Dropshipping Ad Testing Budget Calculator

Setting it out

A test budget has to buy enough conversion events to distinguish a working product from a failing one. Required spend = target conversions × cost per acquisition, where the target is set by how confident you need to be.

For a rough read, five to ten conversions is enough to see whether a product converts at all. For a decision on scaling, considerably more.

Budget per test = expected CPA × target conversions, with the expected CPA estimated from the category rather than hoped for.

A concrete case

At an expected $12 cost per acquisition, seeing ten conversions costs $120 if the product performs as hoped, and a failing product will spend the same $120 producing two.

That asymmetry is the point: the test spends the same either way, and what varies is what you learn.

A budget of $150 per product across ten products is $1,500 to find perhaps one or two worth scaling. At a 10% hit rate the cost of finding one winner is the full $1,500.

That winner then has to generate $1,500 of contribution, 262 orders at $5.72, before the round is even, which is why testing discipline matters more than campaign optimisation.

What the number hides

Very small budgets do not produce small results; they produce unreliable ones. Ad platforms need conversion data to optimise, and a budget below the level that generates it will underperform in a way that says nothing about the product.

Judging a test in its first day also rejects products that would have worked once the platform's optimisation had data to learn from.

Where to go from here

Set the budget in conversions rather than dollars, and run each test long enough for the platform to exit its learning phase. Cutting early is the most common way a viable product is rejected.

Then apply the stopping rule mechanically. A test that has spent its budget without producing the conversions is finished, regardless of how promising the click-through rate looked.

Testing creative separately from product

A failed test can mean the product has no demand or that the advertisement was poor, and those require opposite responses. Conflating them means abandoning good products and persisting with bad ones.

The way to separate them is to test several distinct creative approaches on the same product before concluding anything. If none of three genuinely different angles produces engagement, the product is the problem; if one does and the others do not, the product works and the creative was the constraint.

That costs more per product tested and produces far better decisions, which is why experienced operators test fewer products more thoroughly rather than many superficially. The budget is the same; what changes is how much of it buys information rather than noise.

Record the result of every test, including the ones that failed and why. Operators testing continuously accumulate real knowledge about which angles, price points and categories work for them, and that knowledge is the only durable asset the testing produces.

Without the record, each test starts from the same place as the first one, which is why some operators improve their hit rate over time and others test for years without getting better at it.