Empirical research means gaining knowledge by observation or experience. In content strategy, we typically gain knowledge by similar means, whether it’s a content audit or some kind of user research. And that’s why empirical (social) research and content strategy are a great match: the research helps to quantify evidence or make sense of it in a qualitative form. There are several tools from the discipline of social research to use in content strategy, but more interestingly the process of empirical research can fuel content strategy.
Working with the empirical cycle
Adrianus Dingeman de Groot was not only a Dutch chess master and psychologist, he also conducted some of the most famous chess experiments of all time. He has developed the empirical cycle, which I will use in the following to illustrate how an empirical approach can be used in the field of content strategy. In this context, I’ll exclude any explorative character, which is a similar but different topic.
This is where our research begins. We observe a phenomenon and think about its causes. In the context of content strategy such a phenomenon could be for example that on rainy days more users are consuming content on a website. Or, if we use a certain wording, the users are more likely to convert. This might sound familiar from the Growth Hacking space, where marketers work with experiments. Actually, it’s not that different. You make an assumption about things you observe.
Now it’s time to formulate your hypotheses. This sounds easy but it can be really hard to formulate a hypothesis that generalizes the observed phenomenon. A hypothesis itself is not just a random question, it can be seen as an educated guess that can be tested.
Some hypothesis in the field of content strategy I’ve created during our course “Basics of Empirical Social Research” at the Content Strategy Master programme of the FH Joanneum Graz are:
- Target-group based lead magnets increase conversion rates of email subscriptions.
- Sticky CTAs lead to higher click rates on mobile devices.
- Faster loading times decrease the bounce rate of a web page on mobile devices.
- A personalized headline improves the CTR of an article on a website.
- Personalized content improves time on site, conversion rates and retention rates.
To test the hypotheses you have to formulate some experiments. They can either confirm the hypotheses or refute them. It describes the process of (deductive) reasoning from one or more statements to reach a logically certain conclusion.
That’s the actual process of testing the hypothesis and collecting data. These tests are often also called experiments. For example, such an experiment could be an A/B-test of a newsletter, any creative user research method or a “classic” method like an online survey.
In line with social research, an experiment should include two kinds of variables:
- An independent variable, which is manipulated by the researcher
- while the dependent variable is measured.
Moreover, we should randomly allocate the test subjects to neutralize our bias (if that’s possible). Although, I think that this claim might be too over-optimated in a business context, it absolutely makes sense to think about the designated variables as well as possible bias and how to avoid them.
Eventually, you try to make meaning of the data and formulate a theory. Such a theory should be a reasonable explanation for the selected phenomenon. Besides generating valuable insights, the last stage should also include reflection and maybe also recommendations to make educated decisions. Furthermore, this last stage of the empirical circle can fuel further research.
For example, if it can be proven that faster loading times decrease the bounce rate of a particular web page on mobile devices, you could outline recommendations on how to optimize the website for better loading times.