The most innovative marketers routinely pick up new things, try them out, and succeed or fail. Why are marketers struggling so much to adapt to artificial intelligence and machine learning? In this series, we’ll explore machine learning and artificial intelligence to build a foundation for understanding the field – and how it applies to marketing.
Types of Machine Learning
Machine learning comes in two varieties: supervised and unsupervised. Both kinds of machine learning are important and have different, distinct roles. Let’s investigate each.
Supervised Machine Learning
Supervised machine learning teaches algorithms and software to recognize something. For example, when you use your smartphone to take a picture of something and it automatically recognizes someone’s face and focuses on it, that is a piece of software that has been trained to recognize faces using machine learning.
Supervised machine learning begins by giving a piece of software a known training library and teaching it to recognize things in that library. For example, we’ve used supervised machine learning as consumers for a very long time in our email inboxes. For years, users have marked messages as spam in their inboxes, and machine learning algorithms have learned through training what is and isn’t spam. Every time we mark an email as spam, we help add more to the training library. Over time, assuming we mark messages as spam or not correctly, the algorithms learn what is and isn’t spam.
How would we use supervised machine learning in marketing? We might teach a piece of software to recognize corporate logos and images. We would give it a massive library of corporate logos with logos marked clearly and each photo for the machine to analyze. Over time, the software would recognize more and more logos in the training library until it achieved nearly perfect image recognition of the known good logos.
Once trained, we would then use images that might or might not have logos in them with the algorithm to identify where our corporate logo was used. We would continue to add new logos to the training library so that the machine would continue to learn and grow its capabilities to recognize logos. We’d end up with software which could recognize our logo if someone posted a photo of it on Instagram and didn’t tag us.
Any situation in which we need to recognize or find something is a situation where we would benefit from using supervised machine learning:
- Logo identification
- Mentions of us in social media
- Sentiment analysis
- Influencer identification
- What makes content rank well for SEO
Unsupervised Machine Learning
In comparison to supervised learning, unsupervised learning is all about sorting and classification. We hand a large bucket of stuff to the machines and ask it to sort and categorize it as many ways as possible so that we can then analyze it. Unsupervised machine learning is especially important in today’s giant data world. We create so much content that there’s no way we could ever read even a fraction of it. Only with technologies like unsupervised learning do we stand a chance of analyzing all the content we produce every day.
Imagine you are sitting in front of a table filled with lego blocks. How many different ways can you categorize those blocks?
- You could sort them by color
- You could sort them by shape
- You could sort them by size
- You could sort them by your own personal preferences
If I put 50 blocks on the table, that would be a manageable exercise. If I put 50 billion blocks on the table, you’d run out of lifetime before you are done categorizing and sorting. However, if we gave an image of that table to a computer, it could use unsupervised machine learning to visually categorize the blocks in minutes or even seconds.
How would we use unsupervised machine learning in marketing? Any task which requires sorting and classification is a candidate for unsupervised machine learning. For example, we might collect all the blog posts written about our company and use natural language processing to find out what topics our company is associated with. Are people writing about us using the messaging that we want to be known for? Unsupervised machine learning would break apart all those posts and transform them into a list of topics.
Some other uses for unsupervised machine learning in marketing include:
- sorting images
- understanding social media conversations about us, competitors, or our industry
- clustering and sorting customers to find our best customers
- digesting the day’s headlines to find trending topics to write about
Integrating Machine Learning Types
We would rarely use only one kind of machine learning; most of the time, we will be doing combinations of supervised and unsupervised machine learning. We might, for example, ingest all the social media conversations about a given topic and classify those conversations into sentiment and topics using unsupervised machine learning. We would then use supervised machine learning to find how many of those conversations are about us.
By understanding the different types of machine learning and their uses, we can better understand how they fit into our overall marketing strategy. In the next post in this series, we will take a look at deep learning and how artificial intelligence maps to our own intelligence.
The AI Primer for Marketers Series
- Part 1: Why AI and Machine Learning are harder for marketers
- Part 2: Algorithms and Machine Learning basics
- Part 3: Types of Machine Learning
- Part 4: The AI landscape
- Part 5: AI-Powered Content Marketing
- Part 6: AI-Powered Conversation
- Part 7: AI-Powered Analytics
- Part 8: AI-Powered Distribution
- Part 9: Preparing Your Career for an AI Future
Be sure to catch my keynote talk about this topic as well.
You might also enjoy:
- How to Set Your Public Speaking Fee
- The Evolution of the Data-Driven Company
- Best Practices for Public Speaking Pages
- How to Set Your Consultant Billing Rates
Want to read more like this from Christopher Penn? Get updates here:
Get your copy of AI For Marketers (2019 Edition)