Continuing down the “roadmap of exploration” on how artificial intelligence and machine learning may impact marketing, I have discovered 4 videos which serve as a good foundation for thinking about the opportunity. These videos provide 4 important concepts for me as a marketer:
- Which applications of AI in marketing could I focus on first
- What are the right questions to pose when framing an AI-for-marketing challenge
- How does machine learning work and what is an easy way to talk about it
- How does a marketing/tech geek think about AI and machine learning
A.I. for Marketing & Growth - Where do I start?
This 5:02 video from Growth Tribe maps AI application to marketing and business along a maturity curve. That curve outlines which applications are more ready or more accessible for marketers from “must haves” to “nice to haves.” The simple rating system will be different for different businesses and marketers, and it’s a simple way to think about the potential behind AI in marketing. I particularly like David Arnoux’s (@darnocks ) rapid, no-nonsense approach to explain all of the topics from Growth Tribe.
Must Haves:
- Predictive Analytics
- Clustering/Customization
Should Haves:
- Recommendation Engines
- Natural Language Processing
Nice To Haves:
- Psychographic Personas
- Image Recognition
AI for Marketing & Growth #1 - Predictive Analytics in Marketing
Specifically focused on the needs of the marketer, David quickly (3:16) outlines AI’s application to use cases like focusing on leads with the highest ROI and which offer to recommend when and to which customer. Most usefully, the video outlines how to think about an AI challenge via 3 simple questions:
- Which questions am I trying to ask
- Which metrics am I trying to forecast
- Which future behavior am I trying to predict
The 7 Steps of Machine Learning
This 10:00 video from Google Cloud and @YufengG (Yufeng Guo) outlines the basic steps needed to get to machine-based predictions. This is foundational knowledge meant to help non-tech geeks understand the concepts and have a productive conversation with tech and data teams.
The 7 steps include:
- Gathering Data
- Preparing that Data
- Choosing a Model
- Training
- Evaluation
- Hyperparameter Tuning
- Prediction
Machine Learning in Marketing from the Conductor event, C3 2017
In 32:10, Dharmesh Shah, CTO at Hubspot (@dharmesh) offers a way to think about Machine learning in marketing. He quickly answers the question of what is the difference between data-powered and machine learning-powered marketing. His concept of “Autonomous, Self-driving, Marketing Automation” is simple and powerful. He grounds the talk in real marketing use cases like smart lead routing (sending the right lead to the right salesperson) and persona development. He covers some useful basics like the difference between unsupervised machine learning vs. supervised machine learning and makes a great point that success will be about data not the algorithm. He also offers a glimpse of his sandbox for chatbots which he sees as the future of Web sites (growthbot.org).
Comments