Skip to content

How much should I spend testing a dropshipping product?

adsUpdated 2026-08-18

Short answer

$50 a day for five days, so $250 to a first verdict. At an $18 US CPM that buys roughly 2,800 impressions a day, which is enough to read cost per add-to-cart by day 3. Below $25 a day the data cannot distinguish a bad product from an unlucky week, and under $400 total you are better off running organic.

Testing budgets and what each one actually buys you. US CPMs, outside Q4.
Daily budget5-day spendWhat you can readVerdict
$10$50Almost nothing. Under 800 impressions a day.Do not. Save it and run organic.
$25$125Cost per add-to-cart, roughly. Purchases are noise.One shot. Test a single product only.
$50$250CTR, cost per ATC and a usable cost per purchase.The standard test. Use this if you can.
$100$500Creative-level winners emerge, not just campaign-level.Faster answers, three products in parallel.

Testing budgets and what each one actually buys you. US CPMs, outside Q4.

The standard test: $250 over five days

One Advantage+ Shopping campaign, broad targeting, $50 a day, four to six creatives against a single product. Every toggle in that campaign is listed in Advantage+ Shopping settings. That is the test. It costs $250 and returns an answer you can act on.

Why $50: at a US CPM of $12-$25 it buys roughly 2,000-4,000 impressions a day. Across five days that is enough traffic for cost per add-to-cart to mean something, which is the metric that tells you whether the product has a pulse before purchases accumulate.

What you judge, and when

Day 3 — read the top of the funnel. Outbound CTR and cost per add-to-cart. Kill individual creatives below 0.8% CTR. Leave the campaign itself alone.

Day 5 — make the decision. No add-to-carts after $150-$200 of spend means the product or the offer is wrong, and no targeting change fixes that. Move to the next candidate.

Days 6-10 — iterate the winner. Let winning creatives spend and add two new hooks against the same product. Judge on gross profit per order, not ROAS in isolation.

The 72-hour rule

Every edit resets the learning phase and throws away the data you just paid for. New campaigns look terrible for 48-72 hours — that is the platform calibrating, not the product failing.

If you cannot leave it alone, launch on a Friday and delete the app from your phone until Monday. This is not a joke; it is the single most common self-inflicted wound in paid acquisition.

When your budget is smaller than $400

Run organic instead. A half-funded paid test produces an ambiguous result, and ambiguous results are worse than no result because they cost money and teach nothing.

The organic route is 3 TikToks and 2 Reels a day for 30 days on one product. It takes 15-20 hours a week and 4-10 weeks to a first sale — the full plan is in the zero-budget section of the ad blueprint.

Budget for the second product, not just the first

Roughly one product in five works. A $250 test that fails is not bad luck, it is the base rate. If your entire budget funds exactly one test, you are betting the business on a coin flip you lose four times out of five.

The realistic total for a proper attempt is $1,200-$1,500, which funds two or three tests plus the store, samples and creative. The full breakdown is in what it costs to start.

Before you launch anything

Run the product through the break-even calculator. If break-even ROAS is above 1.8, the test will fail regardless of how good the creative is, and you will have spent $250 confirming a spreadsheet error.

Related questions

How long before I judge a Meta campaign?

Three days minimum, five for a decision. Every edit to budget, audience or creative restarts the learning phase, so a campaign judged on day 2 is being judged on the platform still calibrating. Write the launch date down and do not open the ads manager until day 3.

Should I test five products at once with a small budget?

No. Splitting $250 across five products gives each one $50, which is below the threshold where any signal exists. You end up with five ambiguous results instead of one clear answer. Test one product properly, and keep the other four in a list for later.

This is one question out of a much longer guide. The full breakdown lives onthe full ad blueprint, and you can run your own numbers in thebreak-even calculator.

Keep reading