The How and Why of Data Science in Marketing

How many data scientists do you have on your payroll? Some large corporations have dozens. Some have none. Even if you don’t have someone like that on your staff, you need to make certain you have someone who is thinking like a data scientist. In fact, you really want your marketing team to think this way. Thinking like a data scientist allows your team to move from guessing about information to knowing it. But how can you do that? We have a few tips that will help your marketing department think like data scientists.

When you make statements that start with “I think…” you’re showing that you don’t really know. You’re basing conclusions on guesswork and estimations that may not be correct at all. When you start with “I know” and have hard, unbiased data to back up your statements, you’ll make much more of an impression. You’ll be treated more seriously.

Why Do You Need Data Science in Marketing?

Data science can be defined as harnessing your data to help your company make better business decisions. These decisions can be made at any level and by anyone from the CEO down to an assistant, but when they’re used in marketing, you’re able to make better decisions and predictions for your campaigns. This allows you to make your campaigns more targeted and to create better content. If you go to your CEO with a marketing plan based on data, it’s much more likely to be approved. It’s also more likely to succeed, which makes you and your team look better.

Being able to make better predictions affects everything from which brands are going to do well to what stocks increase. In many corporations, you as the head of marketing are expected to be able to make very accurate predictions about your customer base. It’s not always easy, but it’s what you’ve been charged with. That’s why thinking like a data scientist is so important. You can’t simply think like someone building a marketing campaign using what you think is true. You have to know it and have facts behind you.

Who Is a Data Scientist?

While the above definition of data scientist sheds some light on what they do, the question is who are they? Data scientists are usually found in areas of business, mathematics, and technology. They analyze the data they collect and organize in three different ways:

• Descriptive Data
• Prescriptive Data
• Predictive Data

What Do Marketing Experts Do with this Data?

Marketers are most often concerned with descriptive data. They use Google Analytics and other tools to see how customers and potential customers interacted with their marketing plans. This data is historical in that it looks only at past interactions by gathering information on click-through rates, cost per click, and other data. It gives you an idea of where you’ve been and what you’ve done.

Predictive analytics, on the other hand, is a way of showing you where you’re going. It takes this historical data and combines it with real-time information to predict what customers might do. It can generate Marketing-Qualified Leads (MQLs) that you can pass over to the sales department. Since predictive analytics are updated with real-time information, you can adjust your plan midway through if new data comes in. While most marketers deal with descriptive analytics, fewer work with predictive analytics.

Fewer still work with what’s called prescriptive analytics. This type of data usage goes beyond predictive analytics so that you can determine what’s going to occur and, more importantly, what methods you need to employ to take advantage of what you expect will happen. You can predict what’s going to occur and the best way to deal with that outcome.

One example of prescriptive analytics can be found in Watson, the AI created by IBM. When Watson asked to create ads on social media and track their performance. Several weeks later, Watson predicted that one image wasn’t providing the click-through rate needed and recommended replacing it.

What Thinking Like a Data Scientist Does for You

When you start to think like a data scientist, you’ll be able to better think out your content, refine it, and measure the outcomes through predictive modeling. This modeling lets you use various techniques and models that you couldn’t make use of otherwise.

Planning Content

When you think like a data scientist, you can create predictive models that can steer your content in a more effective direction. These models will be able to predict a number of things:

• Leads
• Demand
• Audience segmentation
• The Total Addressable Market

Refining Your Content

Testing is vital to refining your content into the best it can be. Like any scientist, you’ll begin with a hypothesis, or content that you believe your audience will like. Then you test that content through serial testing, A/B testing, or other types of tests. At this point, gathering data is actually more important than your content reaching your target audience.

Once you’ve gathered data and analyzed it, you now have hard facts. You can say with authority that content A did not perform as well as content B. You can then create content C from this information and test it against A. The cycle of testing, refining, and testing again continues.

Data scientists continue this cycle for as long as they need. Sometimes completely new content is thrown in to see if going in a very different direction would produce better results. Being creative is just as important as narrowing down the data and refining content.

Measuring Results

To measure your content’s or marketing campaign’s results, you need a model that takes into account all factors. This model is usually a type of algorithmic multi-touch content-attribution model. That’s a mouthful to say, but it’s a fairly easy concept to understand. This model looks at what your customer viewed and which content (or combination thereof) was effective. This provides hard numbers that CEOs and others love to see.

With this information, you can confidently present data at your next board meeting that shows how you’re progressing against your budget. You can list off your marketing-qualified leads, sales-qualified leads, customers, and more. All of this can be done easily and without creating anything really new or different. In fact, most of it is thanks to JavaScript. This programming code is used in multiple social media sites and can be harnessed to let you track all the data you need.

You may need to consult with an actual data scientist here in order to build a model that perfectly fits all of your needs. However, there are some things you can do to get started:

• Know what numbers your CEO and other executives want to see.
• Look at how to measure the performance needed to gather those numbers.
• Work with a data scientist to create a model or use a predictive-modeling SaaS platform.

What Numbers Your Company Executives Want to See

Business involves numbers. There’s no way around that. To really impress the leadership at your business and make them see you as someone to take note of, you need to be able to talk numbers. Specifically, you need to be able to talk about your return on investment or ROI. This is often the bottom line. It’s hard for someone in content marketing to figure out how to talk this language, but it can be done.

First, you need to forget ROI and look specifically at ROMI, or your Return on Marketing Investment. ROI is usually only used when looking at something that has been sold. Marketing doesn’t really deal with that. Instead, you can look at the cost and the profit that relates directly to marketing or a marketing campaign.

Consider ROMI to be the incremental net income divided by the cost of your marketing campaign times 100 percent.

Then you need to talk about CLTV or the Customer Lifetime Value. This number is an idea of how much one customer will bring in over their entire lifetime with the company. If you’re investing more to bring in a customer than what the average CLTV is, you’re likely losing money.

How do you calculate CLTV? Take the average amount of income per customer, multiply it by their lifetime, and then multiply that by the gross margin. You can determine your gross margin by figuring the ratio of total income to the total cost of goods sold minus the costs of services.

That’s a lot of math, and as a content marketer, it might be more than you want to deal with. There’s no shame in bringing in a data scientist to handle all of this. That’s especially true if your marketing team is fairly small and you’re needed to keep on top of the latest campaign. You might not have time to do this work.

How Do You Measure Performance?

Measuring your content’s performance also requires a model. In fact, another multi-touch algorithmic attribution model is needed. This model can take information such as the UTM data you add to website URLs. This data tells you what website the user came from. For example, you can use this to see who is coming to your website from your Facebook profiles. That’s just one of the most basic ways to use UTM codes. Once you have this data, you can use the model to determine how well that part of your campaign is performing.

Always keep one thing in mind when you’re determining how to measure your campaign’s performance: what will you do with the information you get? That’s necessary to design the model correctly and to determine what data you need to collect.

Working with a Data Scientist

The easiest and best thing to do to gather all of this data, create the needed models, and draw the right conclusions is to hire a data scientist. This way, you have a creative person on hand to build your models. If that’s not an option, you can look at SaaS platforms that make use of predictive modeling. You’ll have to spend time learning how to use these models, but they can be a more affordable option.

Think like a Data Scientist

Overall, if you’re a content marketing expert, you have to know how to examine data and draw conclusions from it. You don’t have to be a data scientist, but you need to know how to think like one.