Have you ever read a marketing book prior to the rise of office computing? It was quite a different world. Marketing was a much more intuitive thing based on experience and focus groups. Those of you who like TED talks may remember the talk on spaghetti sauces that showed marketers the power of data-driven marketing.
These days the pendulum is far in the other direction. Marketers are drowning in data. Even if the internet hadn’t started, the spaghetti story would have pushed more data into marketing. But now every marketer has to learn a raft of analytics and metrics and how to weigh them. There’s A/B testing and statistical analysis. And now we’re using AI just to keep track of it all.
Marketers can’t escape data, nor should they. But in order to use data effectively, marketers need to learn a skill called data literacy.
What Is Data Literacy
Gartner defines data literacy as “the ability to read, write and communicate data in context”. It includes an understanding of data sources, constructs (e.g. graphs), and analytical methods. It also includes knowing which data tools are best to use in which situations.
That’s quite a lot to unpack, but it’s easy to think about. Literacy is the ability to read and write a spoken language. Languages are made up of components like sounds, words, tone of voice, syntax and grammar rules, just at a basic level. A literate person can translate that spoken data into a written form and then pull the data out again through reading it.
A literate person can also put together disparate pieces of spoken language data and draw new conclusions from them. We can remix what we’ve heard in new ways and come up with new ideas. People who can do this well are called authors and writers and literary critics, among other names.
Data literacy is just like this, only instead of spoken language we’re using the language of data. We can’t just listen to the stream of data. We have to capture it correctly and analyze it in the right way before it will reveal its secrets to us. However, this doesn’t mean you have to become a data scientist.
Why Data Literacy Is Important
Marketers are relying more and more on data-driven decisions, whether that’s done through interpreting signals, reading metrics, creating lead scoring systems, or using AI to make marketing decisions. However, before marketers and managers can trust these decisions they have to understand the data and the tools used to visualize and interpret that data.
Also, they must understand the flaws that can arise from poor data practices or using the wrong tools to measure the wrong things. The wrong marketing decisions can lose companies a lot of money. Increasing your data literacy can help you gain confidence in your technology stack. It can also point out flaws and improvements you need to make.
What Makes Good Data
ChartMogul uses the acronym CAUSE to describe good data. Good data is:
You have to trust the source of the data. Imagine the chaos that would happen for marketing if Facebook or Google was measuring data incorrectly!
Data that is actionable is practical and relevant to the questions you’re seeking to answer. This also applies to data visualization techniques.
Data can be biased for a wide variety of reasons, from seeing what we want to see to interpreting data in a way that supports some other interest. For instance, data visualizations can be tweaked to bias the interpretation one way or another to support a company’s argument. Data literacy helps us see these flaws.
Statistical literacy is a separate but related topic to data literacy. As a simple example, if you’re going to use statistics there has to be enough data to achieve statistical significance. Small sample sizes (not enough data) can lower the accuracy of an interpretation.
Easy To Interpret
Clean data visualization is a lot like good writing. The message is easy to grasp. If data looks confusing, someone may be trying to hide something.
If any of these parts is weak then it can cast any conclusions you make about your data into doubt. Therefore, if part of your job is to explain marketing data to others then you should be prepared to answer questions about these in advance. Where did you source your data and why is that data important? How did you remove bias from your interpretation? Are your methods statistically relevant? Can you explain how to read your reports to someone who doesn’t know your field?
Data Literacy Reduces Reporting
One of the benefits of data literacy in your company is a reduction in the number of reports you have to make. Some of the data you collect is meaningful, but much of it isn’t. However, we fear missing something important in the noise of the unimportant and try to find some signal buried inside it that can give us an edge. The result is a barrage of reports that waste time and confuse the issue.
We risk bringing up a bad memory from school but try to think of the last time you read literature from the 19th century. The writing of this era has prose that’s really complex and describes all sorts of things that are rather unnecessary to the plot.
To some people, this sort of complexity is wonderful. They love to pick apart the words and spend time imagining the scenes with the help of all that extra description. But that’s not likely to be your upper management who just want to know how to make their next decision on where to take their marketing.
When you’re able to understand the data and ask the right questions from it, you can cut out the fluff and show your teams the metrics that truly matter. Reporting becomes far simpler.
If you’re going to graph your data, the right kinds of charts have to be used with the right kinds of data. There are good reasons why there are so many graph types in Excel. It’s not just for variety. This post from Hubspot goes into the major types and what they are used for.
Good charts have the following pieces:
- A clear title explaining what the chart shows
- Labels on the axes, along with units.
- Unbiased starting points for the axes. (e.g. bar graphs starting at zero)
- A legend if you’re comparing a lot of different values or trends
- Not too many items at once. Choose the most significant!
Finally, when you’re making your charts, be wary of causation errors. Just because two things are trending in the same direction (correlation) does not mean that one thing is causing the other. This is a common error. Try to find other reasons why both data sets are moving in the same direction before jumping to conclusions.
You don’t need to be an expert in data science to use data smartly, just like you don’t need to be a meteorologist to read a weather report. You just need enough to see if your data and your conclusions fit the five criteria above. If you can do that, you can confidently make decisions on that data.
Subscribe to clever content club.
Subscribe to our newsletter for more on content marketing and artificial intelligence