Now that you've decided who can see your experiment, that traffic will be randomly allocated into different variations. You can control traffic in two different ways:
Percentage of visitors included in the experiment - This is the more common way that traffic allocation is used. By default, your experiment will run for 100% of your traffic. You might want to adjust the traffic allocation to a lesser percentage if you’re experimenting on a high-traffic page, and you want to minimize your risk by letting only a small subset of visitors see the variation.Generally speaking, though, the more traffic you can safely include in your experiment, the faster you will reach statistically significant results. If you’re experimenting on a low-traffic page, reducing the traffic allocation may prevent you from getting meaningful results.
Percentage of visitors in the experiment who see a certain variation - By default, your visitors are bucketed into the different variations roughly equally, but you can also allocate different percentages of your visitors to different variations. People often use this feature to serve a winning variation to 100% of their audience.