The following examples illustrate the appropriate traffic allocation and variation exposure settings for each use case:
Use case 1: The hero banner
The hero banner of the website is managed with Dynamic Content. There are always 3 promotions running. All 3 are replaced every week.
Traffic Allocation: Dynamic
This is an ongoing, continuous optimization campaign in which variations are updated frequently. Because changes are frequent and you can't wait 2 weeks for a winner, this method yields the best results.
Variation Exposure: Session Stickiness
Because the hero banner is prominent and has a high impact on the visitor’s experience, it’s important to keep the experience consistent throughout the session. It’s less critical that future sessions provide the same experience, and furthermore, if the current variation didn’t lead to a conversion, it’s important to give a different variation a chance.
Use case 2: Email capture
To increase the number of subscribers to a newsletter, you launch an overlay campaign that's served to non-registered users. Uncertain which incentive will perform best, you decide to run an experiment with two competing variations:
Traffic Allocation: A/B Test
The overlay will remain in effect in the long run, so making a conclusive decision is paramount.
Variation Exposure: User Stickiness
Different promotions with conflicting messages can confuse your visitors, and for conversions, make it hard to understand to which variation the subscription should be attributed. Each visitor should, therefore, only see one version.
A good rule of thumb is that A/B tests usually call for user stickiness, and dynamic allocations call for session stickiness.