Christopher Penn, also known as the “Marketing Ninja,” is one of our industry’s favorite writers on the topic of marketing technology and data-driven marketing. Earlier this year, Penn wrote a thought-provoking article on why learning Python or R could help marketers better understand the scientific mind-set behind using data to inform marketing strategy. Penn believes that mastering data science and the accompanying coding languages will help marketers craft definitive testing and verification processes, rather than operating on guesswork and gut feelings. While we agree that closer proximity to data makes you a better marketer, you won’t need to learn Python and R to operate like a data scientist, as mature marketing intelligence platforms that incorporate artificial intelligence will more easily enable data-driven decision-making.
Penn outlines the following process for applying the scientific method to marketing:
- Ask intelligent questions of data
- Define variables and locate supplementary data sets
- Formulate correct hypotheses
- Design statistically valid tests
- Gather and analyze our data well
- Refine a hypothesis
- Create a general theory
As a result, he advocates learning Python or R — what he calls "gold standards for data analysis and machine learning” — to enable these practices. We agree 100% that the scientific method is a vital skill for marketers to learn. But marketing intelligence can also help you to test hypotheses and analyze performance, most readily in a single-view platform.
This comment from Mark Schaefer (another highly esteemed writer and influencer in marketing) is a great starting point to discuss how AI will impact the marketer's relationship with data and the scientific method:
He took the words right out of our mouths!
One could liken this argument to that of driving a car. Sure, training to become a mechanic will help you understand the car, as learning Python would help you understand your marketing data, but isn't it a more valuable use of your time to focus on just driving the car better? The point is, while marketers would certainly benefit from understanding Python, they can understand and use data without it as long as they have the relevant vehicle.
Let's power through Penn's list using an AI-powered marketing intelligence platform to fuel us. AI can help you identify patterns in performance data that enable you to ask the right questions of the data, define relevant variables, and determine hypotheses to test when activating your next marketing campaign.
Here's a sample hypothesis: "My next search campaign can be improved by increasing the dollars associated with unbranded keywords.” A brand manager who hopes to test this hypothesis would identify historic campaigns that are similar to the campaign in question in order to view the nature of branded to unbranded keywords across campaigns. Finally, she would gather data and use AI methods within a marketing intelligence platform to determine the optimal use of unbranded keywords. In using marketing intelligence technology to apply the scientific method to her marketing strategy, she will have not just a theory, but action-oriented, tactical recommendations to create better performing search campaigns.
Learning code and data science will still be valuable no matter how far technology advances. If you understand Python you have a good foundation for understanding marketing intelligence. If you enjoy coding and running statistical models in Python, marketing intelligence can uncover relationships and patterns visually, so that you have a point of reference to start with. And while marketing intelligence delivers holistic insights on marketing performance, you may prefer to run more piecemeal analysis using Python. For all other marketers who find coding frustrating, marketing intelligence will get you from point A to point B just the same.
(Feature image by Jonathan Simcoe)