files catalogue changes and keeps the reports honest. Sellers keep everything that happens once a customer decides to buy.

Who runs the listings

Three kinds of owner can take the job. A team inside the brand knows the products best, although it often sits far from the marketplaces and has other work on its desk. One of the larger sellers knows Amazon well and is there every day. An outside team knows Amazon and comes in without a history in that range.

With a seller in charge, the problem is trust rather than skill. Whoever controls titles, content and merges could tilt them toward its own offers, and every other seller would assume it does. Sellers who suspect that start protecting themselves with pages of their own, and the duplicates come back.

Whoever takes the job, the brand should ask about conflicts of interest before it grants access. An outside team that also works for one of the sellers should say so, and the two jobs stay separate. We work for none of the sellers.

Who could run the listings Scoring of three owners on two axes: trust of the other sellers and speed to set up. A team inside the brand: high trust, slower to set up. One of the larger sellers: fast, low trust because it competes with the other sellers. An outside team: fast, and placed high on trust only when conflicts of interest are disclosed before access. The placement is our judgement, not a measurement. The brand keeps approvals and a written handover in every case. Which fits depends mostly on whether the brand has people free for the job. Who could run the listings Two things decide it: whether other sellers trust the owner, and how fast it can start. Placement is our judgement. Speed to set up → Trust of the other sellers → Brand team One of the sellers Outside team if conflicts are disclosed Brand team Knows the products best. Often far from the marketplaces, with other work on its desk. One of the sellers Knows Amazon well. Competes with the others, so other sellers protect themselves. Outside team Knows Amazon, fast to start. Trusted only when conflicts of interest are disclosed. Who could run the listings Scoring of three owners on two axes: trust of the other sellers and speed to set up. A team inside the brand: high trust, slower to set up. One of the larger sellers: fast, low trust because it competes with the other sellers. An outside team: fast, and placed high on trust only when conflicts of interest are disclosed before access. The placement is our judgement, not a measurement. The brand keeps approvals and a written handover in every case. Which fits depends mostly on whether the brand has people free for the job. Who could run the listings Two things decide it: whether other sellerstrust the owner, and how fast it can start.Placement is our judgement. Speed to set up → Trust of the other sellers → Brand team One of the sellers Outside team if conflicts are disclosed Brand team Knows the products best. Often far from the marketplaces, with otherwork on its desk. One of the sellers Knows Amazon well. Competes with the others, so other sellersprotect themselves. Outside team Knows Amazon, fast to start. Trusted only when conflicts of interest aredisclosed.
Who could run the listings.

Between the brand’s own team and an outside team, the trade-off is speed against continuity. People who have done the work before get going faster. The brand keeps the standards and the approvals, and a written handover lets its own team take the job over later. Which of the two fits depends mostly on whether the brand has people free for it.

A key cut to the job

Two Amazon systems hold the keys. The first is Brand Registry, Amazon’s programme for brand owners. A brand can join it even if it does not sell on Amazon. The Rights Owner and Administrator roles stay with the brand, and the Administrator assigns roles to others. An outside team can get the role Amazon provides for third parties who report suspected rights violations for a brand. Tools such as A+ Content and Stores also need a selling account.

The second key is that seller account. An outside team needs the catalogue there, not the rest of the account. If the brand holds its own seller account, its owner can add users under Settings, User Permissions. Each section is then set to None, View or View & Edit. The outside team gets View & Edit for inventory (Manage Inventory / Add a Product) only, with no access to payments, reports, advertising or account settings. The same right would let a user change prices and stock, and the permission does not rule that out. So the limit rests on the agreement: the outside team leaves both alone, since in this model they sit with the sellers. In our design every filed change is recorded with its approval and date. The brand can see what was changed. The brand can remove the access at any time.

Some brands prefer to file every change themselves, and some hold no seller account of their own. For them a slower route needs no access at all. The outside team prepares each change with its evidence, and the brand’s own team files it. That adds coordination time to every single change.

Either way, each change follows the same path. It goes out with its evidence, the brand approves it and it is filed. Afterwards someone checks that the page shows what was intended. A rejected filing returns to the queue with the platform’s reason attached.

One catalogue change, from proposal to check Process for one catalogue change. A proposal with evidence goes to the brand for approval. If approved, the outside team files it with inventory rights only. The same right would allow price and stock changes, so by agreement the outside team leaves prices and stock untouched. Afterwards the page is checked against what was intended and the result recorded. A declined proposal goes back with the brand's reason, and a filing the platform rejects returns to the queue with its reason. Alternative lane without outside access: the outside team prepares the change with its evidence, the brand's own team files it, and the page is checked afterwards. One catalogue change, from proposal to check Every change passes the brand's approval. Access is limited to inventory. By agreement, prices and stock stay untouched. Proposal Change plus its evidence Brand approves? yes File Inventory rights only, revocable Check Page shows what was intended Record Result and date kept no: back with the reason platform rejects: back to the queue with its reason Route without outside access For a brand that files itself or has no account. Same approval, same check. Costs coordination time on every change. Outside team Prepares change, evidence, wording Brand files Its own team, its own account Page checked The live page afterwards Process design. Timings are agreed with the brand. One catalogue change, from proposal to check Process for one catalogue change. A proposal with evidence goes to the brand for approval. If approved, the outside team files it with inventory rights only. The same right would allow price and stock changes, so by agreement the outside team leaves prices and stock untouched. Afterwards the page is checked against what was intended and the result recorded. A declined proposal goes back with the brand's reason, and a filing the platform rejects returns to the queue with its reason. Alternative lane without outside access: the outside team prepares the change with its evidence, the brand's own team files it, and the page is checked afterwards. One catalogue change, fromproposal to check Every change passes the brand's approval.Access is limited to inventory. By agreement,prices and stock stay untouched. Proposal Change plus its evidence Brandapproves? yes File Inventory rights only, revocable Check Page shows what was intended Record Result and date kept no: back with the reason platform rejects: back to the queue with itsreason Route without outside access For a brand that files itself or has no account. Same approval, same check. Costs coordination time on every change. Outside team Prepares change, evidence, wording Brand files Its own team, its own account Page checked The live page afterwards Process design. Timings are agreed with the brand.
One catalogue change from proposal to check. Process design with the approval gate and the no-access alternative.

A few products, then the catalogue

The work goes live first on the products with the most visits, not on the whole catalogue. Processing times and rejection reasons differ by kind of change, and nobody predicts them well. A small start measures them.

With those measurements, the full catalogue can be planned on real durations instead of guesses. After that, the work moves market by market. Each new market reuses the product records, the content standard, the rules and the report built for the one before.

After the rollout the job continues, since new pages keep appearing. So the report is repeated, and the job stays staffed rather than ending after one cleanup.

Where the money comes from

A product page earns from its visits, the share of visitors who buy and the price. Well-run listings work on visits and on that share. One page per product can bring together reviews and sales history that were spread over several, and ad spend concentrates on one page. Amazon merges duplicate pages only for identical products, and a merge request can be refused.

The clearer effect is on effort. The first market carries the full setup of records, content, rules and report. Each further market reuses most of that, so the effort to open it should fall with every step.

Effort to open each further market Bar chart of the effort to open markets one to five, indexed to the first market at 100. Without one owner for the listings, each further market costs almost as much as the first. With one owner who reuses product records, the content standard, rules and the report, the effort falls with each further market. Effort to open each further market Index, first market = 100. The shape is the point, not the values. Market 1 Market 2 Market 3 Market 4 Market 5 Markets in the order they open Effort to open (index) Each market set up on its own One owner, reuse What gets reused Product records and links, the content standard, rules and the report. What stays per market Language, local rules, sellers with stock, the platform's processing times. Effort to open each further market Bar chart of the effort to open markets one to five, indexed to the first market at 100. Without one owner for the listings, each further market costs almost as much as the first. With one owner who reuses product records, the content standard, rules and the report, the effort falls with each further market. Effort to open each furthermarket Index, first market = 100. The shape is thepoint, not the values. Market 1 Market 2 Market 3 Market 4 Market 5 Markets in the order they open Effort to open (index) Each market set up on its own One owner, reuse What gets reused Product records and links, the contentstandard, rules and the report. What stays per market Language, local rules, sellers with stock, theplatform's processing times.
Effort to open each further market, with and without one owner for the listings.

Most of the margin on these sales stays with the sellers, who hold the stock and set the prices. The first thing the design aims at is one page per product in every market the brand opens. Whether more of its demand then reaches its own products is something only the first markets can show.

What listing work cannot fix

Some of what a product loses in a category has nothing to do with listings. Shoppers filter search results by features. When a product lacks a feature buyers filter for, it drops out of those results, however good its page is.

We say this early, since listing work can hide a product problem for a while. Clean listings still help in that case. They make the product gap visible. The next product then launches onto a clean catalogue, with content and links already in place.

What well-run listings lead to Growth path in five steps: a checked report without the worst mismatches, the most-visited products live, the full catalogue planned on measured times, the next markets reusing the setup, and the next product launching onto a clean catalogue. A side branch shows what listing work cannot close: a product that lacks a feature buyers filter for drops out of filtered results, and that gap goes to product. Steps two to five are design intent, not measured results. What well-run listings lead to Each step reuses the one before. One gap stays with product. 01 A checked report Worst mismatches out, reasons kept 02 Top products live Real processing times and rejection reasons 03 The full catalogue Planned on measured processing times 04 The next markets Records, standard, rules, report reused 05 The next product Launches onto a clean catalogue What listing work cannot close A product without a feature buyers filter for drops out of filtered results, however good its page is. That gap goes to product. How to read this Step 1 is built. Steps 2 to 5 are design intent. No sales result is shown here. What well-run listings lead to Growth path in five steps: a checked report without the worst mismatches, the most-visited products live, the full catalogue planned on measured times, the next markets reusing the setup, and the next product launching onto a clean catalogue. A side branch shows what listing work cannot close: a product that lacks a feature buyers filter for drops out of filtered results, and that gap goes to product. Steps two to five are design intent, not measured results. What well-run listings leadto Each step reuses the one before. One gapstays with product. 01 A checked report Worst mismatches out, reasons kept 02 Top products live Real processing times and rejection reasons 03 The full catalogue Planned on measured processing times 04 The next markets Records, standard, rules, report reused 05 The next product Launches onto a clean catalogue What listing work cannot close A product without a feature buyers filter fordrops out of filtered results, however good itspage is. That gap goes to product. How to read this Step 1 is built. Steps 2 to 5 are design intent. No sales result is shown here.
What well-run listings lead to, and the gap they leave to product.

For a brand in the same position

  1. Check the report before you plan. A price picture built on wrong matches sends the whole team after the wrong problem.
  2. Before you hand over your listings, ask about conflicts of interest with your sellers.
  3. Keep the owner roles in Brand Registry, grant inventory rights only in the seller account and keep every approval with the brand.
  4. Start with a few high-traffic products and let them tell you how long each change really takes.
  5. Keep the report running after the rollout, since new pages keep arriving.

Where we would look first

If you sell through independent sellers, open your own price report for one marketplace and sort it by the deepest discounts. Then look at what the bottom rows really are. If spare parts or accessories sit there as your main products, fix the matching before the figure reaches any decision. If the rows are genuine, the report is ready for the brand’s own decisions on list prices and pages.

Credits

Our team at Capcelerate built the report and its audits, and designed the operating model. The price report builds on anonymised public marketplace information.

Tell us what you want to grow. The intro call is free.

Request a free intro callFree, up to 30 minutes. No obligation.