Like any buzzword, analytics can often be overused and hard to define from an exact source. According to the Oxford Dictionary, the textbook definition of analytics is “the systematic computational analysis of data or statistics, in order to describe, predict, and improve business performance”. Gartner defines it more broadly as “statistical and mathematical data analysis that clusters, segments, scores, and predicts what scenarios are most likely to happen.”

Analytics is commonly known to branch out into four pillars or areas: descriptive analytics, diagnostic analytics, predictive analytics, and prescriptive analytics.

In essence, analytics is the act of extracting meaningful and actionable insights from data by using a set of techniques and tools paired with domain knowledge. Raw data, however large it may be, will not be a silver bullet in your quest for insights in marketing. Neither will advanced techniques and a lack of domain...

The quote at the beginning of this chapter illustrates one of the fundamental questions of the marketing manager in their day-to-day activities. The best way to evaluate where to spend and target their efforts to achieve their ultimate target is to obtain new customers or retain current ones.

Marketing analytics is nothing more than the application of analytical methods to said goal, bringing a quantifiable way of guiding investment or consumer targeting decisions. As with any new and growing domain, it is hard to pin an exact definition of it, but we can define it as a “technology-enabled and model-supported approach to harness customer and market data to enhance marketing decision making”. Being a domain in the larger field of data analytics, it looks to use mathematics and statistics together with computational tools and techniques to find meaningful patterns and knowledge in data. In this book, we will strive to focus on only...

As we have seen earlier, analytics is a broad term covering four different pillars in the modern analytics ladder. Each plays a role in how your business can better understand what your data reveals and how you can use those insights to drive business objectives.

The following diagram will help you visualize how the pillars relate to one another:

Figure 1.2 – The analytics maturity model

The first step in the process is to always understand the fundamental questions you are trying to answer. All analytical questions can be boiled down into the following categories:

• What happened and when did it happen?
• Why and how did it happen?
• What will happen in the future?
• How can I make something happen?

These categories will define the different areas of analytics involved, which will inform our decision about what tooling and techniques to apply.

Analytics can be split into four areas...

You might wonder why we need a book on Marketing Analytics Using Python. Surely you can do the same thing using the trustworthy combination of Excel, some VLOOKUPs, and some PivotTables. This is a widespread misunderstanding, and the problem stems from not realizing what the entire analytical process should look like and why. The following diagram shows the process in a simplified way:

Figure 1.3 – Analytical process

As an analyst, you should have the preceding workflow that will generally go through the following tasks:

1. You should, first and foremost, scope out the question. You need to understand what is being asked of you clearly. Remember that your stakeholders have immense business knowledge and a problem they need to solve, but more often than not, the question might not be clearly defined.
2. You must extract the correct datasets to explore the problem space. This might be as easy as extracting a CSV from...

Python offers a marketing analyst many benefits. First, it is an easy but powerful programming language with a great ecosystem of tools and libraries for data analysis and statistics. Second, as a programming language, it is easily testable, and the code can be made in such a way as to be generalizable and reusable. Do not underestimate this second point. Reusability is a great asset to have. You can reuse them for other datasets or testing purposes, which will massively increase your productivity in the medium to long term. Third, it handles massive amounts of data with modern libraries such as pandas and NumPy. The limit is essentially the physical memory in your machine.

Some of you might wonder, “Why not R?”. It is a matter of personal preference. Most marketing analytics was derived from the field of applied econometrics. R is one of the prime tools in econometrics and statistics. As a language, it was built for statisticians who did not want to learn...

Marketers and marketing analysts have had the chance to swim in a world of data in the last 20 years. In fact, that was one of the main drivers of the spread of marketing analytics in the field, especially in the area of digital marketing, which accounts for almost two thirds of all marketing spending worldwide. However new trends in the attitude toward privacy and tracking online are making it harder for us, as analysts, to quickly derive insights from available data.

As we can see from Figure 1.4, for years the trend was clear. The largest proportion of marketing budgets went to online and digital marketing:

Figure 1.4 – Evolution of digital marketing spend as a percentage of global marketing spend

At our disposal, we had highly granular and vast datasets on our users, encompassing their behaviors, which ads they saw, when they saw them, and how they reacted to them. Clickstream data...

When moving past toy examples, data wrangling and transformation is neither easy nor something to be taken lightly. As described, since most digital marketing spending and interactions are of a digital nature, you are essentially swimming in a sea of data. Your job as an analyst is, as described at the beginning of this chapter, to generate insights in a timely manner. Here, efficient data accessibility is going to become a topic, especially if you work for a larger company with multiple large data sources. While a deeper and thorough walkthrough of data engineering is outside of the scope of this book, as an analyst, you should have a basic understanding of what it is and why it matters.

Don’t moonlight as a data engineer

Excel gives us a bit of a bad habit as analysts: we use it simultaneously as a database and a tool to generate insights. The use of it as a database gives us a false sense of understanding of the need for...

In this chapter, we delved into what we mean by marketing analytics. We broke down the types of analytics—from descriptive to prescriptive— discussed the value they add to businesses, and learned what questions each of them answers.

We investigated the fundamental questions you are trying to answer—what happened, when did it happen, how did it happen, what will happen in the future, and how can I make something happen. We also covered how they relate to each sub-domain of analytics, and we can now distinguish clearly between descriptive, predictive, diagnostic, and prescriptive analytics.

First, analytics is a complex field that can be summarized as the intersection between statistics, computer science, and analysis. You need to understand that marketing analytics distills those tools and techniques to the efforts of marketing to better optimize spending and obtain a return on investment.

Analytics can be split into question categories that map...

• https://www.lexico.com/en/definition/analytics
• https://www.gartner.com/en/information-technology/glossary/analytics
• Performance implications of deploying marketing analytics, Germann, Frank, Lilien, Gary L., & Rangaswamy, Arvind, 2013. International Journal of Research in Marketing, Elsevier, vol. 30(2), pages 114-128.
• https://en.wikipedia.org/wiki/Prescriptive_analytics
• https://www.forbes.com/sites/forrester/2021/06/25/google-delays-the-cookiepocalypse/