Content Intelligence Blog

5 min read

A MarTech buyer’s cheat sheet to understanding AI lingo

Dec 12, 2018 9:26:07 PM

Yes, this is another article about the meaning of artificial intelligence (AI) in marketing.

No one needs to look far to find examples of AI applied to marketing, but as a marketer, what new tech is worth your time and your company’s dollars? I want to share some personal tools that help me interpret what I read about MarTech solutions “powered by AI.

AI is easy to find yet harder to understand and even more difficult to define. The Google news feed on my phone feels a little like my microwave, I don’t really know what is happening behind the scenes and sometimes the suggested articles don’t come out exactly how I would expect, but I know if I swipe right on my phone, it is a super speedy and no-mess way to get some news updates.

Google has developed an effective way to inventory, categorize, and label content and then consistently recommend relevant content to me based on my past browsing behavior and the behavior of others like me.

As marketers we are not only consumers of AI. Increasingly, we rely on AI for our own marketing activities.

Finding meaning in the jargon

Artificial intelligence

So many definitions swirl around AI. Eek! My preferred definition for AI in marketing technology is paraphrased from one I’ve heard used by Paul Roetzer of the Marketing AI Institute:

AI is technology that automates a task previously done by a person.

Every time you see AI in the context of marketing technology, simply substitute “Intelligent Automation.

These days intelligent automation in marketing comes in two flavors:

  1. Making a recommendation about a marketing activity — for example, software that serves up content topics for blog posts or email subject lines
    or ...
  2. Automating a routine task — for example, programmatic ad buying or journey-driven email campaigns

When it comes to AI, a recommendation is another way of saying a prediction. A piece of marketing software will recommend the action predicted to lead to the best outcome. Recommendations are the stepping stones that allow software to automate routine tasks.

When a piece of marketing software is able to fully automate a routine task, it is usually because:

  • The goal of the task is unambiguous, i.e., two different people would come to the same conclusion about what is to be accomplished.
  • The steps to achieve the goal follow a set of rules or can be reliably predicted.

Some marketing activities are better suited for total automation than others. As the quality of available data improves, and as marketing technology advances, the range of marketing activities that can be automated will grow.

Data science

If you talk to someone who actually builds AI for marketing technology, they may be more comfortable talking about their work as Data Science, rather than slapping the AI label on everything.

Data Science is the technology that enables artificial intelligence. In other words, data science is the umbrella term that encompasses all the techniques for collecting, analyzing, and applying data to intelligently automate tasks.

When it comes to applying AI to a specific marketing task, there are three components of data science that are important to understand:

  1. A consistent, repeatable process that can be measured to act as the information input. This may be referred to as the Data Source, customer behavior, measurable activity, or a variety of other names.
    Example: An email being delivered and then opened is a consistent, repeatable process that can be measured. If we wanted to recommend email subject lines, then we would want to measure open rates for past emails with a variety of subject lines.
  2. Access to a large number of observations from a specific process. This may often be referred to as Big Data.
    Example: Credit card companies use credit card transaction records to provide a large number of observations to power their fraud detection technology.
  3. A specific approach to identify patterns in data. Much of the work at the frontier of Data Science is in developing new approaches for organizing and analyzing unstructured data such as written content or making predictions about future data based on past data. These days it is most common to bundle the many available techniques under the umbrella of Machine Learning.
    Example: Image recognition uses machine learning techniques for pattern recognition in order to tag individual elements within an image so the image database can be searched using keywords.

Some have described Machine Learning as teaching machines how to think or learn like humans. That is a long-term goal of some artificial intelligence researchers. Today, as AI is applied in marketing technology, it is more appropriate to think about Machine Learning as a set of analysis techniques using machines (i.e., computers) to help us (i.e., humans) learn something specific about the world.

One reason it is difficult to build machines that learn and then react like humans is that we are great at processing many different types of data hitting us at the same time to assess a situation and take action. In contrast, computer applications shine at processing an enormous quantity of similar data in order to find the patterns for a particular repeating task.

Our online behavior is frequently repeated and easy to measure. This makes it ripe for AI. This is especially true for AI in marketing as consumers start or complete more shopping journeys online.

Marketers themselves similarly plan, create, execute, and measure marketing activities in front of a screen. Even though work in AI goes back decades, I believe we have reached a tipping point for AI technology now because so much human activity is being consolidated around easy to measure activities on our digital devices.

MarTech decoder

To break down the next flashy article or ad for a new marketing technology powered by AI, ask yourself the following questions:

  • Do I understand which marketing task is being intelligently automated in the new product and is that marketing task relevant to my day-to-day responsibility?
  • Does the new technology have access to its own source of Big Data in order to be helpful to me, or will I be responsible for providing a large source of data?
  • If I am responsible for bringing my own data, what will it take to access and connect this data to the new marketing technology?
  • What evidence do I have that the new technology makes good recommendations or effectively automates one of my tasks?

The answers to these questions will help you assess the value of the new technology without getting caught up in the AI tornado.

If you can’t answer these questions yourself but are still intrigued with the technology, request a demo or speak with a sales rep. If they can’t answer these questions in plain English it may be too early to take the leap.

On the other hand, when you find a technology that can demonstrate its value in your day-to-day life, you likely have a winner.

Download our eBook, "Marketing Gets A New Direction With Artificial Intelligence," to learn more about how you can put artificial intelligence to work in your content strategy.

Bart Frischknecht, PhD
Written by Bart Frischknecht, PhD

Vice President, Product Strategy, Vennli
Bart is all about building marketing technology to help business leaders achieve growth goals. He is passionate about using data to put customers’ needs and choices at the center of strategic decision making. Bart’s background is a blend of design, marketing, and engineering, which provides a unique perspective on a company’s role to create, communicate, and deliver value to its customers.

Post a Comment


Subscribe to Our Blog

Stay up to date with the latest and greatest marketing and content planning tips, tricks, and news.