Marketers have relied on automated messages for years. The concept is not new, an email is triggered based on some action (or inaction) taken by a shopper. This could be signing up for emails, making a purchase, or anything in between.
What has changed is the marketer’s ability to automate messages, based on complex data models and fueled by customer intelligence, that were once only accessible (and understood) by data scientists.
Rather than depending on a call-and-response paradigm, automated messages can now be sent based on predictive models that will anticipate a shopper’s next action, which products may motivate them to buy, and the offer that will seal the deal.
While these advancements burst open the doors to marketing opportunities that were once unimaginable, many marketers will fall prey to some common missteps that lead to automated messages failing, or not reaching their full engagement and sales potential.
Often, it’s not the type of message we are sending that leads to lackluster performance. As marketers, we are aware of the key marketing moments. Exploring product pages, filling the mobile shopping cart, browsing in a physical store, perusing the app. Our shoppers are finding new ways to engage, and following paths to purchase that are less linear than in the early days of e-commerce.
Your automated messages need not only to meet the shopper during these key phases of shopping, but also to speak to the customer in a way that is relevant, maintains the momentum that will lead to a purchase, and anticipates what actions they may take.
Here are three components to use in your campaign planning that will help you to avoid marketing automation missteps.
When do I send my automated messages? This is one of the most frequently asked questions in digital marketing. It’s also one of those questions that has no simple answer. Most marketers will take an unscientific approach and use their gut to determine timing. Cart reminder message? Let’s try 45 minutes after the shopper leaves. Why not?!?
Others will use aggregated data based on what they see from other retailers and competitors. Cart reminder message? All the cool kids are waiting 15 minutes. Let’s try that!
Some may take a more scientific approach, and look at data around their messages to determine baselines and benchmarks. Cart reminder message? Most of our shoppers come back within 12 hours on their own. Let’s send after 12 hours to catch the stragglers.
These strategies may have led to bumps and dips in performance over the years, but all relied on various amounts of human involvement. Perhaps it’s time to use the massive hordes of data you have collected to determine message timing based on the individual user? This is where customer intelligence and machine learning can take the lead and determine optimal timing for automated messages, allowing you to focus on other ways to improve and optimize your marketing.
As you move toward timing messages based on user data and engagement, you will need to consider the content of those messages. While the path to purchase is not always linear, there are moments in the journey that have some predictability.
The theme of certain messages may be defined to an extent; for example, using the cart reminder mentioned above, you would likely want that message to reference the actions the shopper made in the cart. The exact content of the message, however, can be more of a challenge to predict and prepare.
You may be using static content for all recipients, “Thanks for signing up. Here’s 10% off your first purchase.” or some basic dynamic content “Thanks for signing up. Here’s 10% of women’s summer styles.” This can be an excellent way to give these messages some muscle, but there’s more you can do to make the experience even more relevant and engaging.
Take a more predictive approach to the content, products, and offers in your automated messages by combining customer data from all channels and devices. Offer the shopper a truly customized experience by populating content that references their device and channel preferences, and includes offers that machine learning calculates will give you the highest probability for conversion. This may mean in-app product recommendations rather than a cart reminder email.
As you begin to include more machine learning in your automated message strategy, you should also start to shift the way you measure performance. Rather than myopically focusing on how well one particular campaign performed, begin to analyze shifts and trends in your overall shopper behavior.
Are new shoppers progressing to first-time buyers? Are you building loyal customers who come back to buy? Are you finding ways to reconnect with defecting shoppers?
One of the benefits of relying on machine learning is that you are able to spend more time analyzing the broader performance of your marketing efforts. These shifts in shopper trends will help you to identify weak links in your automated messages, without simply focusing on whether a particular triggered email is working.
After all, the inbox may not always be the best way to reach your shoppers. Understanding how and where your shoppers are engaging will help you to determine how you expand your automated strategies beyond the inbox, and build more robust cross-device, cross-channel strategies that keep your shoppers coming back.