By Ryan J Barr, Andrew Williamson, Emily Finnegan, and Carter Andrues
Machine Learning and Predictive Analytics for Digital Marketers and Business Owners
There’s a lot of talk these days in marketing about “machine learning”, “data science”, “artificial intelligence”, and “predictive (marketing) analytics”, but what do they all mean? Are they just buzzwords? No, not at all. They’re incredibly important concepts that you must be aware of and prepare for to stay on top of digital marketing.
Machine Learning is part of the field of artificial intelligence which builds algorithms that help computers learn. Yup, people can teach a system to learn and perform tasks. To clarify a common misconception, machine learning is not artificial intelligence; rather, it is only a type of artificial intelligence.
Showing a return on investment for digital marketing efforts isn’t complicated or difficult, yet it is a massive pain point for many marketers and business owners. In this two-part blog, we’ll introduce you to machine learning, our favorite up-and-coming tool in the marketing world. Then, in part two, we’ll share three ways it can ‘superpower’ digital marketing efforts.
Predicting the Future By Looking at the Past
Predicting the future is hard, but not impossible. It requires large amounts of data. Nobody can pull data from the future, but historical data can be used to make a strong calculated prediction about the future. In other words, making predictions about the future is possible if there is enough data to analyze.
So how does a marketing manager or business owner make better decisions about marketing tactics? Have you ever created a marketing campaign that you were really excited about, but your client didn’t like it? Has your boss ever told you that your idea was silly and that it wouldn’t be appealing to customers?
Marketing by intuition and experience-led decisions isn’t dead; however, marketing has grown up and graduated to a new level. Using data analytics will allow you to prove to your boss or client that your campaign idea will work. Or, you should ask that your employee/marketer provide you with data to prove their campaign will work.
In Other Words…
Think of it this way. Say you and a competitor are competing to sell seats for a rafting trip in your local area and your competitor is using data science; who would win?
Your strategy is for your creatives to design an amazing 1200×628 pixel Facebook ad and deploy it to an audience they believe (aka “hope”) will like it; the ad is targeted to people who have liked your page, and specified “hiking” and “adventure” as interests.
Meanwhile, your competitors analyze past data to see who has been clicking on ads. They look at Google Analytics to see who is visiting their website regularly. They build a Facebook ad targeting a people between the ages 18-34 because their analytics show it’s the group who is engaging the most with their previous ads. Who wins? You, winging it? Or your competitor, using data?
It is likely that neither of you win. You’re using your years of experience and they’re using regular Marketing Analytics. You both get impressions, clicks and a few conversions. Neither of these take advantage of machine learning. We’ll cover machine learning with three examples in part two. For now, let’s break down the differences.
Here Are a Few More Key Concepts
- Marketing is the act of creating, communicating, and delivering your ideas, products, and experiences to customers.
- Marketing Analytics is how one measures, manages, and analyzes their marketing performance. Marketing Analytics measures return on investment.
- Data Science is a field where scientific methods, systems, and processes are used to find knowledge and insight from data in its various forms. Data science is not how big your data is–a 100GB spreadsheet isn’t any cooler than a 100MB spreadsheet. Data science is about answering questions using science, not intuition, feelings, or emotions.
- Predictive Marketing Analyticsis using all of the above to create, communicate, and deliver your story, idea, product, and experiences to the right customers at the right time. With Predictive Marketing Analytics, businesses can maximize their return on investment and massively reduce wasted spend. We’ll get into the proofs for that in part two.
- Note: Throughout this writing “Marketing Analytics” is different than “Predictive Marketing Analytics”.
Don’t Panic! As a business manager or marketing manager, you don’t need to be a master of Marketing Analytics; rather, you just need to be familiar with it. We suggest having a data scientist as part of your organization. Data scientists are often not marketers or creatives; rather, they are the type of people who love .CSV files, speak “R” and “Python”, and get excited about pivot tables in Microsoft Excel.
As business managers or marketing managers, you want to understand enough about Marketing Analytics to communicate effectively with the data science team. Together, you can work as a team to maximize the effectiveness of your marketing efforts.
What’s Everyone Else Doing?
The CMO Survey “…collects and disseminates the opinions of top marketers in order to predict the future of markets, track marketing excellence, and improve the value of marketing in firms and in society.” (In the following paragraphs, don’t get too caught up on how the number of surveyed responses change; we’re reporting the highlights from a massive survey and each question had a different number of responses.)
The August 2017 CMO report found that only 5.5% of firms’ total budget goes to Marketing Analytics (210 firms surveyed). Spending on Marketing Analytics is expected to increase by 229% over the next three years, though.
Out of 213 firms surveyed, only 37.5% of projects used Marketing Analytics to make decisions.
When asked about access to marketing talent(on a scale of one to seven, with seven being “has the very best talent” and one being “does not have the best talent”), 31 of 214 firms (or roughly 14.5%) said rated themselves a one (“does not have the best talent) and only four rated themselves as a seven ( “has the very best talent”). The mean of the 214 firms fell at 3.7, or roughly the scale midpoint—neither best best nor the worst talent.
Okay okay! Math, numbers, statistics! Enough! What does it all mean? It means that to get ahead of competitors, you have to use data to target customers more effectively. To maximize your return on investment in marketing, you have to embrace data as part of your future digital marketing efforts. Marketing by intuition isn’t enough anymore.
To get an edge, empower your marketing with data science. Again, CMO says Marketing Analytics is expected to grow 229%! If this trend is a moving train, it’s already leaving the station. You have to reach out right now and grab the caboose or you’ll be left behind!
In our next entry we’ll examine three ways that machine learning can superpower your marketing efforts. Click here to continue to part 2.
 Raúl Garreta, A Gentle Guide to Machine Learning, 2015, https://monkeylearn.com/blog/gentle-guide-to-machine-learning/
 AMA, Definitions of Marketing, 2017, https://www.ama.org/AboutAMA/Pages/Definition-of-Marketing.aspx
 Vasant Dhar, Data Science and Prediction, 2013, http://cacm.acm.org/magazines/2013/12/169933-data-science-and-prediction/fulltext
 Jeff Leek, The key word in “Data Science” is not Data, it is Science, 2013,
 6. Matt Ariker, Alejandro Diaz, Christine Moorman, Mike Westover, Quantifying the Impact of Marketing Analytics, 2015, https://hbr.org/2015/11/quantifying-the-impact-of-marketing-analytics
 Thomas H. Davenport, A Predictive Analytics Primer, 2014, https://hbr.org/2014/09/a-predictive-analytics-primer