Profit on ad spend beats ROAS for ranking channels because it subtracts each channel’s own costs before crediting the ad. On Triple Whale’s 2026 benchmarks, TikTok’s $4 CPM still needs a higher break-even ROAS than Meta, because fixed per-order costs weigh heavier on smaller orders.
A cheap impression is not a cheap customer, and a cheap customer is not a profitable one. Only one of those three numbers pays your bills.
TikTok ads cost a median $4.07 per thousand impressions for Triple Whale customers between October 2025 and September 2026, against $15.48 on Meta and $15.93 on Google. Those figures come from Triple Whale’s TikTok CPM benchmarks, published October 12 and built from more than 6,000 stores and about $962 million of tracked TikTok spend. Read on its own, that is a fourfold discount on reach, and it is the kind of number that ends up in a budget meeting as a reason to move money.
A day later, Wayflyer published the argument that should sit next to that number in the same meeting: your highest ROAS channel might be your least profitable one. Wayflyer’s case is sound, but every figure in it is explicitly made up, which makes it easy to agree with and hard to act on. Triple Whale has real benchmarks and no profit math. Put together, they answer a question neither answers alone: what does a cheap channel actually earn once the costs that ad dashboards never see are taken out?
Triple Whale sponsors the eCommerce Fastlane podcast, and its benchmarks describe its own customers, who skew toward brands that spend seriously on paid social. I treat its figures as vendor network data, useful and not neutral, and I lean on them here because they are the largest public dataset I know of with matching TikTok and Meta benchmarks.
ROAS ranks channels in the wrong order because it divides revenue by ad spend and ignores every cost that sits between the sale and your bank account, and those costs are different on every channel. The ad platform knows what the click cost and what the order was worth. It does not know your product cost, your shipping cost, the payment fee, the discount the customer used, or how often that channel’s buyers send things back.
Wayflyer’s illustration makes the point cleanly. It sells the same $60 product on two channels. Channel A, paid social to the brand’s own store, shows a 4.0 ROAS. Channel B, a social marketplace, shows a 5.0 ROAS. After marketplace fees, creator commission, platform-funded coupons, shipping, and higher returns, Wayflyer’s numbers have Channel A keeping about $10 per order and Channel B losing $4.60. The channel that wins on ROAS is the one to cut. Wayflyer labels those figures as invented, and so do I, but the mechanism is real and it is the one most merchants at the $500K to $2M stage never check.
I would add a second mechanism Wayflyer’s example does not stress: costs that are fixed per order. Picking, packing, and shipping cost roughly the same whether the order is $40 or $90. That cost takes a much bigger bite out of a small order, which means a channel that produces smaller baskets needs a higher ROAS just to break even. That is exactly where a cheap CPM can mislead you.
Profit on ad spend is the profit left after every cost tied to a channel’s orders, divided by that channel’s ad spend. A POAS of 0.5 means each ad dollar returned 50 cents of profit after paying for itself; zero is break-even; anything negative means the channel loses money on the orders it drives.
The calculation has three steps, and you only need a spreadsheet. First, work out contribution before ads for an average order on that channel: average order value, minus product cost, minus shipping and fulfillment, minus payment fees, minus the average discount, minus an allowance for returns at that channel’s return rate. Second, subtract the channel’s cost per acquisition to get profit per order. Third, divide profit per order by CPA to get POAS.
Two numbers fall out of the first step that I find more useful than POAS itself. Contribution before ads is your break-even CPA: the most you can pay for an order and not lose money on it. And average order value divided by contribution is your break-even ROAS: the platform ROAS you need to see before a channel is paying for itself. Our guide to why Shopify reports don’t show real profit walks through where each of those cost lines lives, because Shopify’s own reports stop at gross figures.
One honest limit: this is first-order math. A channel that loses money on the first order can still be the right investment if those customers come back. That is a real argument, but it needs your own repeat rate by channel behind it, not hope. Our piece on the five numbers to lock before BFCM includes the 60 and 90 day repeat rate for exactly this reason.
Here is the calculation run on Triple Whale’s published benchmarks, with one large caveat stated up front. The CPA and average order value come from Triple Whale’s TikTok ads benchmarks and its Meta ads benchmarks, both covering August 2025 to July 2026, across more than 5,900 and more than 40,000 brands respectively. Triple Whale labels its TikTok figures as overall benchmarks and its Meta figures as medians, and numbers from different brand sets do not describe any single store. The cost assumptions are mine and illustrative: product cost at 35% of order value, $8 per order for pick, pack, and shipping, payment fees of 2.9% plus 30 cents, and a returns allowance of 8% of order value.
Three things stand out. First, the cheap channel needs the higher bar. Because the same $8 of fulfillment lands on a $50 order instead of a $73 one, the TikTok order needs a 2.66 ROAS to break even against 2.34 for Meta. The fourfold CPM discount does not carry through to profit; it gets spent on smaller baskets.
Second, neither platform’s reported benchmark ROAS clears its own break-even line under these assumptions. Triple Whale reports an overall platform ROAS of 1.51 on TikTok and a median of 1.88 on Meta. Either the typical brand is losing money on first orders, it has much better margins than my assumptions, or the platforms undercount the sales they drive. Triple Whale itself notes that last-click attribution undercounts delayed purchases on TikTok, and that some downstream revenue never reaches the pixel. All three are probably partly true, and that is the point: you cannot know which applies to you without running your own costs.
Third, the TikTok margin is thin enough to disappear. At $1.85 of profit on a $17.07 CPA, a 20% rise in acquisition cost during BFCM, when auctions get crowded on every platform, pushes CPA to about $20.48 and profit to roughly minus $1.56 an order. A channel that looks like the efficient bet in October can become the one losing money on Black Friday, at exactly the moment most merchants raise its budget.
Before raising any channel’s budget for BFCM, I would set a break-even CPA for each channel from your own costs, then set a holiday ceiling below it with a margin for auction inflation. If a channel’s current CPA is within about 20% of its break-even, I would not scale it into November on the strength of its ROAS or its CPM. That is a call you can disagree with, and some merchants with strong repeat rates will reasonably accept a first-order loss, but it should be a decision made with the number in front of you.
The subtraction matters as much as the addition. In my years as a Senior Merchant Success Manager at Shopify, the pattern I saw most often at the $500K to $2M stage was premature complexity: a third ad channel opened because the second looked efficient in a dashboard, before anyone had checked whether the first two were profitable. If the math above says a channel is not earning its keep, the right BFCM move may be to cut it back and put the money into the channel that is. Our analysis of what branded search ROAS hides found the same pattern from another angle: the most flattering ROAS in an account is often the least incremental.
Promotions change the numbers too. A 25% sitewide discount is a cost line in the calculation, not a separate decision. Our breakdown of Q4 2026 margin math and holiday promo planning covers the inputs that moved since June, and every one of them flows straight into contribution before ads. Run the calculation with your Black Friday discount in it, not your everyday price.
A spreadsheet is enough to calculate POAS by channel until your channels, products, and discounts multiply past what you can update weekly by hand. The calculation does not change; what changes is how often the inputs move and how many combinations of product and channel you need to watch.
If you are under about $1M with two ad channels, use a spreadsheet. Pull 30 days of spend, orders, and revenue per channel, add your cost assumptions, and update it every two weeks. That is an hour of work and it will change at least one budget decision.
If you are between $1M and $5M across three or more channels, a profit analytics tool starts to pay for itself, mostly by keeping product cost, shipping, and returns current by channel and by SKU without manual work. Wayflyer’s article points to BeProfit, which is a Wayflyer partner. Triple Whale offers profit reporting inside its own platform, and our StoreHero review covers another option built around contribution margin for brands in this range. Whichever you choose, check that it can show profit by channel and by product on the same screen, because Wayflyer is right that one without the other is half the picture.
Above $5M, the question moves from per-order profit to incrementality: whether a channel’s orders would have happened anyway. That needs testing, not just accounting. Our pre-Q4 audit before scaling holiday ad spend starts with proving conversion tracking, which is the first input every one of these calculations depends on.
ROAS measures revenue per ad dollar, while POAS measures profit per ad dollar after every cost tied to the order. ROAS divides the revenue an ad platform attributes to your ads by what you spent, so it ignores product cost, shipping, payment fees, discounts, and returns. POAS subtracts those costs, then divides the remaining profit by ad spend. A POAS above zero means the channel made money on its orders; below zero means it lost money. Because those costs differ by channel and by order size, two channels with the same ROAS can produce very different profit, which is why POAS is better for deciding where budget goes.
Calculate break-even ROAS by dividing your average order value by your contribution before ads for that order. Contribution before ads is the order value minus product cost, shipping and fulfillment, payment fees, the average discount, and an allowance for returns. For example, a $50 order that keeps $19 after those costs needs a ROAS of about 2.6 to break even, while a $73 order keeping $31 needs about 2.3. Work it out separately for each ad channel, because order size, discount use, and return rates usually differ by channel, and compare each channel’s platform ROAS against its own break-even, not a single store-wide target.
TikTok is not automatically more profitable than Meta, even though its ads are much cheaper to show. Triple Whale’s 2026 data puts the median TikTok CPM near $4 against about $15 on Meta, but TikTok orders were also smaller, with a typical order value near $50 against about $73. Fixed costs like shipping take a bigger share of a smaller order, so TikTok can need a higher ROAS to break even. Whether it wins for your store depends on your margins, order size, and return rates by channel. Run a break-even CPA for each channel with your own costs before moving budget.
Include every cost that changes with each order a channel drives: product cost, picking and packing, outbound shipping you pay or subsidize, payment processing fees, the average discount or coupon used, marketplace or platform fees, creator or affiliate commissions, and an allowance for returns at that channel’s return rate. Then subtract the channel’s ad cost per order. Leave out fixed overhead such as salaries and software, which belong in your broader profit and loss rather than a per-channel comparison. If you sell the same product on several channels, calculate each one separately, because fees, discounts, and return rates usually differ.
Raise a channel’s Black Friday budget only if its cost per acquisition sits comfortably below its break-even CPA, not just because ROAS looks good. Ad costs usually rise during BFCM as more advertisers compete, and your Black Friday discount reduces the profit each order leaves behind. A channel that earns a small profit per order in October can lose money in November once both effects hit. Calculate break-even CPA with your holiday discount included, set a ceiling below it, and leave room for higher auction prices. If a channel is already close to break-even, scaling it into peak week is the riskiest place to put new budget.
