Before ordering another batch of ads, decide what you need to learn and whether the media budget can support that decision. Producing more files does not automatically produce more usable evidence.
There is no universal number of ads every ecommerce brand needs each month. Product economics, the differences between concepts, purchase volume, and the way the test runs all matter.
Count questions, not exports
A concept is a distinct reason to buy or a distinct way of explaining the product. A resized export is not automatically a new concept. Three hooks on the same demonstration may answer a narrower question than three completely different demonstrations.
Put the question next to each proposed asset. For example: does showing assembly address the hesitation better than showing the finished product? That gives the production team a job beyond “make something fresh.”
If the existing ad appears to be weakening, first use our creative-fatigue diagnostic. A landing-page or stock problem can make the next batch look disappointing for reasons unrelated to the creative.
Keep production and media costs visible
Production pays for the work and rights needed to make the assets. Media pays to deliver them. A proposal should make both visible, along with what the team will review after delivery.
Here is a hypothetical planning example, not a recommended spending threshold. Suppose a team reserves $1,200 for exploratory media and proposes six concepts. An even planning allocation would be $200 per concept. At a hypothetical $40 cost per purchase, that represents about five purchases per concept if performance and delivery followed those assumptions.
That calculation is not a significance test or a prediction. Actual spend may be uneven, purchase costs can differ, and a few outcomes may leave substantial uncertainty. Its purpose is to expose a mismatch before commissioning more concepts than the team can meaningfully assess.
The response could be fewer questions, a longer test, a different budget, or accepting a less conclusive result. “Test everything” is not a plan.
Fill in a decision brief
Use one row for each concept:
| Field | What belongs here |
|---|---|
| Buyer question | The hesitation or reason to buy being tested |
| Creative difference | What actually changes between versions |
| Production scope | Formats, footage, editing, permissions and cost |
| Media plan | Platform, audience, objective, delivery approach and limit |
| Primary outcome | The business outcome used for the decision |
| Diagnostic metrics | Attention, clicks or page arrivals that help explain it |
| Review point | When the team reviews results and what could stop the test early |
| Decision | Continue, revise, stop, or inconclusive, with the reason |
Define the outcome before seeing the result. If the business needs purchases, a strong hook rate is a clue about attention, not a substitute for sales.
Know what kind of comparison you are running
Exploratory delivery can help identify promising ideas. A controlled split test answers a more specific comparison. Do not describe every campaign with multiple ads as a controlled experiment.
TikTok’s split-testing documentation allows a single selected test variable, including creative or creative assets. Compatibility varies; its current matrix lists TikTok Shop as incompatible with that tool. Check TikTok’s split-testing variables and compatibility for the exact campaign before promising that method.
For any platform, make the test design explicit. If audience, offer and landing page change at the same time as creative, the result cannot isolate the creative change alone.
Buy a learning process you can use
Ask a prospective partner how it turns results into the next brief, what happens when a test is inconclusive, and which production costs or rights sit outside the retainer. At BPM, retainers are not calculated as a percentage of ad spend. The scope still needs to explain the work included.
BPM’s creative services and paid social management should fit the same operating plan. Bring your current assets, product economics and available media budget to a scoping conversation. The useful output is a set of questions you can afford to investigate.
