Let’s talk about your copy. Search, display, social, email, landing page – doesn’t matter. First of all, I’m sure it’s great. Your copywriter knows the audience and the writing is effective. But here’s the question… could it be better? If not, return to Facebook and go about your day. Now, if you feel there’s some room for improvements, let’s talk.
What could be better?
Could you use a more engaging tone? Are your key messages effective? Are you asking users to read too much or do they want to know more?
There are a lot of variables at work. Most marketers view copywriting as a left-brain activity, but when we isolate these individual factors, we can introduce a touch of right- brain that connects your copy to your goals. So what’s the answer?
A/B testing.
I know, I know. Calm down and let me explain; It’s more than a buzzword. Marketers generally have one of two thoughts about A/B testing.
I say to both of you: A/B testing can be incredibly easy and incredibly easy to overcomplicate. For us, it comes down to three considerations: Isolate, tag and test. Remember those three things and you’re well on your well to better conversion rates.
When you look at copywriting as a sacred, creative process that must be organic, it’s hard to isolate. And it makes sense – like colors and brushstrokes in a painting, all these factors work together to make a final product. But when you take a step back, it’s still just brushstrokes and colors. What if this stroke was broader or that color was brighter? Pick one area that may be weak, and try something new.
Naming convention is crucial to testing. As campaigns get more and more complex, we need short-hand to keep track. Your ads, emails and pages all have a lot of elements. For example, you have email campaigns that address several audiences and you want to test a call to action (CTA). Try something like:
Edition_Number-Audience-KeyTheme-CTA
Now we can test the following CTAs and clearly determine the winner. More importantly we can test as many CTAs as we want over the next year. As long as we are consistent, we always have clear data.
This can be the tricky part. How do you implement the test. For landing pages, Google Analytics has a tool that will automatically serve your test pages. Some email systems link MailChimp and ExactTarget have the feature already built in. And any good ad platform allows you to ad and test a lot of variations.
Social media can be a little more difficult to run a test because we don’t want to post the same thing over and over again. There are ways to adapt your content calendar to test a single variable over time. More on that another time.
One last caution: sample size. The goal is a sample large enough for statistical significance. I know audience size doesn’t always allow for a large enough sample. So when you’re dealing with smaller numbers, tread lightly. Feel free to reach out for more on that one (don’t want to bore all these copywriters with too many numbers).
But for now - go forth! Isolate, tag and test!
