Dropshipping ads, margin first

Dropshipping ads are judged by ROAS more than any other kind, and ROAS is the number least able to say whether a dropshipper made money. Shopify's November 2025 guide to dropshipping puts the margins on products from open marketplaces between 10% and 15%, and between 20% and 50% through vetted supplier networks such as Shopify Collective. At a 15% margin, break-even ROAS is 6.67x; a 3.00x that would delight a brand with a 60% margin spends a third of revenue on ads against a 15% contribution, a loss on every order.

The second trap is whose ROAS. Meta reports the revenue its attribution credited to its ads under its own window; Shopify records the orders that happened. The two differ by design, and a product can show 3.00x in Ads Manager and 2.20x on store orders. For a product test, the store's figure is the one that pays the supplier.

So the first number is not a ROAS target borrowed from a forum. It is your own break-even, computed from the product's price and costs before the first dollar is spent, and every budget in this guide is a multiple of it.

Start from one order at the selling price. Take out the landed product cost, the shipping you pay the supplier or carrier, and the payment fee: Shopify's January 2026 article on card processing fees lists Shopify Payments' online rate at 2.9% plus 30 cents on the Basic plan in the US, falling to 2.5% on higher plans, and its pricing page shows different rates by country. What remains is the contribution per order, and the contribution divided by the price is the margin.

Break-even ROAS is the price divided by the contribution, which is the same as 1 divided by the margin. The worked example above lands at 1.95x on a 51.2% margin: a product with a healthy spread between price and landed cost. Swap in a $20 product cost on the same $39.90 price and the margin falls to 31.2%, break-even rises to 3.21x, and the most you can pay for an order drops to $12.44.

Then choose the margin you want to keep after ads, because break-even leaves nothing for the refunds you did not model, software, or your own time. Keeping 15% of revenue in the example gives a target ROAS of 2.76x and a maximum CPA of $14.45, and that CPA, not the ROAS, is what the test budgets below are built on. The break-even ROAS calculator does the arithmetic from your own figures.

How to run ads for dropshipping comes down to spending in stages your break-even can afford. Launch one Sales campaign with one ad set per creative, three to five creatives, the purchase event as the optimization event and a daily budget per ad set of about one break-even CPA: $20 here. Three days later, an ad set that has spent twice the break-even CPA without a single store order is killed, whatever its clicks say; one whose store-recorded CPA is under break-even survives.

Survivors run four more days at the same budget. The confirm rule is a store CPA at or under the target across the full seven days, on at least ten orders, because a CPA computed on three orders is noise. Meta's Help Center says an ad set typically leaves the learning phase after about 50 optimization events in the week after its last significant edit, and that one unlikely to reach them is marked learning limited. At $20 a day and a $14 target CPA you will not get there in a test, and that is acceptable: the test asks whether the product sells at all, not whether delivery is stable.

The kill rule is the point of the method. Written before the launch, it removes the conversation about giving the product one more day. A product that holds target CPA for seven days moves to the scaling step; one that does not is paused with its figures kept, so the next test starts from what this one taught.

• Test: 3 to 5 ad sets at one break-even CPA a day each, 3 days
• Kill: 2 × break-even CPA spent, no store order
• Confirm: store CPA at or under target over 7 days, 10 orders or more
• Scale: 30% steps, 3 days apart, while CPA holds

A dropshipped product rarely has a brand to lean on, so the creative carries the whole sale and wears out faster than the product does. Watch each ad against its own earlier week: frequency rising while click-through rate falls and the cost per order climbs is the fatigue pattern, and a low click-through rate from launch at low frequency is a weak message, which is a different fix. The creative fatigue guide sets out the diagnosis table.

Plan the turnover rather than reacting to it. Keep two untested creatives ready for every product in the scaling stage, and replace a fatigued ad with a new execution of the angle that sold, not a new angle, until the angle itself stops converting. In Adrails, Creatives lists each image and video once across every ad that uses it, with a signal: Fatigued when the analysis finds its results wearing out, Wasted when it spent with no result, Learning under 1,000 impressions. A Creative fatigue automation can post that list every week.

The rules above use store-recorded orders, so the store must be readable next to the ad account for the same dates, product by product. Ads Manager cannot show it, and a spreadsheet built on Monday is wrong by Tuesday when a refund lands.

Adrails reads Shopify orders and ad spend for the same dates. Orders arrive in real time through Shopify's webhooks and are reconciled every fifteen minutes, and Analysis sets net sales, orders and sessions next to Meta spend for the chosen period, each source in its own row, so the platform's revenue and the store's are both visible. A campaign's Tracking tab lists the customers each ad brought and what they spent, for campaigns whose links carry the ids, which campaigns created by Adrails add automatically. Tag ad names with a product label, and Creatives filters every figure by product.

Read the ratio between Meta's attributed revenue and the store's each week. A steady gap is attribution doing what it does; a sudden one is tracking, consent or a checkout change, and it should be checked before any budget moves on it.

Meta's Help Center lists what counts as a significant edit: pausing the ad set, or changing its optimization event, audience or creative, with budget and bid changes significant depending on the magnitude of the change. A significant edit may restart the learning phase, and Meta's learning phase page advises realistic budgets and warns that frequent budget changes can send an ad set back into learning. Doubling a winner overnight is therefore the most common way to lose it.

Raise in steps instead: at most 30% per change, then three days before the next read, with the rule that store CPA must still sit under target. Keep the creative and the audience untouched while the budget moves, and add new creatives as new ads rather than editing the ones that work. When a step pushes CPA over target, step back once and hold; a product has a ceiling, and finding it is the job.

Adrails holds the same line by construction. A budget change the Media Buyer prepares is capped at 30%, an increase needs at least one result and a day's budget of spend behind it, the evidence shown is at most 24 hours old, and the change is applied in one click by an owner or admin, with an undo while the platform still holds the state Adrails wrote. Everything Adrails creates on Meta starts paused, so a new ad set at a higher budget never spends before you have seen it.

The ad side is half the business; the other half is a supplier that ships on time at the cost the break-even assumed. DSers sources and fulfills from AliExpress, 1688 and Alibaba with price and stock sync and has no advertising features at all, so the supplier price you settle there is the product cost line of the formula. CJ Dropshipping runs sourcing agents, warehouses and carriers with no monthly subscription, and can shoot product videos meant for Facebook ads: footage of the item you will actually ship.

Zendrop adds an AI ad generator whose videos are downloaded, not published, and AutoDS automates import, price monitoring and fulfillment for the store. None of them runs the ads. Each meets Adrails in your Shopify store: the orders they fulfill are the orders Adrails reads next to spend, and the videos they produce are the creatives you drop into the conversation to launch the next test.