Compare Marketplace Fees and ROAS Before You Scale

A channel is only good if the net survives

Marketplace demand can hide weak unit economics for a while. Direct storefronts can hide traffic cost. The useful comparison is the same product through multiple channels with all costs included.

That usually means platform fee, payment fee, shipping support, returns allowance, and ad spend or acquisition cost per order.

Compare channels with one product

ROAS without margin discipline is misleading

A campaign can show decent ROAS and still produce weak net margin if fulfillment, returns, and fees are heavy. That is why fee math and ad math need to live together.

Before scaling spend, test whether the product still works after realistic fees and traffic cost. If not, fix the offer or the price first.

• Track contribution margin by channel.
• Use average ad cost per order, not only headline ROAS.
• Reprice before scaling if the margin is thin.

A worked comparison across two channels

For an illustration, use $60 seller revenue, $20 product cost, and $7 shipping in both rows. If a marketplace statement shows a $7.80 fee and $3 attributed advertising cost, contribution is $22.20. If a direct-store statement shows $2.10 processing and $9 attributed advertising, contribution is $21.90. These are stated assumptions, not current platform defaults.

The $0.30 difference does not establish a winner. Repeat the reconciliation with actual fee bases, refunds, acquisition attribution, monthly platform overhead, and completed-order volume for each channel.

Signs a channel is not worth scaling further

• Net margin keeps shrinking as ad spend rises, even though revenue looks fine.
• Return rate on that channel is meaningfully higher than on others.
• Most of the visible ROAS is coming from repeat buyers you would likely retain anyway.
• Fulfillment or support cost per order is rising faster than order volume.

A worked ROAS-after-fees comparison

$500 of attributed ad spend and $2,500 of attributed seller revenue is 5x actual ROAS by definition. Subtracting marketplace fees does not create a different ROAS unless the business deliberately defines and labels a separate nonstandard metric.

To test profitability, keep the 5x ROAS and calculate contribution separately: revenue minus product, fulfillment, platform, payment, refund, and advertising costs. Do not rename fee-adjusted revenue as ROAS.

Attribution windows change the comparison

A 7-day click attribution window on one channel and a 30-day window on another will report different ROAS for genuinely similar performance, since the longer window credits more of the eventual sales back to the ad. Match attribution windows before comparing two channels side by side, or the comparison itself is misleading.

Blended ROAS hides channel-level problems

Looking only at a blended ROAS across all channels can mask one channel quietly losing money while another compensates. Break the number out by channel monthly, even if it takes a few extra minutes, so a declining channel gets caught before it has eaten a quarter's worth of margin.

The same logic applies within a single channel across different campaign types, branded search, generic search, and display or social placements typically have very different real returns once fees and cost of goods are factored in, even when the platform reports them under one combined dashboard number.

Seasonal spikes distort short comparisons

Comparing two channels' ROAS over a single week around a sale event or holiday period can produce a misleading snapshot, since conversion rates and ad costs both shift temporarily. Compare channels over at least a full month, and ideally the same month in a prior period, before drawing a conclusion about which one performs better after fees.

Keep a simple monthly scorecard

A single spreadsheet tab per month, listing each channel's spend, gross sales, fee-adjusted revenue, and resulting ROAS, takes a few minutes to maintain and turns this whole comparison from a one-time exercise into an ongoing habit that catches problems early rather than at year-end review.

Keep the comparison honest

Use the same fee-adjustment method for every channel in the comparison, applying commission and processing costs consistently rather than adjusting one channel carefully and eyeballing another, since an inconsistent method is what usually produces a misleading winner.