This article was written by expert industry mentor in Marketing Analytics and Insights- Jack Golding
There are many paths to becoming a marketing analyst. My personal one was through mathematics. I chose to study math because of how powerful it is as a language for problem solving. A mathematician describes a problem, then using tools like calculus to produce it's solution. In first year maths, you solve a problem and then sign your work with quod erat demonstrandum (QED), Latin for "which had to be demonstrated." Q.E.D is a fancy way to say you proved something. The real world is not so elegant, but taking a logical approach can help us provide solutions.
An example of this in digital marketing are users. We want to know how many people engage with our websites, but we can't count them like we could in a retail store. We use tools like Google Analytics to represent a page loaded on a browser as a person visiting our website. Then we can collect data and begin to answer trickier questions. A famous quote from statistician George Box is "All models are wrong but some are useful."
Very few marketing analysts come from a technical background. In my experience, most study marketing and were not afraid of spreadsheets. This opens the door to the significant amount of data that businesses generate every day. One accesses this data through a myriad of different software and tools. The issue is these are so complex that we don't learn about the underlying model and its limitations. This can lead to awkward remarks from stakeholders. An example is how can our website, have more unique visits in a month then there are Australian people?
Artificial Intelligence throws another spanner in the works. Tools are becoming easier to use but harder to comprehend. The more products that vendors create, the further we move away from the underlying model.
Which leads to the ultimate question - how can you prove that viewing ad resulted in a sale? Analysts call this attribution. In most organisations allocating marketing budgets is a political decision. All the math in the world can't help us here, but interpretability can. A great tool we now have for interpretation is Large Language Models (LLMs.) Marketers now have access to a wealth of knowledge like when Web Search was first launched. LLMs are able to marketers work through an understanding of this complex field within the context of their own business. This excites me a lot both as a practitioner and an educator.
Understanding the underlying models we use in digital marketing is paramount. Without this understanding we cannot communicate our results and get stakeholder buy in. AI will create limitless opportunities for marketers in the future. For now, the best use of these tools is further developing their own understanding and taking the business along for the ride.
If you’re interested in learning more about Marketing Insights and Analytics, we offer our short course here: https://online.rmit.edu.au/course/sc-marketing-analytics-and-insights-dmk201