Digital marketers looking for a better way to measure their ROI might want to think back to a lesson from their elementary school days. I'm talking about the Scientific Method, the eight-step method of inquiry that we all used on countless chemistry and science projects. It turns out to be a pretty effective tool for digital marketers too -- if you're willing to tap back into those scientific roots.
One of the first things we’re taught as marketers is to test. Regardless of the medium, Marketing 101 dictates that we test once, twice, and a thousand times -- and react to the results accordingly. This idea, while basic enough, can get jumbled. Technology, our friend and sometimes subtle foe, can often trip us up. What, exactly, are we testing? Which metrics are important to us?
At Grainger, we frequently test just about every single email we send and meet regularly to plan our tests and to look at results. Through our email marketing testing, we've learned, for instance, how the use of video can boost click-to-open (CTO) rates by 29 percent and how special discounts can more than double an email marketing campaign's open rates. We know too that customer testimonials and the placement up high of Twitter and other social media icons are highly effective.
There are other lessons we've gleaned from our rigorous testing about what works -- and what doesn't -- in email marketing. Here, however, I take a more detailed look at how the 'scientific method' works at Grainger for one popular way of trying to engage customers: the use of icons such as stars, hearts or market-specific symbols like an airplane or a car.
1. Define the question
Our strategy and production teams work closely with our creative team and Responsys to brainstorm test ideas and execute a plan. The brainstorms produce ideas and important questions. Here's one typical question we might ask -- Will icons in subject lines improve open rates? -- and how we tackle that scientifically.
2. Do background research
You may already have a solid base of knowledge about your product and customer, but it's important to do a bit more homework specific to the question. In this case, we need to understand what kind of icons we're hoping to use. What are some icons that have traditionally done well for emails? Which would make sense for our specific campaign? The knowledge we compile will help create the hypothesis we are going to test.
3. Form a hypothesis
The hypothesis will be the answer to our question. And don't forget from your grade-school days -- your hypothesis should have a measurable component to it. In this case our hypothesis would be: If we put this icon in our subject lines, then our click rates will improve (here, the measurable piece is the open rates). Keep these tests simple. You can run many at a time -- concurrently testing engagement and revenue, for example.
4. Test the hypothesis
Keep in mind that this test may be conducted again -- maybe six months down the road and by a different team -- so we must form an experiment that can be easily reproduced later on. This step is pretty straightforward: Commit to the amount of time and frequency you plan to send your emails (in this case, an email with an icon in the subject line) and record the data.
5. Analyze the data
In this case, we found that icons in subject lines increased open rates by 14% when compared to open rates for the emails without icons.
6. Form conclusions
With this data we were able to conclude that the hypothesis was correct and were able to move forward accordingly with this campaign by including icons in the subject line.
7. Communicate results
Science-haters rejoice! This step doesn't have to come in the form of a science report. We share our results with our teams so that they can see what worked for this specific campaign. Equipped with numbers, our conclusions are more compelling and are generally more apt to be put into practice.
This is probably the most important step of all. Our products and consumers are ever-changing. As you know, what works well one month could fall flat the next. Test everything -- frequency, offer, content, placement, call to action, landing pages, and personalization -- to be sure all of your components are helping the emails perform best.
Of course, results will vary based on particular industry, your database and other variables. The bottom line? Marketers need to test and track results to keep their audiences engaged and to improve ROI. Getting a little more scientific about that will go a long ways.
Janine Salmon is an Email Marketing Supervisor at industrial supply company W.W. Grainger, Inc.