How many variants do you launch?
The documented 2025 methodology launched 12 variants across 3 angles, then ran a second round of tests. Another case study launched 20 campaigns on day 1. The right count depends on the market and question.
What does each variant test?
A meaningful variable: the angle, offer frame, audience, or wording of the ask. The goal is to learn which principle drives the response, not to create random copy volume.
When do you stop testing and scale?
When positive replies exceed 12% of total replies. Below that gate, the right move is another sprint — not more sends.
What if the first angle is wrong?
That is useful information early. The sprint gives the team a way to change the frame before a weak message consumes the campaign's volume.