Amazon global selling: a 100-point entry guide

By Naeela

Amazon global selling is worth pursuing only when one new marketplace scores at least 75 out of 100 on demand, contribution margin, compliance, logistics, inventory, competition, and operating capacity. For brands generating $1M+ in annual revenue, the safer first move is a 90-day pilot in 1 market with 3 to 5 proven SKUs.

International sales are not growth when tax, landed cost, cash, and stock risk live in separate spreadsheets. Score the market before moving inventory. Then make the launch earn its next tranche of cash.

What is Amazon global selling?

Amazon global selling is Amazon's system for reaching customers in stores outside a brand's home market. A US seller can use a North America and Brazil unified account for Canada, Mexico, and Brazil, then create regional or country-specific accounts for Europe, Asia-Pacific, the Middle East, and North Africa.

Amazon's current US-to-international guide describes the operating sequence as account setup, product selection, requirements, listing, fulfillment, and ongoing management. Amazon also provides cross-listing, supply-chain, shipping-provider, currency-conversion, Marketplace Product Guidance, and Sell Globally tools.

Those tools reduce administration. They do not decide whether the market deserves your cash.

A product that works in the US may face a different retail price, VAT treatment, return cost, package rule, language, delivery promise, CPC, or review profile elsewhere. The listing can go live while the business case remains dead on arrival. That is why the first question is not, "Can we switch this market on?" It is, "Can this market repay the inventory, compliance, and attention it consumes?"

Why is market size the wrong first filter?

Demand matters, but market size alone rewards the loudest opportunity rather than the most executable one.

Imagine a leadership team comparing Canada, the UK, and Germany. The deck shows population, ecommerce growth, and category sales. Germany wins the top-line slide. Then the launch team discovers that packaging needs work, the listing needs real localization, tax ownership is unclear, and the forecast assumes inventory can move freely without changing registrations or cash timing.

The market was large. The plan was not ready.

Amazon says its Europe account can reach shoppers in 28 countries and that Pan-European FBA can distribute units across EU countries. That reach is useful only after the brand understands where inventory may be stored, which registrations follow, what the local offer must earn, and who owns each compliance decision.

Market size belongs inside the decision. It should not control the decision.

We use a 100-point scorecard because expansion is a connected operating choice. A strong demand score cannot cancel a negative contribution margin. Easy registration cannot rescue weak product-market fit. A good launch plan can still fail if the home business cannot spare the stock or senior attention.

How should you score a new Amazon marketplace?

Score the evidence, not the enthusiasm. Use the same 7 criteria for every candidate market so the favourite does not receive a kinder test.

The thresholds are an operator's decision rule, not an Amazon policy. Change the weights when the category demands it, but freeze them before scoring markets. A regulated children's product may deserve 25 points for compliance. A bulky product may need more weight on logistics. Do not change the rules after seeing which country wins.

The score also needs vetoes. Do not launch if any of these is true:

1. The SKU is contribution-negative at the expected local price.
2. Product legality, labelling, tax registration, or responsible-person coverage is unresolved.
3. The business cannot fund both the pilot and the home-market replenishment plan.
4. No named person owns the weekly market decision.

A 78 with an unresolved legal gate is not a pass. It is a spreadsheet trying to overrule reality.

How do you prove local demand without fooling yourself?

Use 3 layers: Amazon demand data, the live competitive shelf, and your own product evidence.

Start inside Seller Central. Amazon points sellers to Marketplace Product Guidance and the Sell Globally dashboard for country-level demand and recommendations. Treat those forecasts as a starting point. They do not know your landed cost, cash constraint, defect history, or appetite for operational risk.

Next, inspect the local search results as a buyer would:

• local query language and category terms;
• price bands with VAT included where applicable;
• delivery promises and fulfillment methods;
• rating and review-count distribution;
• pack sizes, dimensions, claims, and certifications;
• sponsored density and the brands holding organic visibility;
• local objections visible in reviews and questions.

Then test whether your existing evidence travels. A US bestseller badge does not prove UK price acceptance. Amazon says some verified-purchase reviews from other stores may appear and be translated, but it does not guarantee transfer or merger of the US review count. Eligibility and display can vary by product, brand data, store, and variation structure. Check each live ASIN and model conversion without assuming every US review will appear. A hero SKU may also rely on a size, claim, ingredient, plug, or use case that does not transfer.

Choose 3 to 5 SKUs that already have stable conversion, repeatable supply, controlled return reasons, and enough contribution to absorb learning costs. A 40-SKU catalogue creates more translations, compliance records, forecasts, and stranded-inventory paths before the team knows whether one offer works.

Which costs belong in the market-entry model?

Build the model from the customer price backward. A converted US P&L misses too much.

Start with the local selling price, then subtract:

• VAT or sales tax treatment;
• Amazon referral and fulfillment fees;
• inbound freight, duties, brokerage, and receiving;
• storage, aged inventory, removals, and disposal;
• returns, refunds, and local reverse logistics;
• currency conversion and settlement costs;
• localization, compliance, and tax-administration costs;
• advertising required to learn and launch;
• expected promotions and price protection;
• product cost and contribution required to justify the cash.

Run at least 3 cases: base, downside, and exit. The downside case should combine lower conversion, higher CPC, slower sell-through, and a weaker exchange rate rather than changing one cell at a time. The exit case should price the cost of removing, liquidating, returning, or disposing of inventory.

There is no responsible universal setup budget, launch timeline, or minimum margin for global selling. Amazon's current fees vary by store, category, size, weight, fulfillment path, storage profile, and optional program. For Remote Fulfillment, Amazon says the program cost and destination-store referral fee replace the US FBA cost and referral fee for that order. Build each SKU from the current fee preview instead of borrowing a range from a blog post.

Tax needs its own owner. EU guidance last checked July 13, 2026 says VAT obligations vary by where a business buys and sells, whether it trades goods or services, and whether the customer is a VAT-registered business or a final consumer.

The EU One Stop Shop can consolidate eligible cross-border consumer sales into 1 registration, return, and payment. It does not erase domestic inventory obligations. Amazon's current Pan-European FBA instructions require a VAT number in each country enabled for storage and at least 2 enabled storage countries. Confirm the storage footprint, stock transfers, registrations, returns, and effective dates before turning placement on.

For qualifying online-marketplace consignments outside the UK at the point of sale, HMRC's guidance generally makes the marketplace responsible for VAT when the total intrinsic consignment value is £135 or less. Above £135, normal import VAT and customs rules apply. B2B sales with a supplied UK VAT number, excise goods, non-commercial imports, Northern Ireland, and importer facts can change the treatment. When an overseas seller imports inventory into the UK, import VAT and customs duty also belong in the model.

If the model cannot show contribution after those costs, the market is not ready. A positive ROAS will not rescue a negative order.

What must be settled before inventory moves?

Compliance begins before the purchase order leaves the warehouse.

Amazon's Europe guidance tells sellers to review VAT, product compliance, safety, and listing requirements for each country. It points to Manage Your Compliance, Compliance Reference, and a vetted Service Provider Network. Useful tools, but the seller still needs a named internal owner who understands what the provider has and has not accepted.

For consumer products in the EU, the General Product Safety Regulation is part of that gate. Regulation (EU) 2023/988 covers online sales targeted at EU consumers, requires safe products, and creates duties around technical documentation, product identification, traceability, safety information, and an EU-established responsible economic operator for relevant products. Online offers also need the manufacturer, responsible-operator details where applicable, product identification, and required warnings. The rules have applied since December 13, 2024.

GPSR is not a universal substitute for category law. Food, cosmetics, electronics, toys, medical devices, chemicals, and other categories can carry separate requirements. CE marking applies only where the relevant harmonized legislation requires it. A responsible person does not make an unsafe or incorrectly labelled product compliant.

Before launch, create one evidence folder per SKU with:

1. applicable laws and category rules;
2. test reports and technical documentation;
3. manufacturer, importer, and responsible-person details where required;
4. packaging, label, warning, and instruction files;
5. marketplace approvals and listing evidence;
6. tax registrations, filing owner, and effective dates;
7. recall, complaint, and suppression escalation paths.

If the answer to a compliance question is "Amazon let us create the offer," the answer is incomplete. Listing acceptance is not legal advice and it does not transfer ownership of the product.

Which packaging and CE rules travel with the product?

Treat packaging, conformity marking, and product safety as separate workstreams.

EU guidance on CE marking is blunt: CE is required only when harmonized EU rules specifically mandate it, and it must not be used otherwise. Applicable products need the correct conformity assessment, technical file, declaration, and marking placement. GPSR does not replace those steps.

Packaging adds a different layer. The EU Packaging and Packaging Waste Regulation generally began applying on August 12, 2026, while many detailed requirements phase in later. Its current implementation guidance retains national producer registers. Registration turns on where a producer first makes packaging or packaged products available, not simply where Amazon stores inventory. The UK has separate threshold-, role-, and activity-based packaging-EPR rules.

Before launch, map the product, retail pack, shipping pack, inserts, claims, language, producer role, registration IDs, reporting owner, and evidence by destination. Do not put a CE logo on every box or assume 1 EU registration covers every packaging obligation.

Which fulfillment model should you choose?

Choose fulfillment by contribution, delivery promise, and inventory risk together.

Amazon's Remote Fulfillment with FBA currently uses eligible US FBA inventory for orders in Canada, Mexico, and Brazil. It handles export logistics, customer service, and returns, while eligibility, destination rules, and fees remain product-specific.

For a Canada decision, that leaves 3 basic paths:

1. Local FBA inventory in Canada.
2. Remote Fulfillment with FBA using inventory held in US fulfillment centres.
3. Fulfilled by Merchant through the seller's own operation or shipping providers.

Remote Fulfillment can test demand without sending a separate Canadian inventory pool. That lowers one form of risk. It may still carry a different fee, price, delivery promise, returns, and conversion profile than local FBA. Confirm the current ASIN eligibility report and SKU-level economics in Seller Central before choosing it. Do not promise a launch in days or a fixed delivery speed.

For Europe, Amazon describes local FBA, with Pan-European FBA as an additional distribution option, and merchant fulfillment. Eligible products can also use certain UK-to-EU or EU-to-UK Remote Fulfilment lanes without destination inventory. Enrollment, price caps, compliance, delivery, import duty, and VAT economics still apply. Local inventory can improve the delivery proposition but creates placement, tax, compliance, and cash decisions. Cross-border fulfillment can limit stranded stock but weaken the customer promise or contribution.

Inventory deserves a 13-week cash plan by marketplace. Include units on hand, units in transit, lead time, minimum order quantities, customs delay, forecast error, safety stock, and the point at which the pilot can no longer reorder without stealing cash from the home market.

The recognizable failure is a global dashboard showing healthy total stock while the winning UK variation has 9 days left, Canada holds 120 days, and the next factory order is 70 days away. One total hides three different decisions.

How should listings, ads, and pricing change locally?

Cross-listing is administration, not localization.

Amazon's Build International Listings tool can create or connect offers across stores and can help synchronize prices. That does not remove the need to decide which product language, proof, price, and claim will win locally.

Localization should cover:

• the terms shoppers use for the product and its attributes;
• units, dimensions, pack sizes, and spelling;
• category-specific warnings and legal copy;
• images whose claims and labels are valid in the market;
• A+ Content and brand-store paths;
• customer-service and return expectations;
• price presentation, VAT treatment, and promotion depth.

Use the Amazon listing checklist as an execution reference, then apply it to the local search shelf rather than copying the US page. The Amazon SEO guide explains why relevance, choice, inventory, price, and advertising have to be read as one system.

Ads need a local learning plan too. Begin with bounded budgets and search-term discovery, then move spend toward queries that can produce contribution after the full market cost. Do not import US bids, budgets, and targets as if customer demand and CPCs crossed the border untouched.

What should a 90-day pilot look like?

A pilot is a decision system with an exit, not a small launch that drifts into permanence.

The 90-day clock is an ALFI review schedule, not a profitability promise.

Before day 1

Freeze the 3 to 5 pilot SKUs, scorecard, price ladder, contribution model, compliance file, inventory cap, ad budget, owners, and stop conditions. Capture the baseline for conversion, returns, stock, review profile, and expected contribution.

Day 30: prove the offer works

Check whether listings are eligible, traffic is relevant, the offer earns clicks, and orders arrive without unexpected suppression, tax, or fulfillment problems. Fix catalog and compliance defects before adding spend. Do not read a tiny sample as a final verdict, but do not ignore an obvious structural failure either.

Day 60: prove the economics can improve

Separate launch learning from a permanently weak model. Review contribution after actual fees, ads, returns, currency, and operating cost. Compare local query conversion, price response, delivery promise, and stock cover. If the market needs endless discounting to move, call that what it is.

Day 90: expand, hold, or exit

Release more inventory only when contribution, sell-through, replenishment, compliance, and team capacity support it. Hold when the evidence is promising but one repair remains. Exit when the offer cannot earn acceptable economics without assumptions the team has already disproved.

Use the Amazon agency pricing guide when comparing the cost of external support with the internal coordination load. The cheap-looking model can become expensive when tax, catalog, logistics, and ads have different owners and nobody owns the combined result.

Where does ALFI fit, and where does it not?

ALFI is built for established brands that need one senior team to connect marketplace selection, unit economics, inventory, catalog, creative, advertising, and weekly decisions. Naeela remains in the strategy and accountability conversations, while experienced operators own the work across the system. The 18-client ceiling protects that continuity.

We would refuse a five-market launch when the business can only support one clean pilot. We would also refuse responsibility for profitable expansion while controlling only PPC. That costs us isolated project revenue, but accepting a fragmented mandate would make the promise dishonest.

ALFI is unusually right for brands generating $1M+ a year on Amazon, or $1M+ DTC brands ready to build Amazon as a serious channel, when expansion has become an operating and profit decision. A tax specialist, compliance firm, localization partner, or internal hire is the better fit when the need is narrow and the brand already has one capable owner joining the pieces.

The reciprocal standard matters. We owe candour, senior judgment, continuity, and early risk escalation. The client must provide unit economics, inventory truth, access, and the willingness to stop a market that does not earn more cash.

What should you do this week?

1. Choose 2 candidate marketplaces and freeze the 100-point scoring criteria before researching either one.
2. Build local contribution models for 3 to 5 proven SKUs, including tax, duties, fulfillment, returns, ads, and currency.
3. Name the owner for tax, packaging/EPR, product compliance, catalog, inventory, advertising, and the weekly go-or-stop decision.
4. Inspect the live local shelf and document price, delivery, reviews, pack size, claims, and sponsored density for 10 competing ASINs.
5. Write the day-30, 60, and 90 stop conditions before inventory moves.
6. If the score clears 75 and no veto remains, see how ALFI owns Amazon growth.