Multivariate flags are used to experiment with multiple states of a new feature.
It’s easy to think of feature flags as Boolean values only. A feature is on or it’s off.
But what about a situation where life isn’t that simple? What if you want something more nuanced than turning something on or off?
For example, let’s say that you have a “Buy Now!” button. And you want to run an experiment to see what color of button prompts people to click the most frequently: red, yellow, or blue.
Multivariate Flags
To accomplish this, Rollout supports multivariate flags. Instead of a Boolean, you use string values as your flags: “red”, “yellow”, and “blue”. This will trigger traffic to be directed to a screen with each different color of button, allowing you to track results for each one.
You aren’t just limited to strings, though. You can use other data types as well. Feature flagging is more than just turning things on and off. Multivariate flags allow you to run sophisticated experiments in order to learn things about your users and your application.
