Your Amazon Growth Might Be Your DTC Decline: The DTC vs Marketplace Question Nobody Measures

The DTC vs marketplace picture in most reporting stacks contains a number that does not exist: total online growth. Amazon reports its growth. The direct channel reports its own. Nothing in between establishes whether a customer who bought on Amazon this quarter would have bought direct last quarter. Both lines rise, the sum looks like expansion, and part of it is migration.

• A share of marketplace growth in any omnichannel brand is displaced demand rather than created demand.
• Separate reporting makes that share structurally invisible, not merely hard to find.
• A dollar that migrates from direct to marketplace arrives with substantially less contribution attached.
• The question is answerable. Most brands have simply never been asked to answer it.

The DTC vs Marketplace Question Is Not Where to Sell, It Is Which Demand Moved

Almost every article written on this subject answers the wrong question. Search for DTC vs marketplace and you will find the same trade-off rehearsed repeatedly: margin against scale, brand control against traffic, first-party data against reach. That framing assumes the decision is where to sell.

Search marketplace vs DTC and you get the same list in the opposite order. For a brand already doing both, that decision was made years ago. The live question is different and considerably more uncomfortable: of the demand arriving through the marketplace, how much is new, and how much simply changed address?

The distinction matters because the two produce identical entries in a report and opposite entries in a P&L. New demand is a customer who would not have bought at all. Displaced demand is a customer who would have bought anyway, through a channel that returns more per unit and leaves behind data you can use again.

Nobody in the organisation is lying about this. The Amazon vs DTC comparison that would reveal it is not being calculated, because no single report contains both sides of it.

Two Channels Reporting Separately Mechanically Produce a Sum Larger Than Reality

This is arithmetic before it is anything else, and it is worth walking slowly because the conclusion is counterintuitive.

Marketplace reporting counts orders placed on the marketplace. Direct reporting counts orders placed on the site. Neither is wrong. But a customer who would have purchased direct and instead purchased on the marketplace appears once in the marketplace line as growth, and appears in the direct line as nothing at all. Not as a loss, which would be visible. As an absence, which is not.

An absence does not generate an alert. A direct channel that grows 4% when it would have grown 11% without displacement reports a good quarter. Nobody investigates a good quarter.

Add the two lines together and the total describes a business that grew more than it did, because the migration was counted once as a gain and never once as a transfer. The error is not in either report. It is in the addition, and the addition is what goes to the board.

In multichannel ecommerce this is not an edge case. Three conditions make it near-universal in brands above roughly $50M online:

• The two channels have separate owners, separate targets and separate dashboards. Each is measured on its own line.
• The marketplace passes no customer identity back. There is no shared key on which to join the two datasets, so the overlap cannot be observed even in principle without deliberate design.
• Nobody is accountable for the consolidated number except one person, usually the VP of ecommerce, who receives two reports that were never built to be compared.

A Migrated Dollar Loses 30 to 50 Percent of Its Contribution on the Way Across

Displacement would be a reporting curiosity if the two channels returned the same margin. They do not, and the gap is large enough to change what the growth is worth.

The marketplace fee stack is public. Referral fees run from 8 to 15% of sale price depending on category, with most categories at 15%. Fulfilment, storage and inbound placement layer on top, and industry accounting commonly puts referral plus fulfilment at 25 to 40% of revenue before any advertising. Advertising is technically optional and practically mandatory: most sellers spend 10 to 25% of revenue on it, and a healthy total advertising cost of sales sits at 8 to 15%. Marketplace Pulse has put the total effective take, ads included, at up to 50% of seller revenue.

The direction of travel is set. Amazon confirmed that 2026 fulfilment fees would rise by an average of $0.08 per unit, after holding rates flat in 2025.

Now apply that to a migrated dollar. Revenue is unchanged, so nothing in the top line signals a problem. Contribution is not unchanged. The same customer buying the same product through the marketplace instead of direct delivers materially less margin, and delivers no email address, no purchase history and no ability to be marketed to again without paying for the privilege a second time.

This is why the invisibility is expensive rather than merely untidy. The brand is not just miscounting growth. It is reinvesting behind the lower-contribution channel in the belief that it is funding expansion.

Your Amazon Team Is Optimising Exactly What It Was Asked to Optimise

It is worth saying this plainly, because the argument is routinely misread as a criticism of marketplace teams, and it is the opposite.

An Amazon team given a revenue target on Amazon will grow revenue on Amazon. It will bid on branded terms, because branded terms convert. It will improve conversion on product pages, because that is the lever available. It will scale spend where return on ad spend holds. Every one of those decisions is correct against the brief it was given, and a team that did anything else would be underperforming.

The brief is the problem. Your team was asked to optimise a channel, not a business. Nobody asked it to establish whether the incremental order it just won was incremental to the company or only to the channel, and it has no mechanism to find out even if it wanted to.

The same is true in reverse. A direct team fighting for its own target has no visibility into how much of its softness is displacement rather than underperformance, which makes it very hard to defend a budget it may well deserve.

Both teams are operating rationally inside a structure that guarantees the consolidated question goes unasked. That structure is an executive decision, not a team failure.

The DTC vs Marketplace Overlap Is Measurable, and Three Signals Tell You Where to Look

The reason to raise this is that it is answerable, and the first read costs less than most people assume. Before commissioning anything, three signals in data you already hold will tell you whether displacement is plausible in your business.

Branded search against marketplace growth. Track branded search volume and marketplace revenue on the same monthly timeline. If marketplace revenue climbs while branded search stays flat, the marketplace is not creating demand. It is capturing demand that already existed and would have landed somewhere.

Direct channel growth by geography. Compare direct growth across markets where marketplace presence differs. Comparable markets should behave comparably. Where direct growth diverges in the markets with the strongest marketplace presence, you have a pattern worth explaining.

New customer ratio on the direct channel. Displacement does not usually take new customers, it takes repeat customers who found a more convenient route. A direct channel whose new customer share holds while repeat revenue softens is describing migration rather than acquisition failure.

None of these three proves anything on its own. They are not designed to. They are designed to tell you whether the expensive question is worth asking, and to let you arrive at that conversation with a hypothesis rather than a worry. Establishing causation requires a deliberate test design, and that is a different exercise with a different cost.

Nobody Asks a Marketplace to Prove Its Own Incrementality

There is a reason this gap persists, and it is structural rather than conspiratorial.

Every measurement system built by a platform reports the value the platform created, as measured by the platform. No marketplace has a commercial reason to volunteer the proportion of its reported growth that its seller would have earned anyway, and none is asked to. The same has been true of every advertising platform for twenty years. Marketplaces inherited the pattern rather than invented it.

There is a smaller and more literal version of the same problem that most brands do discover eventually: the marketplace bidding on the brand’s own name in search results, so that the brand pays a platform to intercept demand it had already created. It is the same mechanism as displacement, visible only because it leaves an auction record. The cross-channel version leaves no record at all.

What is different here is scale and stakes. A distorted advertising report misallocates a media budget. A distorted channel report misallocates the entire growth strategy, because it shapes where inventory goes, where pricing is set, which team gets headcount, and which channel the business believes is its future.

This is what separates the displacement question from ordinary channel conflict, which is a pricing and distribution argument between partners. Nothing here is a dispute. It is an absence. The correction is not to distrust the marketplace. It is to stop accepting any channel’s self-reported contribution as a business result, which is a standard the brand applies to itself and not something the platform can be expected to supply.

Answering the DTC vs Marketplace Question Changes Where the Next Dollar Goes

Displaced demand and created demand look identical in a report and behave nothing alike in a P&L. As long as the two channels report separately, the difference between them stays invisible, and the consolidated growth number is larger than the business it describes.

The three signals above will tell you within a fortnight whether displacement is plausible in your numbers. What they will not tell you is how much, and that answer requires isolating the effect deliberately rather than inferring it from two charts.

That discipline is not new, it is simply usually applied one layer lower. Detecting when branded advertising has stopped being incremental, or flagging when the marketplace itself bids on your brand name in search, are the same question asked inside a single channel. Primelis Market runs that logic across more than 100 brands and over $1.2B in Amazon revenue. Asking it across both channels is the version that decides where the growth budget goes.