Traffic Distribution
Control how traffic is allocated between experiment variants.
Traffic distribution determines how users are assigned to experiment variants.
Sample Rate
The sample rate controls what percentage of eligible traffic is included in the experiment. For example, a 50% sample rate means half of your visitors will be enrolled in the experiment.
Variant Split
Within the sampled traffic, the variant split determines how users are distributed across variants. For a simple A/B test with equal distribution:
- Control: 50%
- Treatment: 50%
You can adjust these percentages to run experiments with uneven splits (e.g., 90/10 for a cautious rollout).
Multi-Variant Tests
Mojito supports experiments with more than two variants. The total distribution across all variants must equal 100% of the sampled traffic.
Changing Distribution
You can adjust traffic distribution on a running experiment. Changes take effect on the next SDK configuration publish.