Amazon Inventories, Plural: What Changes When You Sell Into More Than One Marketplace

Ask an operations lead how much Amazon inventory sits behind a given ASIN and you will usually get one number back, pulled from whichever Vendor Central tab happened to be open. That answer holds up fine while the US is the only market you sell into, but once Canada or the UK or Germany is also live, the number covers one position out of several and gives you no way to tell which one you are looking at.

The plural in "Amazon inventories" is literal for any vendor past its first marketplace. Amazon issues a separate login per marketplace, and it forecasts and reports on each one by itself, so none of those positions net against each other anywhere inside Vendor Central. What follows is what that separation does to day-to-day inventory work. For the mechanics of a single account's cycle, our post on Amazon inventory management on Vendor Central walks the whole loop, and everything below assumes you have more than one of that loop running at the same time.

The separation starts at the sign-in

Vendor Central is not a single address. The US portal is vendorcentral.amazon.com, Canada has its own and Germany and the other EU markets have theirs, and the regional portals differ in more than the domain, since the interface, the features available, and the local tax and reporting requirements can all vary from one region's portal to the next.

Underneath each sign-in sits a vendor group, which is the account that login belongs to. A vendor group can hold several vendor codes, and those codes are how Amazon distinguishes separate business agreements, brands, and categories inside that one account. Codes and marketplaces run on different axes: the codes divide the business underneath a single login, while the marketplaces are the logins themselves. A brand carrying four vendor codes in the US still has one US position expressed four ways, whereas a brand with a US login and a German login has two positions, and neither account has a screen that acknowledges the other one exists.

So "how much Amazon inventory do we have" stops having a single-number answer on the day the second portal is issued. Every habit built while there was one account is worth a second look once there are two.

Each marketplace forecasts on its own

Forecasting is where the account structure starts costing you money. Vendor forecasting data is requested one marketplace at a time, and that is enforced rather than merely defaulted: ask for the forecast across more than one marketplace in a single request and it comes back an error, because there is no combined forecast sitting on Amazon's side to return.

That matters for planning, because a demand shift Amazon's US model has picked up on an ASIN tells you nothing about what its UK or Canadian model is doing with the same ASIN that week. The forecasts are built from different sales histories against different customer bases, so they move independently, and a vendor watching the US forecast climb on an item and committing supply against it can be walking into a stockout in Canada at the same moment, with no hint of it on the screen they were watching.

There is a second constraint on top of that one: only the most recent forecast is available for a given marketplace, and historical forecast data cannot be requested. Last week's forecast is gone once this week's replaces it, and that happens per marketplace, so if forecast-versus-actual analysis matters to your planning, the weekly save has to cover every market you are live in, including the small ones nobody opens most weeks. Miss it in a smaller market for a quarter and you have no way to reconstruct whether Amazon's demand picture there was ever right.

Blending the markets hides the newest one

Vendor Central reports are pulled individually by marketplace, a separate report per country, so every performance figure you end up reviewing started out as its own file. What happens to those files afterwards is your call, and what usually happens is an average.

Take a team that has run the US account for six years and opened a UK account eight months ago. They are looking at a mature operation and a young one, and the young one is where the learning curve is still being paid for. Average the two on a slide and the new market's trouble disappears into the old market's volume, which is precisely the stage at which it would still have been cheap to fix.

So keep the operating numbers in a per-market view. A vendor who confirms orders on time in the US and lets them slip in a newer EU market will look fine in aggregate for as long as the US volume is large enough to carry the average, and that can easily run a year before anyone asks a question about it.

Nothing rolls the markets up for you

Pulling per market also means a vendor live in five markets runs five pulls to answer one question and then joins the results by hand. (Every Vendor Central report, defined covers which report holds which field, if you are deciding what to pull in the first place.)

That joining work is where multi-marketplace inventory management quietly turns into a data project. Someone has to pull each market's files on a schedule, normalize what comes back, keep straight which item in one market corresponds to which item in another, and turn all of it into something a planner can read in one pass. It is recurring work that appears on no org chart, and it usually lands on whoever built the first spreadsheet, so when that person takes two weeks off the roll-up stops and nobody notices until a planning meeting opens with no numbers in it.

It is worth being precise about what is actually documented here, because most material written about selling across Amazon regions describes third-party selling and carries over badly to a vendor account. What the vendor-side documentation establishes is the separation itself: marketplace-specific logins, with reports and forecasts scoped to one marketplace per request. Plan the combined view as something your own team builds and maintains, because Amazon hands back the pieces and leaves the assembly to you.

What to actually track across marketplaces

Five things belong in a multi-marketplace inventory review, and teams tend to skip the first one:

• The count of live marketplaces, written down. Not from memory, and re-checked whenever a market is added, so that every number in the review has a marketplace attached to it.
• Open purchase order quantity and sellable position, side by side, per market, from the same week. A thin sellable position next to a large open PO reads very differently in a market you launched last quarter than in one you have supplied for years.
• Confirmation and ship-notice discipline, one line per market. Blending those lines turns a newer market's slippage into a rounding difference on the aggregate.
• A saved forecast file for every live market, every week. Amazon keeps only the current one and will not hand back last week's.
• Cross-market checks on a demand signal before anyone dismisses it. An unexplained lift on an ASIN in one region is worth ten minutes against the other regions before it gets written off as noise, and that check only happens if someone has access beyond their own market.

Give the roll-up a named owner and a named backup. These reviews usually break for a dull reason: the one person who knew how to assemble the picture takes three weeks off, and nobody looks at the smaller markets until they are back.

Where InsightLeap fits

The assembly problem in the last two sections is what InsightLeap is built around. Managing brands and regions in one place is the product's stated position, and its features page lists broader country support among its advantages. The unlimited-users-with-tiered-permissions model matters here too, because the people who need the combined view usually outnumber the people holding Vendor Central logins, and per-seat access math is often what keeps a planner covering one region from ever seeing another.

Start with the count

Before the next Amazon inventory review, settle one question ahead of any number on the agenda: how many Amazon marketplaces are we live in right now? Then require every figure in that meeting to name the market it came from. A vendor operating in four marketplaces is carrying four inventory positions and four forecasts that will not agree with one another, and a review that treats them as a single picture will keep missing whichever market is quietly going wrong.

If you want the single-account mechanics underneath all of this, start with Amazon inventory management on Vendor Central, then run the checklist above once per market you have live.