How does Dynamic Motivation work?
Dynamic Motivation generates a non-deterministic, realtime-generated response to a campaign prompt, using user attributes (either session or profile data) for the current user. Dynamic Motivation evaluates custom-written rules against user attributes to determine, at run-time, the proper campaign, assets, and message to deliver via “open time” emails, web browser pages, and mobile applications.
Dynamic Motivation allows for various campaign “topics” to be chosen by these rules. These “topics” are essentially full-blown campaigns (each with its own assets, messages, etc.) that can be displayed to the current user based on the previously mentioned evaluation.
Dynamic Motivation uses contextual multi-armed bandit* to determine a particular Variant to use based on the user logged in (if known) and their user attributes, which are used to segment the user into a particular audience segment.
*Multi-armed bandit experimentation is a statistical technique used in online experiments to optimize the allocation of resources between multiple options (or "arms"). The technique involves balancing exploration and exploitation of different options to maximize overall performance or reward. In a multi-armed bandit experiment, each option is assigned a probability distribution that determines the likelihood of selecting that option. The probability distribution is updated based on the outcomes of previous selections, so that the algorithm can learn which options are more likely to lead to higher rewards. The goal of multi-armed bandit experimentation is to make better-informed decisions and maximize the impact of investments in online experiments. Going a step further, in a contextual multi-armed bandit, each option (or "arm") is associated with a set of contextual features that provide additional information about the environment in which the decision is being made.
Why won't I get a winning Variant?
Variants are served across multiple segments, based on different rules sets implementing individual user data–which also varies based on what data we’re able to access, depending on permissions. MAB and CMAB (see question above) constantly rotate Variants in and out of serving as well, always putting forward the Variants predicted to perform best for individual users at that specific time based on previous data. Due to so many factors, and the fact that Variants are served based on data for each specific user, there is no 1 “winner.”
How is Dynamic Motivation content created?
A library of dynamic content specific to your company is created by Persado’s Brand Content Strategists. The content itself is based on 2 inputs:
Relevant topics (as mutually agreed upon, e.g., 'Decreasing Cart Abandonment’')
Learnings from your company’s past Experiments, including Personalized Language Profile (PLP) learnings, if available and applicable.
This library of Dynamic Motivation content is reviewed and approved by your editorial and brand teams before being served to your audience.
Are you optimizing the content served to consumers over time?
Yes. Over time, content that is lower performing will be removed and new messaging with the latest insights will be suggested to be added into the rotation.
How does Persado monitor content/offers to ensure users are receiving the correct content? For example, ensuring users with no loyalty points don't receive a 10% loyalty point discount.
Persado works with you from the start to define any exclusions or inclusions based on data sets.
What are the prerequisites for setting Dynamic Motivation content live?
Prerequisites:
Persado works with you to determine the most valuable topics (i.e., “decrease cart abandonment”).
We determine the data necessary to decide which topic a user should receive (i.e., if a user leaves items in their cart, serve a message to remind them to complete their purchase).
We set up the necessary data feed with you to ensure relevant data is available for decisioning.
We confirm web setup configuration details together and conduct a smoke test.
Persado creates language Variants for each topic for you to review and approve.
Once content is approved, Persado creates preview links to test the content onsite and confirm tracking. The experience will not be live and will only be viewable through a specific link.
After everything is confirmed and approved, the content can be launched.
How does Dynamic Motivation work with Persado Language Profiles (PLP)?
PLP can be used in combination with Dynamic Motivation but is NOT required. If available, the PLP can be used in the Contextual Adaptive Algorithm to determine the right language Variant for a user. However, Dynamic Motivation can run without it.