Contribution Margin by Channel: Why Your Blended Number Hides Which Customers Actually Lose Money

Contribution margin by channel is the per-channel profit left after subtracting every variable cost — COGS, payment fees, shipping, returns, and the advertising spend tied to that specific acquisition path — from the revenue that channel produces. A blended company-wide contribution margin hides the truth: some channels are subsidizing others, and the channels you celebrate in board meetings may be the ones quietly losing money on every order. This guide walks through how to calculate contribution margin by channel, how to spot the cross-subsidy, and how to act on what the numbers reveal.

Why the blended contribution margin lies

When Wayfair reported a Q1 2026 contribution margin of 15.0% of net revenue, the math underneath was clean on paper: a 30.5% adjusted gross margin, minus 3.8% in customer service and merchant fees, minus 11.2% in advertising, according to its April 30, 2026 8-K filing with the SEC. That single number reassures the market. What it cannot tell you is which incremental customer — the repeat buyer arriving via direct traffic, the new buyer captured by a Meta retargeting ad, or the marginal customer pulled in by a deep promo on Wayfair Professional — actually generated positive contribution.

The blended number is an arithmetic mean across wildly different cost-to-serve cohorts. If your direct and organic channel runs at a 35% contribution margin and your paid social channel runs at negative 5%, a 60/40 mix produces a blended 19% — a number that looks healthy enough to keep funding the channel that is destroying value. This is exactly the pattern that quietly broke Casper and Blue Apron.

According to a teardown of Casper's S-1 filing by Justine Moore, the company's blended LTV/CAC ratio settled around 1.4x — but the unit economics worked out to a roughly $157 loss on every mattress sold once you allocated paid acquisition spend properly to the cohort it pulled in. Blue Apron showed the same shape at a different price point. Daniel McCarthy's analysis of their public filings found CAC climbing to $169 with break-even requiring 8+ months of retention — a bar two-thirds of acquired customers never cleared.

Practical takeaway: If your CFO can only show you one contribution margin number, you are flying blind. Build the channel cut before the next board meeting.

The contribution margin by channel formula, step by step

Strip away the jargon and the calculation is a straight subtraction. Per channel, per cohort, per month:

Channel revenue: Net revenue (gross sales minus refunds, discounts, chargebacks) attributable to that channel.
Minus COGS: Unit cost of goods, allocated to the orders that channel produced.
Minus variable fulfillment: Pick, pack, ship, last-mile, returns processing.
Minus payment and platform fees: Stripe, PayPal, Shopify, Amazon referral fee, FBA fulfillment fee, marketplace storage.
Minus channel-specific advertising: Every ad dollar that targets that channel, including the agency fee or affiliate payout.
Minus channel-specific headcount allocated as variable: If you have a dedicated Amazon manager, that fully-loaded cost is variable to the Amazon channel.

What remains is contribution profit. Divide by channel revenue to get contribution margin percent.

The hardest line is item 6. Jordan Glickman's writing on contribution margin for paid media argues that ignoring channel-specific overhead understates the true cost — $200K in fully-loaded headcount on $2M of channel revenue erodes 10 percentage points of margin in a single line item, but that line is invisible in most ROAS dashboards.

Practical takeaway: Build a spreadsheet model that goes channel by channel — Direct, Organic Search, Paid Search, Paid Social, Email, Affiliate, Amazon, Wholesale, Retail. Force every variable cost into one of those columns. The columns that absorb the most overhead are usually the ones the blended number is hiding.

The marketplace tax: when Amazon revenue looks like growth but acts like a drag

Marketplace channels carry a structural cost stack that DTC channels do not. According to Stack Influence's 2026 marketplace fee math, a $75 Home & Kitchen product on Amazon carries an ~$11.25 referral fee (15%), $5.50 to $8.50 in FBA fulfillment, and roughly $0.45 in storage allocation — that's $17 to $20 of variable cost before a single sponsored-product ad dollar is spent.

Luca's 2026 ecommerce benchmarks show the consequence: the average profit margin of top Amazon sellers fell below 10% in 2023, while Shopify DTC merchants in the same period maintained 22%–35%. A product with identical COGS will show a 13–20 percentage point lower contribution margin on Amazon than on a Shopify DTC site before PPC is even allocated.

That does not mean Amazon is the wrong channel. It means Amazon is a customer acquisition vehicle priced like one, and treating it as a profit channel without doing the math is the mistake. The pattern Luca's benchmarks recommend — DTC at 60–70% of revenue, marketplaces at 20–30%, wholesale at 10–20% — works because each channel is playing a different role, and the contribution margin for each is being measured and managed separately.

Practical takeaway: If Amazon is more than 25% of your revenue and your blended margin is below 20%, run the channel cut this week. The mix is probably the problem, not the products.

The SaaS version: inbound vs. outbound contribution

The same disease shows up in B2B SaaS with different symptoms. SaaS gross margins look uniformly excellent — 70%-plus is standard — which lulls operators into thinking channel mix doesn't matter. It does, because the variable cost is not COGS, it's CAC.

Public benchmarks compiled in Digital Applied's 2026 CAC benchmark report, drawing on aggregated data from HubSpot, OpenView, ChartMogul, and ProfitWell/Paddle, put the median B2B SaaS CAC at roughly $702 for self-serve and $11,400 for sales-led. That is a 16x spread. A blended CAC at a SaaS company doing both motions is meaningless — it averages a free signup against a six-month enterprise sales cycle.

Tomasz Tunguz's classic teardown of HubSpot's S-1 is still the canonical example of how to read this. HubSpot's inbound motion produced customers with substantially better LTV/CAC than its outbound motion, and the company's strategic conviction came from being able to see that channel-level number — not from the blended one. More recently, in its Q1 FY2026 earnings call covered on Yahoo Finance, Stitch Fix flagged nine consecutive quarters of improving LTV for new client acquisition, explicitly tied to disciplined channel-level pacing of ad spend.

Practical takeaway: SaaS operators should compute contribution margin per acquisition motion, not per product. The variable cost is the human (or the ad) on the front end, not the marginal compute on the back end.

A five-step playbook to expose hidden losers

This is the order of operations that actually works in a finance-team Excel model or a contribution margin spreadsheet template:

Lock the cohort window. Pick a month or a quarter and freeze it. Channel-level CAC and contribution margin only work as a snapshot; live numbers will move while you analyze.
Attribute every order to a channel. Use last-touch if you must, but document the rule. Multi-touch is better when you can afford the modeling cost; pick one and apply it consistently.
Allocate variable costs cleanly. COGS and fulfillment are easy. Advertising goes to the channel that drove the impression. Platform fees go to the channel that hosts the transaction. Channel-specific headcount is variable to that channel even if your accountant calls it OpEx.
Compute contribution profit per order and per channel. Build the column. Then build a second column for contribution margin percent.
Rank the channels and look at the bottom. If the bottom channel is negative, you have a decision: cut spend, raise prices on goods sold through that channel, renegotiate fees, or accept the loss as a deliberate CAC investment with a documented payback timeline.

The right answer is often the fourth option — but only if you can articulate the payback. Wayfair's reporting in Retail Dive on its earlier years of escalating customer acquisition spend shows what happens when you cannot: the market punishes companies that scale a channel without showing a path to contribution profit on that channel.

Practical takeaway: The five-step pass produces a one-page table. That table is the single most useful artifact in any growth-stage operating review.

What to do with the channel that's losing money

A negative contribution margin channel is not automatically a channel to cut. There are four legitimate responses, ranked by reversibility:

Renegotiate the variable cost. Lower CPMs, better shipping rates, lower marketplace referrals via brand registry, lower payment processing rates at scale. This is the cheapest fix because it does not require changing the product or the customer.
Raise prices or shift mix on that channel. If Amazon margins are thin, list only the SKUs where they work. If paid social only converts on discounted bundles, stop running the SKUs that bleed.
Cap the spend. Treat the channel as a CAC investment with a budget, not a growth lever to maximize. A negative-contribution channel with a 12-month payback can still be the right investment — but only if you can prove the payback with cohort data.
Cut the channel. The fastest path to profitability is usually subtraction. Most operators wait too long because the channel produces revenue, and revenue feels like progress.

Practical takeaway: Run this decision once a quarter. The cost of leaving a value-destroying channel running for one more quarter is almost always larger than the cost of cutting it cleanly.

Stop flying blind on blended numbers

The blended contribution margin is a useful number for the cap table and for the press release. It is a useless number for operating the business. Every meaningful growth decision — where to spend the next marketing dollar, which SKU to push, which channel to scale, which to kill — depends on the channel cut, not the blended one. Casper, Blue Apron, and a long line of D2C casualties did not fail because they couldn't sell product. They failed because the blended number told them they were closer to profitability than they were.

Building this model from scratch is a week of finance team work the first time. A ready-made Excel template with the formulas already wired — channel revenue, COGS allocation, advertising attribution, platform fees, contribution margin percent, and a channel ranking output — turns that into an afternoon. ModelStack's contribution margin and unit economics templates are built exactly for this: drop in your data, get the channel cut on day one, and start having the conversation about which channels actually earn their seat at the table.