Interview with Sandy Hathaway – Co-Founder / CMO, AVARI
Email marketing is becoming extremely personalized with every growing day. Well, thanks in no small part to the availability of information and data. But, a heap of data is of no use. You actually need a sophisticated personalization software and automation tool that helps you achieve your personalization goals easily.
Below is the download of our discussion:
What according to you is a good way to personalize automated email series in this era of email marketing?
Personalization is critical in order for an email campaign to be successful. The reason why it’s so important is because it’s what today’s consumers want. They demand customization because they don’t have a lot of time to waste with irrelevant content.
And when consumers get the experience they desire, they are more responsive to what you have to offer. However, it typically takes quite a lot of work to execute a personalized experience in an automated email series.
You have to define the objective, plan the flow and content, set up the automation and then measure results to see if it’s performing as expected. Many marketers are happy to get that far, let alone think about optimizing content to be more personalized.
A common way to personalize is use first names in the greeting – it’s easy to do and has positive impact, but it’s not enough. A more sophisticated way to personalize is to use certain criteria like gender or zip code to refine an audience into a slightly smaller more targeted group. That also works well, but when you start combining segmentation with automation schemes that need to run indefinitely, the complexity increases dramatically.
That’s why I believe that new methods involving predictive analytics and machine learning are the best underlying technologies for scalable personalization, especially in an automated series.
How do you think predictive content will change the face of email moving further?
Predictive content in email reduces complexity by using a machine to figure out and display the next best content for each person on an individual level, rather than requiring a human to try and think up and program rules for all the relevant scenarios. Not only is it a huge time saver, it’s the most intelligent way to deliver personalization to thousands or millions of people all at the same time.
Furthermore, machine learning is key – a predictive system alone is not enough – it needs to learn from engagement in order to show better content each subsequent time. The reason why marketers should care about machine learning is because it automatically drives content optimization down to the individual level, at-scale. With machine learning, data analysis which used to take days can now be done in just a few seconds in the cloud.
To put it in relative terms, think of AB testing and the drain on resources that can turn into. Jonas Dahl, a business analyst at Adobe, made a great point about this his article “Four Reasons to Lose a Little Control and Trust Automated Personalization”:
“Automated personalization offers ‘always on’ optimization of personalization. It continuously collects data about visitors and the performance of each offer and adjusts the machine-learning models to reflect the most current visitor behavior. This allows it to serve winning offers at any point in time rather than based on data collected over a limited, past testing period as A/B testing does.”
In short, predictive content with machine learning will strengthen the position of email as the number one channel by which consumers wish to receive relevant communications from brands.
Which are some good characteristics of a content recommendation system that uses predictive analytics and machine learning to propel an email marketing program?
Great question. All personalization systems are not created equal. There are some very specific characteristics that one should look for, including:
- The use of implicit as well as explicit data as a basis for the predictions – implicit data enables a 1000x more in-depth understanding of intent
- Hybrid models that use collaborative filtering and content-based filtering algorithms (having both captures patterns in subscriber behavior as well as relationships between content items)
- Real time data capture, continuously updated recommendations, and the speed of delivering fresh content into an email
- Manageable integration requirements in terms of resources and time as well as cost
- An intuitive and easy user interface with quick learning curve
Any final thoughts on personalization?
As email marketers, we know this better than anyone; the content we create is key to establishing connections, engaging our audiences, driving conversion rates, and increasing revenue. We are all working to improve our content and we are all sending more and more email as we compete for share of the inbox.
In reality, the biggest concern marketers face today is the risk of audiences drowning in a deluge of irrelevant content. It’s essential, now more than ever, that the content we create and distribute is relevant. That’s the only way to cut through the email clutter and be a memorable brand. I believe predictive analytics and machine learning are not just an important way, but the only way forward.