This new feature will detect the best performing variations in your tests. It will automatically redirect your visitors to the variations with the best results.
Based on the Bayesian statistical model, the dynamic traffic allocation will gradually divert certain profiles of users to their “favorite variation”, in order to optimize conversions.
How to enable this feature?
- In “Traffic management”, tick “Dynamic traffic allocation”
- Choose the criteria for allocation of the traffic (bounce rate, revisit, etc. )
How can this help you?
The Classic traffic allocation allows you to manage the percentage of traffic you want to send to the variations by yourself. For example, choosing a 50/50 distribution. After some time, when you analyze your results, you will probably see that one variation stands out compared to the other one, based on the goals you’ve set. Therefore you will be able to reorganise traffic allocation manually.
By using dynamic traffic allocation, all of this will be done automatically. You don’t have to do a thing. It will recalculate the best performing segments, and adapt the traffic flow accordingly. If one variation’s performance clearly corresponds to the goals you’ve set for your test, the traffic will gradually divert to this variation until it reaches 100%.
- Your goal is for your users to clicks on the “add to cart” button
- On your variation, this button is red, whereas it’s blue on the original version
- The dynamic traffic allocation spots that your variation (with the red button), leads to +40% of clicks. Automatically, a larger part of users will be guided to variation 1 (where they will see this red button).
- Potentially, this will raise your cart conversion rate, and fulfill your sales goals.
Dynamic traffic allocation:
- Enables to allocate traffic automatically, which is smarter, more responsive and follows your tests performances live.
- Has a high reliability thanks to an advanced Bayesian algorithm
- Allows you to minimize your losses and maximize your conversions
- Is particularly adapted to websites with large traffic
To learn more about the statistical model behind this feature, listen to our expert :