Using an evolutionary algorithm to optimize a consumer bank’s D2C Campaign
Do certain email delivery times lead to higher engagement?
One of the most common questions that email marketers face is when to send their emails to maximize their open rates, click-through rates, and conversions. There is no definitive answer to this question, as different audiences may have different preferences and behaviors. What time zone are they in? What devices do they use to check their emails? What are their daily routines and schedules? How often do they check their emails? These factors can influence when they are most likely to open and interact with your emails.
You can use tools like A/B testing or split testing to compare the performance of different email campaigns sent at different times. You can also use analytics tools like Google Analytics or Mailchimp to track the metrics of your email campaigns, such as open rates, click-through rates, bounce rates, and conversions. By analyzing the data, you can identify the optimal delivery times for your audience and your goals.
When you have a good understanding of the click-through rates and open rates of your customers at different times, the next step is to create an optimal delivery schedule that maximizes these metrics on a segment without negatively influencing them to unsubscribe — a phenomenon that we call “Fatigue” in email marketing.
In this article, I will attempt to solve this optimization problem using Genetic Algorithms — a methodology that isn’t commonly used in the marketing space.
I understand that Genetic Algorithms can be daunting. We will see how simple Pandas manipulations can be used to implement the core concepts in this algorithm.
I introduced ULFC bank, a fictitious but popular consumer bank, in the article on Next Best Action. In that effort, we created a reinforcement model that would suggest the next best offer to roll out to its mortgage customers based on their past responses. With that in place, now ULFC wants the data science team to recommend an…