Is it worthwhile to carry out experimentation if you’re a marketer within a small company? The answer is a definite yes! Experimentation generates data that is turned into actionable insights, which ultimately lead to better marketing decisions. In today’s world, experimentation is crucial for growth and innovation. Below are easily implementable ways to get started on designing experiments and making them a key part of your marketing operations.
Work Backwards (and Make Results Actionable)
To make experiments well worth your time and effort, the experiments need to be able to provide results you can actually implement. Hence, actionable results should always be the goal of any experiment you carry out. Ask yourself what kind of marketing decision do you want to make at the completion of the experiment and let this decision guides how you collect data. This approach, in which you start from the end and work backward, prevents you from conducting an experiment that generates vague, unusable results.
Think Short-Term
The most accurate and actionable experiments deal with actions that can be measured immediately, such as measuring purchasing behaviors to determine if the changes lead to higher profits. You don’t want to be measuring things that would take 20+ years to gather results like lifetime customer value.
Start Small
Experimentation can also be small at first. In fact, it’s highly encouraged as you don’t want to be overwhelmed when doing something unfamiliar. You can use experiments to determine optimal pricing for your product/service or even experiment on which type of email content is going to generate more leads.
Be Incremental
In conjunction with doing small yet manageable experiments, you also want to design experiments aimed at incrementally changing each interaction with your customers. You don’t want to overhaul an entire company website in one go. Rather, start by tweaking the landing page.
Seek Out Natural Experiments
You don’t need to treat experiments like they’re supposed to occur in a lab environment. The key to identifying natural experiments is to find treatment (affected group) and control (status quo group) groups that are influenced by external factors. For example, a new state law requires that you now have to collect tax from online purchases. You can create an experiment out of this external factor by measuring the number of purchases on your website before the online tax initiative against the number after the initiative.
