Creative testing / OMNI Journal
How many ad variations should I test on Meta or TikTok?
Last updated
There is no single right number. Start with a manageable set of genuinely different ideas that your campaign budget can support, then use the results to decide what deserves another round.
Written by Charles D'Souza (co-founder)
What counts as a meaningful variation?
A new opening, audience problem or product benefit gives you a different reason for someone to pay attention. Changing a caption colour may be useful later, but it does not explore the same range of ideas.
For example, one concept could lead with a product demonstration, another with a practical benefit and another with an objection a customer has before buying. These are possible approaches, not promises that one will win.
How should I structure the first test?
Write down what you want to learn before creating the assets. Keep the campaign objective and measurement approach clear. If you change the audience, offer, landing page and ad at once, it becomes harder to know what caused a difference.
- Give each concept a clear message and opening.
- Use placement-appropriate versions of each concept.
- Agree what success means for the campaign before launch.
- Allow enough delivery to assess results rather than reacting to the first few clicks.
What should the next round change?
Look at the whole journey. An ad can attract attention but send visitors to an offer they do not want, or generate clicks that never become purchases. Use campaign and website results together.
Build new hooks and executions around promising ideas while keeping room for a different direction. OMNI produces variations for testing; the right test size and spend depend on your campaign setup.