ML for Marketing Tips Part 1: Building the Business Case for Machine Learning for Marketing

The potential benefits of using ML capabilities are vast. How do you know where to get started? This is part 1 of a 2-part series. Read Part 2 on Selecting Use Cases here

This article series was jointly written by Elizabeth Marshall, Director of Solution Consulting, and Katie Jarrell, Senior Customer Success Manager, with supplemental editing by Matt Parisi, Sr Product Marketing Manager.

Machine Learning is an emerging capability that helps organizations drive key business outcomes. For example, consider some of the outcomes ML is likely driving even in your own life: 

• 75% of what we watch on Netflix
• 35% of what we purchase on Amazon

ML for Marketing is coming of age, and it’s time for marketers to have an ML strategy. This means deciding on technology, processes and use cases with a strategic approach that delivers customer experience value now and into the future as it becomes more commonplace.

While it’s easy to say that ML can transform your marketing efforts, it’s less easy to say exactly how. According to Gartner, the #2 challenge to Machine Learning adoption is simply an understanding of its benefits and uses. That’s why it’s critical to ground your ML strategy in a well-documented business case.  As you read this article, the question to ask yourself is what Business Outcome am I looking to achieve? Then you can align Machine Learning strategies to help your organization achieve it. 

How to Build the Business Case 

Why is machine learning needed in your organization and what benefits will your business derive from it? There are 4 foundational considerations to document before we start putting together specifics of the use case; Justification, Alternatives, Benefit, Risks.

Justification – Define your business problem or opportunity. It is best to connect this to an overarching business initiative.

• What is your business goal?  
• How are you trying to achieve it?

Example: We want to optimize engagement, conversions, and customer retention through data-driven targeting and analysis 

Options (Alternatives) – Identify alternative approaches you may consider and rate them based on identified drawbacks. Gather information about each and analyze. 

For our new ML feature, there are a number of advantages for your ML strategy that you should consider compared to other options (you can watch a demo of our new ML feature here: https://tealium.com/products/tealium-predict-machine-learning/):

• Other ML tools perpetuate data silos. Is this an issue for your organization? It can lead to inefficient operations and uneven customer experience.
• Are you concerned with black-box ML approaches giving no visibility into insight strength or attribute weighting?
• Does the solution enable less technical resources or does it require a statistics degree?
• Do you want to configure your insights or just use an out-of-the-box score that may not be tailored to your business?

Benefit – What is the value returned to the business? Generally for marketing use cases, we will be looking at benefits around revenue, cost savings, and efficiency.

These could be things like:

• Grow revenue by using ML insights to personalize the customer experience
• Drive operational efficiencies by automating segmentation with ML
• Deliver cost savings by using ML insights to suppress poor targets

Risks – What are the risks involved and what are the consequences of the risk actualizing?

• Resource investment— ML efforts require commitment both in resources and technology
• Wasted opportunity— You have already done the hard work of collecting data from your various customer touchpoints, standardizing and enriching it. Not scaling the value with ML is a missed opportunity and potentially opens you to competitive disruption if you don’t mature capabilities soon enough
• Too ambitious with a new capability— Is there a way to start small that can still grow large? With Predict ML, we give our AudienceStream customers 1 free model with 5 trainings so there’s no risk to start, and there’s a path to expand
• Competitive needs— Will getting started now help you differentiate customer experience now and into the future?

Once you have built your business case for using ML for marketing, you can move on to specifically what you would like to do with it.

Continue on to read Part 2 of our ML for Marketing Tips Series, How to Pick Your Use Cases