5 simple skills you need to become more data-driven instantly

With marketing moving towards being more of a data trade than a creative profession, we all need to become more data-driven if we want to stay on top. Since not everyone loves math and analysis as much as I do, I’ve listed five simple skills I think you should learn to get started.

(And no, you don’t need to be a math geek to become more data-driven, if you know some basic math you’ll have more than enough foundation to get started.)


1. Learn how to convert files from .csv to a format readable by humans, like .xlsx

“There’s something wrong with the file”. This comment is by far the single most common I get from clients and colleagues who start to work more with data.

However, there’s nothing wrong with CSV-files. The format is used to store lots and lots of data without ending up with huge files. A CSV-file is a text file where you have one observation or object per row, and list all the related values in a specific order. You separate the data points by a known symbol, most often a semicolon or a comma.

When you want to work with data stored as a CSV-file in spreadsheet software like Excel you’ll have to convert it from a text format into spreadsheet format. Here are two simple ways to do that:

Option 1: When you start with a clean Excel document

  1. Open a new Excel document and navigate to the Data tab
  2. Click “From Text” close to the top left corner
  3. Choose to the CSV file you wish to open from your computer and click “Import”
  4. In the window that opens up, choose “Delimited” and click “Next”
  5. Check one of the boxes next to the different delimiter suggestions – most CSV-files uses either a semicolon or a comma to separate the values
  6. Click “Finish”
  7. Tada!

Option 2: When you’ve already opened your file

  1. Open your CSV file in Excel and select column A in your document
  2. Navigate to the Data tab
  3. Click on “Text to Columns” somewhat in the middle
  4. In the window that opens up, choose “Delimited” and click “Next”
  5. Check one of the boxes next to the different delimiter suggestions – most CSV-files uses either a semicolon or a comma to separate the values
  6. Click “Next” and then “Finish”
  7. Tada!

2. Learn the difference between average, median and mode

When you have a dataset with multiple numerical values, you sometimes need a single number to represent the dataset. A simple method to use is to summarise all your data points into a “typical” data point that represent the “centre” of the dataset.

However, when you calculate the “center” of a numerical dataset you have three different measures to choose from: mean, median, and mode. They each summarise your dataset with a single number, but they are not the same.

Mean 

The mean is the “average” number in a dataset.  You calculate it by adding all your data points together and dividing this sum by the number of data points.

Example: The mean of 37, 10, and 67 is (37+10+67)/3 = 114/3 = 38

Median

The median is the middle number in a dataset. To get the median you order all data points from smallest to largest and pick out the number in the middle. If you have an even number of data points and there are two middle numbers, you take the mean (see above) of those two numbers.

Example: The median of 37, 10, and 67 is 37 because when you organise the numbers from smallest to largest (10, 37, 67), the number 37 is in the middle.

Mode

In any dataset, the mode is the most frequent value – the value that occurs most often among all values

Example: The mode of [2, 4, 3, 3, 3, 1, 1, 2, 2, 2, 4, 1] is “2” because it occurs four times, and all the other numbers occur fewer times than this.

Why should you care? – Mean, median and mode in different datasets

In a normal distribution, the mean, mode and median measures are equal. However, if a dataset is right or left skewed, they are different from each other. Using these three metrics are therefore an excellent way to learn about the distribution of the data points in your data set, and that is often an essential part of analysing a dataset.

Three different data distributions and how the mean, mode and median are impacted byt the distribution

3. Learn what axis is the x-axis and what is the y-axis in a two-dimensional (Cartesian) coordinate system

The X-axis and Y-axis in a Cartesian coordinate system
The X-axis and Y-axis in a Cartesian coordinate system 

The X-axis and the Y-axis when looking at only the first quartile of a Cartesian coordinate system
The X-axis is also called the horizontal axis, and the Y-axis is called the vertical axis

Even though you don’t work with coordinate systems, any data-driven marketer should know which axis is which when someone says “on the x-axis, you can see the time of day” and “on the y-axis, you can see the number of likes”.

Most of the time in marketing analytics the correct terminology to use is “the horizontal axis” and “the vertical axis” since the graphs displayed are not coordinate systems. However, even if this makes more sense to you, some people will use x and y instead, and you’ll have to know what they mean.


4. Learn the difference between correlation and causality

Being data-driven will always include working with data analysis. Analytical skill is something you pick up over time, but there are two key concepts that you need to keep apart from the start: Correlation and Causation. These two concepts are important because if you don’t get them right you won’t get the rest of your analysis right. Additionally, it’s very easy to divide the data-driven people from the non data-driven people based on these two concept.

Correlation = describes the size and direction of a relationship between two or more variables.
Causation = indicates that one event causes the other – i.e. there is a causal relationship between the two events.

One classic causation vs correlation example is that smoking correlates with alcoholism, but it doesn’t cause alcoholism. However, smoking causes an increase in the risk of developing lung cancer.


5. Know how to distinguish the dependent variable from the independent variable

This skill is a little bit more advanced than the others, but if you want to become more data-driven and know what an analyst or data scientist are saying about your A/B-tests, it’s a good thing to remember which is which.

When you test a hypothesis with an experiment, the two main variables are the independent and dependent variable. An independent variable is a variable that you change or control in your experiment to test if this has effects on the dependent variable.

Independent variable = the variable that you manipulate (i.e. the time you post on social media)
Dependent variable = the variable that you hope to impact from the manipulation (i.e. the number of impressions for your social media post) 


Data-driven marketing for the anxious marketer

During a lunch with the CEO of a small start-up, he said: “People tell us that we must do data-driven marketing, but I’m not sure that that’s the most important focus for us at this time”. I tried to explain to him that there’s no way to do marketing without basing it on some data input.

When people talk about data-driven marketing, I find it problematic in several ways. It’s like saying that we should do information-driven marketing, or people-driven marketing, or behaviour based marketing. But I understood his concern since data-driven marketing is what many people are talking about now.

Data-driven marketing is non-sense

Data-driven marketing is just one of the marketing buzzwords that pop up every year. At the beginning of the digital marketing era, everyone was social media experts. After a while, they decided to become content marketing experts instead. Lately, I hear that more and more people claim that they’re experts on data-driven marketing.

At Retune in Berlin this fall, the British programmer Karsten Smith said: “If you focus on a tool just because it’s new, you might become a victim of the rhetorics of newness”. This quote stuck with me. In marketing, we are always hungry for trends, and we need to embrace changes and opportunities to stay ahead. But at times must remind me, and others: nothing is smart just because it’s new. There’s no causal relationship between the two attributes.

What is data-driven marketing?

Definition: Data — “facts and statistics collected together for reference or analysis.”

Definition: Marketing — “the action or business of promoting and selling products or services, including market research and advertising.”

If we start at the beginning: data-driven marketing should be promoting and selling products or services based on facts and statistics. But what if I tell you that facts and statistics is the base for all marketing? To do marketing successfully, you usually consider everything you know about your customer and based on that knowledge you act in ways that serve the business well. This method usually means that you use knowledge extracted from data in one way or another.

New technology gives us more data

New digital technologies have created new ways for marketers to get to know their customers. Most things we do online is trackable and possible to measure. These new technical possibilities make digital data cheap and pretty easy to collect in large volumes. But it is important to remember that data in itself, without proper analysis, is worth close to nothing. And it’s the people who interpret data that offer the value and stands for the most of the cost.

But digital data is not superior to other data types in marketing. Marketing is a complex task, so we need all sorts of information to do it successfully. Digital data can only count as part of this information. Digital data is often lacking more qualitative information, such as the customers’ thoughts, feelings or body language. And while the number of clicks before conversion, or the average time on an individual page, might say something about more qualitative aspects, we can never be sure.

To do effective marketing, you need a variety of information, including several types of digital data. If you blindly trust your digitally tracked data points, you won’t have all information, and you won’t fully understand your customers and their needs. Data analysis is about more than the slope of a curve. It’s about combining the curve with other things you know, and draw conclusions based on all the info you’ve got at hand.

Data-driven marketing is not for everyone

Few marketers know how to read the massive amounts of digital data that they collect. If you’re not familiar with statistics and data analysis, it will be tough to understand it fully at first. But the most important thing is that you should combine your (new) digital marketing skills with the rest of your “classic” marketing knowledge, such as customer interviews, focus groups and gut feeling.

Doing data-driven marketing without enough knowledge can probably be worse than not doing any data-driven marketing at all. If you don’t fully understand the information you’re basing your conclusions on, how can you trust your decisions? How can you create a growth strategy if you’re not sure what you’re tracking and why it’s important?

I studied both statistics, research methods and data analysis at Uni. I did quantitative data analysis for my psychology bachelor thesis and qualitative analysis for my business one. And sure, I believe I remember most of it, but it only makes me smart enough to know when to get help. I don’t mind doing fundamental analysis in Google Analytics; it gives me a lot of information that I need. But I never trust my interpretation when it comes to details.

Whenever I have clients in need of more advanced analysis, I partner with someone who knows more than me. Implementing tracking is, for instance, something I try never to do. But the good thing is how the stage when I need help keeps moving forward.

Don’t forget your qualitative data skills

With a background in behavioural psychology and consumer behaviour, two fields that are heavy on qualitative data, I often prefer qualitative data analysis. I still do a lot of customer interviews and empirical observations as an essential part of my work.

I also talk to salespeople and others who meet the customers. Salespeople talk to “real” customers every day. They often have a closer relationship with the customers than I can ever get from behind my screen. To me, this little chat is also data collection. And although digital data is essential, it is not a good idea for an organisation to trust only this data.

The data you get from a script in a browser is just half the truth. Complement what you see with more (qualitative) information. Try to merge different perspectives to make smarter decisions.

Update your knowledge

We will continue to need well-educated marketers and marketing teams with cross competences. If you don’t have any data analysis knowledge yet, make sure you update your skill set. It will become essential in a few years. A good start is to talk to people who know more than you. (You can often persuade data scientists and programmers with your real interest in their field).

But don’t believe people who tell you that qualitative data analysis is the only skill a modern marketer need. Or that data-driven marketing in the meaning of “big data” is the only thing you should do from now on. That’s just not true.

My three 2015 takeaways for the anxious marketer

  1. Base all marketing on reliable data and analysis that gives relevant information
  2. Don’t focus on what’s new, focus on what’s smart
  3. If you try to follow the buzz, you’ll always be behind everyone else instead of ahead