Is marketing more art or science? We believe it’s a beautiful and confounding mixture of both. There’s certainly a high level of creativity and subjectivity to effective marketing. However, there’s a fair amount of science, too. That’s because digital marketing is a perpetual experiment.
Testing, Testing, Testing
The good news is that taking this perspective allows you to grow and learn. An ineffective marketing campaign wasn’t a failure—it was a test that shows an approach that doesn’t work.
This is why data is valuable in marketing—it’s one way to measure effectiveness. That’s also why it’s so crucial not to make assumptions or stick with something that’s not working.
If you’re a marketer, you’re constantly evaluating results and seeing what could be done better.
Start with a Hypothesis
The scientific method always begins with a hypothesis. That’s simply a guess as to what would happen under specific circumstances.
- I think a shorter subject line will generate more opens on this email than a longer one.
- I predict that a social media post with a video will get more comments than one without.
- I’m guessing that a landing page with larger CTA buttons will generate more conversions.
Your hypothesis can be based on research, something you’ve experienced, or a gut instinct (or maybe all three). The point is to start with a prediction to prove true or false.
Ideally, every hypothesis should directly connect to one of your marketing goals. That could be to increase social following, improve email click-through rates, or generate more e-commerce sales.
Take the time to formally document your guess upfront. That could be within your content calendar or to your team during a campaign meeting. Revisit this hypothesis after the campaign is over to see if you were right or wrong—and there are no hard feelings either way.
Control the Variables
A key to science experiments is running the test on multiple variables simultaneously. However, it’s best to isolate a single variable at a time. Otherwise, you won’t be able to reliably test which one influenced the outcome.
For example, if you’re testing the impact of an email’s subject line length on open rates, don’t also test different send times. Sending different email subject lines at two different times means you don’t know which variable made a difference.
Instead, send the different subject lines on the same day at the same time. That way, it was the subject line that influenced people to open the email and not the time of day. You can certainly test out send times in a later campaign—and while using the same email subject line.
This is what’s commonly referred to in digital marketing as A/B testing. That means sending out version A and version B to see which one was more effective.
Gather Enough Data
Big data refers to a set of data too unwieldy for a human being to process. That’s why algorithms and computers help us do data analysis.
Thankfully, you don’t necessarily need big data to run a successful marketing experiment. However, you do need a certain threshold of data to make the conclusions reliable. That begs the question: how much data is enough?
Unfortunately, there’s no hard and fast rule. Typically, more is better because it averages out outliers. In my experience, at least 500-1,000 data points are a good baseline to start from. That means, gathering 500-1,000 email sends, social media impressions, or website users.
What if your audience isn’t that big? Can you still test out email marketing if your list is smaller? Yes, it will just take sending a few emails to that list to collectively gather enough data to make sure the conclusions are (somewhat) accurate.
There is no perfect formula, so you’ll have to use your instincts, too. The more you think like a scientist and run intentional experiments, the easier this will be.
Draw Conclusions & Learn from the Results
Once you’ve run an experiment and collected the data, it’s time to make a conclusion. Was your hypothesis right or wrong? Were there unexpected things you encountered along the way?
Document your findings in the same place you put your hypothesis. Share these findings with your team and potentially with your audience too. The experiment was futile unless you learn from the experience and make any changes to be more effective.
For example, if you discovered that longer email subject lines frequently outperformed shorter ones, incorporate that practice into your future emails. Then test another variable to make your emails even more effective.
But keep in mind that digital marketing is consistently changing. People’s preferences shift over time. Technology updates sometimes render your findings irrelevant. So always be open to re-experimenting and adjusting your knowledge to match.