For an established consumer brand already selling through Shopify, the next growth investment should go to the specific opportunity the business can support—not automatically to the platform reporting the highest ROAS. Compare a defined Shopify improvement with a defined Amazon opportunity, using compatible product economics, inventory requirements, measurement and operating capacity. The answer may be to prioritize Shopify, test a limited Amazon assortment, operate both with distinct roles, or hold expansion until a named constraint is resolved.
There is no universal Amazon-versus-Shopify budget split. Before allocating resources, separate where the customer buys, who fulfils the order and how the result will be evaluated. A fulfilment change, a marketplace launch and an advertising test are different investments.
Publisher disclosure: This guide presents our editorial recommendations for evaluating channel investments. It is not an independent benchmark, an investment forecast or a client case study. Platform documentation supports the mechanics described below; it does not establish that a particular channel will grow a brand profitably.
Who this guide is for
This guide is for established North American consumer brands, typically around $5M–$50M+ in annual revenue, with proven products and an existing Shopify operation. The decision is how to use the next limited growth resource: advertising budget, inventory, working capital or the team’s time. The brand needs a commercial decision-maker and access to its own channel and operating evidence.
It is not a platform-selection tutorial for a pre-revenue business or a claim that every Shopify brand should launch on Amazon. If product demand is unproven, costs are unknown or nobody can own the new channel, resolve those conditions before treating expansion as the next step. Revenue size alone does not make a brand ready.
If your starting question is how to build a direct channel alongside retail or wholesale, use our retail and wholesale to DTC guide. This article starts later: an existing Shopify operation evaluating a specific Amazon-related next investment.
Separate the purchase destination from fulfilment and measurement
Amazon versus Shopify is not one decision. Three connected questions need separate answers.
| Decision | Question to answer | Evidence to examine |
|---|---|---|
| Purchase destination | Is the intended order on the Shopify storefront or in Amazon’s marketplace? | Shopper journey, offer, assortment and campaign destination |
| Fulfilment | How will the order reach the customer? | Eligible service and market, inventory location, order handling, returns and current costs |
| Measurement | Did the investment produce a worthwhile business result? | Recorded channel orders, applicable costs, attribution definitions and evidence of additional demand |
We recommend describing the intended change in each row before approving the investment. A fulfilment integration is not a marketplace launch. An attribution report is not a profit-and-loss statement. Combined ecommerce revenue cannot explain the health of each channel by itself.
Amazon fulfilment can support a Shopify purchase
Shopify documents an Amazon Multi-Channel Fulfillment, or MCF, and Buy with Prime app that can fulfil Shopify orders without requiring Amazon marketplace sales or Buy with Prime. The documented app requires an Amazon Seller Central account, US fulfilment and customers with US addresses. For MCF outside the United States, Shopify directs merchants to a different integration. A Canadian brand should verify the specific app and fulfilment route it intends to use rather than assume the same setup applies everywhere. Shopify MCF documentation
Amazon’s architecture documentation describes fulfilment and tracking updates returning to Shopify, with Shopify’s catalog and order information remaining authoritative in the illustrated setup. This describes a system; it does not prove lower costs, faster service for a particular merchant or higher conversion. Amazon architecture overview
Our reporting recommendation is to retain the purchase destination when classifying the order and attach the relevant fulfilment costs to it. Changing who handles delivery should not, by itself, turn a Shopify purchase into Amazon marketplace revenue.
Marketplace Connect serves a different purpose
Shopify Marketplace Connect supports connecting a product catalog to Amazon and other marketplaces and managing synchronized listings, orders and inventory through Shopify. Shopify Marketplace Connect
That can help operate a marketplace channel. It does not establish that campaign attribution, customer identity and profitability are unified. Before relying on an integration for a reporting decision, ask the operating owner to show where one order was placed, where it is recorded, how it is fulfilled and where its costs appear. Also establish whether another report credits the same sale.
Compare the cost of the actual order route
Start with the product and the way the customer receives it, not an average fee assumption for each platform. We recommend distinguishing these routes where they are available.
| Order route | What to evaluate | Mistake to avoid |
|---|---|---|
| Shopify purchase with the current fulfilment setup | The specific acquisition or storefront opportunity and its costs | Assuming an owned storefront removes payment, fulfilment, support or acquisition costs |
| Shopify purchase fulfilled through Amazon MCF | The proposed operational change for that Shopify order | Relabelling the purchase as an Amazon marketplace sale |
| Amazon marketplace purchase with the selected fulfilment method | The assortment, selling costs and operating model | Substituting a Shopify fulfilment estimate for marketplace selling costs |
For marketplace orders, use the current fee category and fulfilment inputs for the actual product. Amazon’s US pricing page distinguishes selling-plan charges from category-dependent referral fees and notes that a product’s fee category can differ from its customer-facing category. The Revenue Calculator described on that page compares Fulfillment by Amazon (FBA) with a seller’s own fulfilment using product-specific inputs. For Shopify orders fulfilled through MCF, obtain the applicable MCF estimate separately; do not treat a marketplace FBA estimate as the price of that different order route. These distinctions make a SKU-level review more useful than one assumed Amazon fee percentage. Amazon US selling costs
Make the comparisons compatible. Use the same reporting period and consistent treatment of discounts, returns, geography and currency. Identify the product, fulfilment, payment, marketplace and advertising costs that apply to each route. Account for each cost once: check what a quoted service includes before adding a second allowance for the same activity.
Use actual account charges where available and label estimates for a new route. A US fee schedule is not a Canadian fee schedule, and a calculator output is not an observed margin. If the most influential cost input is unverified, the comparison remains provisional.
Separate order economics from opening a channel
An attractive order calculation does not answer how much work is required before the channel can operate. We recommend showing one-time setup separately from recurring costs: product content, listing preparation, inventory coordination, integrations, reporting and management can all require resources. Include only the activities that apply to the actual scope, with verified costs or clearly labelled estimates.
Keep the treatment of existing overhead consistent across the options and identify new commitments. This is a forward-looking investment comparison, not a replacement for the brand’s financial reporting. A proposed channel can have acceptable order economics and still require more cash or operating capacity than the business is prepared to commit.
Keep attributed sales separate from incremental demand
Amazon Attribution measures eligible non-Amazon marketing activity leading to Amazon destinations. Its documented method uses a 14-day, last-touch click window. The guide also describes new-to-brand purchase measures using a one-year Amazon purchase lookback. Those definitions belong to the named measurement product; do not apply its click rule to every Amazon advertising report. Amazon Attribution guide
The interpretation matters. A credited order does not, by itself, establish that the sale would have disappeared without the campaign. A shopper classified as new-to-brand on Amazon is not proven to be buying from the brand for the first time across Shopify and every other channel.
Keep campaign credit, recorded commerce revenue and causal incremental demand separate. For example, a team should not add a platform’s attributed sales to the commerce records of those same orders and call the sum additional revenue. Use the existing revenue-reconciliation framework to define the comparison before evaluating performance.
For this decision, we recommend identifying what would constitute useful additional business and what evidence can actually test that claim. If the available reporting only shows attributed sales, say so. It can inform a decision without proving incrementality.
Decide where to invest next
Compare concrete alternatives, not platform reputations. “Improve a known Shopify purchase barrier” and “test a limited Amazon assortment” describe investments that can be evaluated. “Grow ecommerce” does not define enough to approve a commitment.
The following procedure is our recommended planning approach, not a validated allocation model or a description of a particular client engagement.
1 Define the opportunity and the alternative
State the commercial problem, the proposed change and the credible alternative. The alternative might be a Shopify improvement, an Amazon test, work on an existing marketplace operation or a temporary hold. State what the business forgoes by committing those resources so that the comparison does not assume expansion is the only choice.
Describe what each option changes: purchase destination, fulfilment, measurement, or more than one. This prevents a promising operational improvement from being mistaken for evidence that a new sales channel will work.
2 Resolve the assumptions that could reverse the decision
Review the product and market, eligibility, applicable costs, inventory availability and operating responsibilities. Identify any material dependency around assortment, offers or channel pricing. Treat unresolved contractual or pricing-policy questions as matters for the appropriate business and qualified advisers, not as assumptions to bury in the growth forecast.
Prioritize the unknown most likely to change the choice. If an unverified fulfilment estimate determines which option looks profitable, verify it before polishing the rest of the plan.
3 Make the resource commitment explicit
Write down the additional spend, inventory exposure, setup work and ongoing responsibilities. Identify who will execute the work and who can decide to stop it. Set a limit that the business can support; do not derive it from a generic percentage split between Amazon and Shopify.
The ad-spend planning guide addresses media affordability. Keep its net-DTC revenue denominator intact rather than silently adding unrelated marketplace revenue to make a spending ratio appear smaller.
4 Define the evidence and review condition
Specify the product scope, market, period, source definitions and commercial measures before launch. Decide when evidence will be sufficiently mature to review, accounting for the purchase cycle and relevant returns or reporting delays. There is no universal test duration in this guide.
Record what would support continuation, require repair or end the test. A useful review can conclude that evidence is insufficient. It should not manufacture a winner because a calendar checkpoint arrived.
5 Select a path and retain its reopening condition
| Decision path | Reason it may fit | Condition that should hold it back |
|---|---|---|
| Prioritize Shopify | A specific storefront or acquisition opportunity has credible evidence and executable economics | Preference for the platform substitutes for an actual opportunity |
| Test a limited Amazon assortment | Product eligibility, current cost inputs, inventory and ownership support a bounded test | Costs, operating responsibility or the test limit are unresolved |
| Operate both with distinct roles | Each channel has a defined purpose, separate economics and accountable ownership | Combined revenue conceals weak economics or overlapping attribution is labelled new demand |
| Hold expansion | A named constraint makes the current commitment unsupportable | The hold has no owner, review condition or path to resolving the constraint |
The decision is not permanent. Reopen it when the relevant evidence changes: a cost assumption is verified, inventory becomes available, a storefront constraint is resolved or operating capacity changes. Document why the chosen option is preferable now and what would make the alternative more attractive.
When outside support is useful
Outside support may help when the brand has a credible opportunity but needs coordination across advertising, creative, storefront conversion, marketplace operations or measurement. Assess the actual scope and who is qualified to own each part. Do not assume that Shopify or paid-media capability automatically establishes Amazon operating expertise.
An agency cannot make an unsupported product margin, missing inventory or unowned implementation plan disappear. Resolve those constraints before treating more marketing activity as the solution. For broader partner evaluation, use the guide to choosing an ecommerce marketing agency, then verify channel-specific responsibilities and evidence with the proposed team.
Frequently asked questions
Should an established Shopify brand also sell on Amazon
Not automatically. Evaluate a defined Amazon opportunity against the next credible use of the same resources. Product eligibility, economics, inventory, measurement and operating capacity should support the decision. Testing, operating both channels or holding expansion can each be reasonable under different evidence.
What percentage of the budget should go to Amazon versus Shopify
This guide does not recommend a universal split. Budget the actual opportunity and its risks, then compare it with the alternative. Revenue size and a channel’s reported ROAS do not establish how much additional investment the business can support.
Is Amazon fulfilment the same as selling on Amazon
No. A Shopify purchase fulfilled through MCF remains a different route from an Amazon marketplace purchase. Document the destination and attach the applicable fulfilment costs without conflating the two.
Does connecting Amazon to Shopify solve attribution
Do not assume so. Catalog, inventory and order synchronization address operating tasks. Verify separately how campaign credit, customer definitions and costs are reported before claiming that attribution or channel profitability has been unified.
Does an Amazon new to brand order prove a new customer for the whole business
No. An Amazon-scoped lookback cannot establish that the shopper has never bought through Shopify or another channel. Keep the platform’s measure separate from a verified first-ever brand purchase.
Should the channel with the highest ROAS get the next dollar
Not on that measure alone. Check what the report credits, the product and customer mix, applicable costs, inventory and the evidence for the additional investment. A healthy blended result does not establish that the next increment of spending will have the same economics.
When is holding expansion the better choice
Holding is reasonable when a material unresolved condition makes the commitment difficult to support. Name the condition, give someone responsibility for resolving it and state what evidence will reopen the decision. A hold should be an explicit commercial choice, not an indefinite absence of action.
Make the next investment specific
An established brand does not need a permanent platform winner. It needs a defensible next commitment. Separate purchase destination, fulfilment and measurement; compare the actual products and costs; expose the uncertain assumptions; and set an owner, resource limit and review condition.
The result may be Shopify, Amazon, both with distinct roles or a temporary hold. The evidence and the conditions for revisiting the choice make the decision useful—not the platform name.