Wayne Lockwood: Why Decision Trees Are Terrible for Omni-Channel Marketing

If you’re in B2C marketing—especially email marketing—you’ve likely seen an incredibly complex decision tree at some point.

Decision trees are often the web of interconnected steps that define how buyers receive email marketing messages. If this, then that. If this other option, then that other option. The more complex the decision tree—the thinking goes—the more “personalized” your marketing strategy.

But personalized customer experiences need to be led by data, and many marketers struggle to create truly personalized campaigns because of siloed customer data. In fact, 31% of senior leaders believe integrating customer data is the greatest challenge their company faces.

Decision trees and flow charts may seem like a way around this challenge, but they can quickly become overwhelming and confusing. Most decision trees are built on assumptions and gut feelings rather than data. And because they’re nearly impossible to measure, optimization is a significant challenge.

Even worse, these legacy workflow diagrams are almost impossible to integrate into a full omni-channel marketing strategy. A shocking 86% of ecommerce marketers still have not implemented a full omni-channel marketing strategy because of the same inability to access and utilize customer data effectively.

Put simply, decision trees make it difficult to effectively measure and optimize B2C marketing.

In this blog, I’ll cover four reasons you should ditch the decision tree and improve your chances of truly connecting with customers—no branches required.

Out-of-Control Complexity

Every marketer remembers their first decision tree. You probably initially created it as a simple flowchart with a few branches you could easily use and decipher. But then, as you considered each option and all the different outcomes, the flowchart slowly became bigger, more complex, and unwieldy.

For example, here’s the logic you might think through as you create a decision tree for your buyers:

If a buyer doesn’t open an email in 3 days <then> send them an email on a different topic entirely.
If they opened the email but didn’t click <then> send them an email with a different CTA.
If they respond to that message <then> you add them to a re-engaged customer flow that lasts for one month.
…and so on.

It takes a lot of time and effort to determine this myriad of outcomes and decide where and when to split up your audience based on their behaviors, demographic information, and more. Some of those decisions may be based on logic, but others may be based on an idea you think may work.

And as your decision tree grows in complexity, it becomes increasingly difficult for other people on your team to understand it. Since only you really get the logic behind it, it’s complicated to help others wade through the many branches and explain why you chose that path. The general rule is that the more decisions you include in a tree, the more difficult it is to understand and the more your own biases and personal preferences can slowly creep in.

Lack of Cross-Channel Visibility

Not only can decision trees be tough to manage, they’re also incredibly limited. Every B2C marketer realizes that email alone is not enough to engage buyers and drive sales. You have to be where your buyers are—including social media, Google search, and more. If you want to implement a true omni-channel marketing strategy, you need to manage campaigns across channels so you can deliver a seamless experience with consistent messaging.

Unfortunately, decision trees rarely provide the necessary visibility for omni-channel. With decision trees, you’re often restricted to email alone or forced to create entirely separate paths for individual channels. While your email campaigns may be doing well, they might perform even better if you swapped out a transactional email for a push notification or a message on Facebook messenger instead. But if you’re relying on a decision tree, you’ll probably never be able to test that hypothesis. Without cross-channel visibility, decision trees can’t help you experiment and optimize campaigns.

Less Accurate Analytics

Experimentation and testing is also a huge part of B2C marketing and a huge miss for decision trees. You need a sizable dataset to be able to accurately come to conclusions and make decisions. However, the more branches you add, the fewer people there are in each branch, and the less data you have to analyze. You’ll eventually reach the point where the data set is too small to reach any reliable conclusions.

And because you can’t easily create campaigns across channels, you also can’t quickly measure and A/B test those campaigns either. As a marketer, you want to make informed decisions about which channels are performing well, which campaigns you should focus on and which need a change of strategy.

For example, you can’t understand:

The conversion rate of that push notification message;
The open rate of the latest email campaign; or
How that conversion rate changes if you swap the order of the two steps.

With a decision tree, you have no way to measure a campaign like this because it includes two different marketing channels and requires a large data set.

Omni-Channel Campaign Management Eliminates Decision Tree Complexity

Instead of building and managing campaigns in a decision tree, marketers should rely more heavily on segmentation, dynamic content, and machine learning to simplify the campaign flow and more easily measure its impact. Instead of complicated branches for each individual marketing channel, marketers should focus on coordinating cross-channel touchpoints in one simple “trunk” campaign. And performance metrics should be visible in the same UI where marketers create campaigns—instead of many separate branches.

For example, a marketer should be able to create a single campaign flow that includes:

An initial email
A related push notification
A similar paid Instagram ad
A final email message

And instead of having a separate “branch” for each new decision, the work of personalization is handled by machine learning, dynamic content, and dynamic segmentation. This makes your life as a marketer far easier and simplifies how you measure marketing results. By streamlining multiple touchpoints in a single UI, you can focus on more strategic decisions about the campaign flow and leave the optimization of smaller, more tactical choices to automation.

Decision trees may have been a fact of life for marketers for years, but they haven’t adapted to the challenges and opportunities B2C marketers face today. By keeping it simple with an omni-channel campaign management approach, you can leverage customer data to create personalized campaigns that can be managed and measured across a variety of channels.

The post Why Decision Trees Are Terrible for Omni-Channel Marketing appeared first on Marketo Marketing Blog - Best Practices and Thought Leadership.