V
Vitamin A
Guest
Yo, I am currently in the middle of optimizing my systen of Splittesting landers. CUrrently the approach is int he following order by doing simple AB splittesting.
After 2500 impressions there should be enough sifnificant data. I will aim for >85% confidence to cut, cause this allows me to splittest faster and pretty effective. No need for 99% confidence in simple adjustable changes.
While doing this, I am just unsure what I should do with the winning controll campaign. If I choosed a winner for an angle and move over to Splittest images, should I let the controlcampaign with the initial image run too, or just run the lander with the new Image?
Is this wrong to do? I just think, if we have already enough data from the initial Lander, why should we run more traffic to it, the numbers wont really change anymore, right?
Also by splittesting Headlines. There are so many variations, I can come up with at least 10+ of them, how many should I try at one time? ONly A/B with headlines is really hard, since there is a lot of room for variations and being creative.
After 2500 impressions there should be enough sifnificant data. I will aim for >85% confidence to cut, cause this allows me to splittest faster and pretty effective. No need for 99% confidence in simple adjustable changes.
- Splittest Design/Angles
- Images
- Headlines
- Small adcopy change
- CTA
While doing this, I am just unsure what I should do with the winning controll campaign. If I choosed a winner for an angle and move over to Splittest images, should I let the controlcampaign with the initial image run too, or just run the lander with the new Image?
Is this wrong to do? I just think, if we have already enough data from the initial Lander, why should we run more traffic to it, the numbers wont really change anymore, right?
Also by splittesting Headlines. There are so many variations, I can come up with at least 10+ of them, how many should I try at one time? ONly A/B with headlines is really hard, since there is a lot of room for variations and being creative.