Marketing as an Experiment
by Karen Janowski of Traction Team
It’s practically a cliché to say that many startup CEOs are skeptical of marketing spending. To some, it’s a mysterious black box, voodoo laced with inscrutable jargon, or just a grand experiment. Well, I’m here to tell you that at least part of your marketing should be an experiment!
Why? Because marketing (particularly the part focused on arousing customer needs) is not a mechanistic process where you put in certain ingredients, turn the crank, and get out a predictable result. There are lots of variables:
- Are you targeting the right decision makers?
- Did you actually reach them?
- Are the product benefits meaningful, relevant, and exciting?
- Are your messages in the right medium?
- Is the timing right?
- Have you understood and addressed purchase obstacles?
- Can prospective customers easily respond, learn more, or make a purchase?
Sure, some of these questions can be addressed by strong advance planning based on customer and market research. But many of these issues can best – and most quickly – be resolved through market testing or experimentation.
Experimentation can be fast and relatively inexpensive. But let’s be clear. I’m not talking about a risky roll of the dice or a crap shoot. I’m talking “scientific method.” Remember your introduction to the scientific method for the fifth grade science fair? Probably not. So, here’s a refresher.
- State the Question or Problem: This could be something like “our lead generation campaigns are not converting into sales revenue in large enough numbers.” Focus on a significant problem and try to state it clearly and objectively.
- Observe and Research: This is a fun step in the process, a place to exercise curiosity and creativity. Have some campaigns worked better than others? What was different about those? Where in the process are things falling apart? Is it a problem of quantity, quality, fit, ability to follow up quickly with the right message? What would success look like? Are there some tried and true methods that work in your marketplace? Or are there some new and different methods that are worth trying?
- Create a Hypothesis: This is an educated guess about actions and results. It could be something like: “If we charge participants $20 to attend our (formerly free) webinar, we will double the percentage of participants taking a live phone call from our inside sales force.”
- Design and Run the Experiment: Think through the structure of your experiment. Outline exactly what will be done, when and how. Try to isolate the main variable that you want to test (or be prepared to analyze multivariate data). For example, if you are testing the impact of $20 vs. free webinars, you may want to control for other factors such as promotional methods or target audiences. Design the experiment so that you can collect relevant information. Marketing experiments tend to fall short on steps 3 and 4. As a CEO, if you hear “we’re experimenting with this marketing campaign,” make sure you ask how the experiment is structured. What is the experiment testing? How will the results be determined? How will the “marketer/scientist” know that the experiment succeeded or failed?
- Collect Data and Analyze: The design of the experiment must take into account how the measure the results. Will there be a before and after comparison? Is it necessary to measure a baseline? Will the experiment be conducted simultaneously with two different groups (i.e., in this example a test group that is charged $20 and a control group that is charged $0 for the webinar)? Can the results be measured in a gross or qualitative way or is a high degree of accuracy needed? Measure the results and compare vs. the hypothesis.
- State a Conclusion: If the hypothesis was proven true, what are the implications for future actions? If the hypothesis is false, what are some of the possible reasons for that? Did certain conditions adversely affect the test and should it be re-run? Or should new hypotheses be generated and tested?
It may not be appropriate or cost-effective to treat your entire marketing budget as an experiment. Sometimes the cost of measuring results or ROI can be prohibitively expensive. Still, it’s a good idea to allocate a portion of the marketing spend to “experimental” marketing. You might discover something extraordinarily impactful. Just make sure you apply the scientific method we all learned back in fifth grade!
What marketing experiments have you run? What did you discover? Share your thoughts!