You Ask, I Answer: What is Data Science for Marketing?

In today's episode, Manina asks, What is data science and why does it matter in marketing and content marketing specifically? Now, this is part one of a seven-part series from uh the content tech summit that I'm doing in uh April. So if you have not uh had a chance to stop by there and uh hopefully see you in April, you should register to attend that show. Uh data science is the practice and profession, it's a profession of deriving meaningful insights from data using scientific methods. Now, that's a really important statement because a lot of what people think is data science tends not to be.

It tends to be uh something related, analytics, for example, or statistics or coding. Those are all components of what data science is, but are not data science by themselves. Analytics, for example, is the extraction of uh or the analysis of data. What happened? What does the data tell us?

It's important, right? We're not saying any of the things are unimportant, they're very, very important. Uh reporting is the uh communication of data to somebody else. Engineering, data engineering is the uh profession of trying to make data easy to work with, easy to store, easy to access. Statistics, of course, is the mathematical field, it's an entire field on its own, uh, that you use in data science, but it is itself not data science.

And artificial intelligence and machine learning, again, use some of the very similar concepts, statistics, coding, etc., but that is for the development of software. AI and machine learning, their fundamental output is software. Data science, the fundamental output of data science is scientific inquiry. It is uh it is new knowledge that you've created from an investigation. And so data science has four key components.

Number one is business skills and acumen. You have to be in tune with business to understand what it is that a business is trying to solve, what kinds of problems that you're trying to solve. And it doesn't have to be in marketing. It can be in finance, it can be in operations. If you're a scientific company, it can be about the science itself.

A data scientist can operate in any profession. It is uh in any industry, it is a horizontal, if you will. The second is statistical and mathematical skills. You have to know statistics, you have to know probability. You probably need to know some linear algebra.

Um, but you need those mathematical skills uh as a data scientist because that understanding of the mathematics, particularly of statistics, is vital to knowing whether or not your scientific experiments are valid or not. Uh are they reproducible? Are they um uh are their error rates correct? Things like that. The third area is the technology and the coding skills.

This is being able to work in a statistical language, uh, R, uh, Python, SPSS. Uh take your pick, but there are a number of different technologies that you need to be comfortable with as a data scientist because you're doing science on data, and the data is gonna live in all kinds of crazy places. You need to know some some SQL, uh, SQL uh programming to access data out of relational databases. You probably need to know some JavaScript to deal with JSON, the JavaScript object notation language. So there's all these technological things that are part of being a data scientist.

But the most important, the most important are scientific skills, understanding and applying the scientific method, and it is this which sets apart data science from everything else. It's literally in the name, as I said, science on data. And science is all about discovering new things, right? It's about proving new things. The scientific method is all about asking questions, establishing a hypothesis, which is a provably true or false statement, and then the system of inquiry that goes around proving a hypothesis true or false.

So a real simple example would be that this would not be data science. Hey, our our conversion rate optimization is lower. Let's let's improve our conversion rate optimization. How do we improve it? Okay, that's that's a very open-ended question.

That is not a provably true or false statement. On the other hand, if you went into your website and said, based on the analytics, based on our exploration of the data, we believe passive voice is the problem. The hypothesis would be converting the website from passive voice to active voice should improve conversion rates by at least one percent, right? That would be our hypothesis statement. And then you would conduct experiments to prove that true or false.

You would convert some pages, not convert other pages, or use optimization software like Google Optimize to show different versions of a page with different language to people and measure the conversion rates. And that is that science, that scientific background that makes data science different than analytics or reporting or statistics or AI, even that hypothesis you can prove true or false. And once you've proven it true or false in a reproducible, uh rigorous way, you now have a working theory. And that we use that word very carefully because it is the scientific meaning, not the general public meaning. A general public meaning is a theory literally means to most people a guess, right?

Which is not what it is. A theory in science is this has been proven, it is open to interpretation, so it is open to further experimentation with a refined hypothesis, but we have a a working model of how we think the thing works. So if you did that experiment and you uh built a website that used active voice only, and you saw a 7% increase in conversion, your theory, your model of website conversion language would be that active voice improves conversion rates by 7%. That would be your model, the your scientific theory that you would put forth. And then again, you can repeatedly test it, you can test it on different websites and different industries and things to see if that model, that theory holds up.

This is what is sets apart data science from everything else. It is that rigorous testing to prove or disprove something. And so that's why it matters in content marketing. That's why it matters in any kind of marketing. If you're using the scientific method along with the stats, the technology, and the business acumen, you're developing models of how customers behave, of how marketing behaves in a way that allows you to reproduce and build on your knowledge base and to continue refining in a reproducible, replicable way.

You can do the experiment over and over again, and you keep getting the same or results. That is why data science matters so much. A lot of marketing people could just kind of guess, right? They're just kind of trying things randomly. And that's not the way to build a scientifically reliable method of doing your marketing.

So really important. You can tell I could go on for hours and hours about this, but really important that that's why the scientific method is so critical to what is data science, right? You have to be using the scientific method. And there's again many, many more rat holes you can dive down, but keep that in mind. It's so important.

Leave your questions in the comments box below. Subscribe to my YouTube channel and the newsletter. And if you're interested in this stuff, join my Slack group. Go to TrustInsights.ai slash analytics for marketers. Uh, we do talk about all things analytics and data science and statistics there.

Stop on by us. It's a lot of fun if you like nerdy math stuff. If you don't like that, it's no fun at all. Thanks for watching. I'll talk to you soon.

Want help solving your company's data analytics and digital marketing problems? Visit TrustInsights.ai today and let us know how we can help you.