PLAYBOOK
How to advertise a dropshipping product.
Sourcing is the easy part. This is the part that decides whether a product works: how many angles you can test, how fast you can turn a winner into variations, and how to not get your account flagged along the way. Here's the version that holds up in practice.
- The five angles worth testing before you touch targeting
- Why creative volume beats creative polish — up to a point
- The compliance mistakes that get product ads rejected
Test angles, not audiences
The instinct is to launch one creative and split it across five audiences. On modern ad platforms that's backwards — targeting is largely automated, and the creative is the targeting. A different angle finds a different buyer far more reliably than a different interest stack does.
So the unit of testing is the angle: the reason someone would want this thing. Five angles for one product tells you something. Five audiences for one creative usually tells you nothing you couldn't have guessed.
The five angles worth testing first
Problem / solution
Open on the annoyance, not the product. Works when the pain is specific and visual.
Social proof
A real review quote as the hook, with the rating on screen. Cheap to make, hard to beat.
Price and offer
Was / now, bundle maths, free shipping. Only honest numbers — the landing page has to match.
Demo
The product doing the thing. For anything with a visible mechanism, this is usually the winner.
Objection handling
Answer the reason people don't buy: fit, materials, safety, returns, "does it actually work".
The follow-up: variations
Once an angle works, change one variable per variant — hook, opening frame, or format.
One product, several angles
The same listing, framed five different ways.
A testing loop that converges
Ship five angles
One product, five distinct reasons to buy. Same budget each.
Kill on the hook
Judge the first three seconds first — most losers lose there.
Fork the winner
Three to five variants of the angle that worked, one variable changed each.
Refresh before fatigue
New creative on the winning angle before CPMs climb, not after.
How many creatives does one product need?
Enough to cover the distinct angles, which in practice is around five to start. More than that up front usually means you're making variations of one idea and calling it a test — you'll spend the budget and learn one thing.
The volume comes later, after an angle proves out. Then variations are worth making in bulk, because you're no longer asking "does anyone want this" — you're asking which execution of a known-good idea performs best.
Why do product ads get rejected?
The most common cause isn't the imagery, it's the mismatch: a price, discount or claim in the creative that doesn't appear on the landing page. Reviewers — human and automated — compare the two. A "was $89" in the ad with no compare-at price on the page is a straightforward rejection, and repeated ones put the account at risk.
The fix is boring and effective: pull every number in the creative from the live product page, and re-check when you change the page. DropToAd works this way by construction — prices, discounts, variants and stock come off the rendered listing, and review quotes come from real customers, so the creative and the page agree by default.
When should I refresh creative?
Watch frequency and CPM together: when CPMs drift up while conversion rate holds, the audience is seeing the same thing too often, and that's a creative problem rather than a bidding one. Refresh the winning angle with new executions before performance visibly drops — recovering from fatigue costs more than avoiding it.
This is where editable output pays for itself. Reskinning a proven creative — new hook, new opening frame, new colourway — takes minutes if the design is still a design, and takes a re-brief if it's a flat file.
What about video versus static?
Static earns its place for price, offer and social-proof angles, where the message is one legible idea. Video earns its place for demo and problem/solution, where the mechanism has to be seen to be believed.
For physical products, UGC-style video — a person actually using the thing — remains the highest-ceiling format, and historically the hardest to produce at test volume. That's the specific bottleneck AI-generated scene video is good at removing.
FAQ
Questions, answered.
Start with about five, each covering a genuinely different angle — problem/solution, social proof, price/offer, demo, and objection handling. Once one angle proves out, make three to five variations of it, changing one variable at a time.
Creatives. On modern ad platforms targeting is largely automated, so the creative effectively selects the audience. Testing five angles teaches you more than testing one creative against five interest stacks.
Most often because a price, discount or claim in the creative doesn't match the landing page. Reviewers compare the two, so every number in an ad should come from the live product page — and be re-checked whenever the page changes.
When CPMs drift upward while conversion rate holds steady, the audience is over-exposed to the creative. Refresh the winning angle with new executions before performance drops, rather than after.
It depends on the angle. Static works well for price, offer and social proof; video wins for demos and problem/solution, where the product's mechanism has to be seen. UGC-style video has the highest ceiling for physical products.
Yes. Creative can be built from the photos already on the product page, so you can test angles before committing to inventory or waiting on a sample.
Stop rewriting briefs. Start testing angles.
Paste the product page. Get five angles. Fork the winner.
powered by Claude · every pixel stays editable