The following is a guest post by Alex Birkett of Omniscient Digital.
No one ever claimed conversion rate optimization was easy.
In many cases, you’re dealing with a dearth of traffic, difficulty prioritizing impactful opportunities and ideas, and a whole lot of noise with not a lot of signal.
Nowhere is this more apparent than in optimizing your copy.
Sure, you can A/B test salient copy on high traffic pages, like your value proposition.
But how do you know if your copy is resonating on other pages, like long form sales pages, pricing pages, or even email marketing messaging?
Over the years of writing copy and optimizing websites, I’ve found a few tried and true techniques for understanding your audience and how resonant and engaging your copy is. You can look at these techniques as a fundamental component to your UX research and design process.
In fact, many of the techniques here are either identical or complementary to how you would normally gather UX insights for website or product optimization.
8 message testing techniques for better copywriting
Copywriting is hard. Testing it is even harder. To truly understand the value of our messaging, it’s best to have tools and techniques to identify two things:
- Where your copy is working or failing
- Ideas as to how to improve your copy at those points
This list will provide both types of techniques. Here are some of my favorite methods and tools
- Wynter message testing
- Session replays
- On-site polls
- Customer surveys
- Website analytics
- Social listening
- A/B testing
1. Wynter message testing
All in, user testing is probably my favorite qualitative research method when it comes to finding conversion rate optimization opportunities.
You can draw a lot of insights on your copywriting from these forms of research, but generally, they’re going to more holistically identify user experience and usability problems with your website or app.
Wynter focuses purely on messaging and copywriting, which is why I’m so excited about this tool.
In the same way you’d run a user test, you get to select a panel of participants from your target audience (Wynter does the selection for you, or you can choose your own participants). But instead of picking website tasks to perform, you show them sections of your website, landing page, or email.
Then they are presented with a series of questions, both at the holistic experience level and the individual section of copy level.
These questions are intent on drawing feedback about the clarity, resonance, and trustworthiness, and persuasiveness of your copy.
We’ve run some message tests for our content agency’s website and found the tool to be incredibly useful.
The cool thing about this technique is it serves both of our research purposes: we can not only identify areas where copy could be improved, but we can come up with ideas as to how to fix it.
But even if you can run A/B tests, this data gives you so many ideas and hypotheses to run with that your new problem is going to be one of gluttony and prioritization; you’ll have so many things to fix and too little time to do so!
2. Session replays
Next up are session replays.
Session replays are another one of my favorite website optimization tools, generally speaking. Instead of directing users through your site consciously, you simply get to watch anonymous users interact with your site as they naturally would.
The benefits of this are apparent: people perform differently when they know they’re being watched. In the case of session replays, you get to see real, unabated user behavior.
The downsides of this method are twofold. First, the time it takes to filter through session replay videos is intense. Second, the precision of your insights isn’t great; two people can look at the same video and come up with different learnings.
To mitigate the first disadvantage, many tools are now proactively identifying signals of user frustration like rage clicks and rapid mouse movements. This allows you, the researcher, to filter quickly and look only at the videos that seem to have provoked frustration.
The second disadvantage, however, is harder to mitigate.
Typically, when I look at session replay videos, I take any insights as ‘preliminary evidence’ or as fodder for optimization ideas. Just because a single user seemed to have struggled to read a certain section of the page or click a certain button doesn’t mean it’s a universal problem.
However, it is a valuable data point and can help you ideate and prioritize experiments.
TryMyUI has great features to help you capture and analyze session replay videos.
Heatmaps are amazing to identify opportunity areas at a high level.
A ‘heatmap’ is really an amalgamation of a few different visual quantitative reports that show you where users are interacting on a page or website. Typically you can see visualizations of:
- Where users are clicking (clickmaps)
- Where users are scrolling (scrollmaps)
- Where users are viewing (mouse tracking)
The implications for copywriting are obvious. The first thing you can do is see what percentage of users are even *reaching* the CTA on a page and where the biggest dropoff points are.
Of course, it’s not always copy that provokes users to scroll further down the page or click on a given CTA. It could also be unclear or distracting design. But in many cases, a sharp dropoff on a scrollmap tells me that the copy gets boring, unclear, or confusing (especially if you can tie that dropoff to exit rates and bounce rates in your analytics).
Hovermaps or mouse tracking maps can also show you where the majority of attention is on a page. This can help you reorganize your messaging in a profound way.
For instance, if you note that a section of your site is getting a majority of the attention, you can shift the placement of your most important and impactful copy to be near that point.
Heatmaps are great at identifying user behavior at a high level and crafting your copy and design to accord to that. They’re also great at identifying dropoff areas where you can potentially optimize your copy for greater engagement.
Heatmap tools include:
4. On-site polls
When it comes to learning about your users and what parts of your copy are resonating or failing to perform, sometimes the simplest solution is best: just ask them.
That’s what on-site polls let you do with anonymous website visitors. You can trigger a poll on certain pages, to certain users, and at certain times, and collect feedback based on questions that you choose to ask.
It’s a flexible technique that can either be useless or powerful depending on how you construct your research design.
Always start designing your on-site polls with a business question.
Really think through the question, “what do I want to learn and what will I do with those insights?” And then craft questions and targeting that have a high probability of eliciting those learnings.
When it comes to testing the effectiveness of your copy, I find these tips to be helpful:
- Don’t add polls to pages unless you’ve thought through your business question and what you hope to take action on
- Think through the targeting and timing of your polls. Someone who takes a poll after purchase is a much different person than someone who takes a poll on a blog post.
- Open-ended questions produce the most interesting insights typically. You can often lead in with a yes/no question and then follow up with an open-ended question.
- I like questions that draw insights on additional information, hesitation, and doubts. For example, “What information were you looking for that you were unable to find?” Or after purchase, “What almost stopped you from purchasing with us today?”
- The hardest part is codifying these insights and taking action on patterns. Use a framework for organizing your voice of customer data.
Tools to help you run polls include:
5. Customer surveys
Customer surveys are a more proactive form of qualitative messaging research.
These have some distinct advantages and disadvantages.
The main disadvantage is that it’s really difficult to design useful surveys. This article gives a great overview on qualitative survey design.
As for advantages, there are many:
- You can often draw deeper and more comprehensive responses from your existing audience
- Your targeting is very clear; you know who these people are
- You can target at very clear stages of the customer journey, which can lead to different questions and answers.
For me, I like to use customer surveys at a few different stages of the journey:
- In drip campaigns for leads and subscribers
- Post-signup for users and customers
- During specific and targeted optimization campaigns
The first one is the easiest. If you’re collecting email leads, you should be following up with a drip sequence. In the 3rd or 4th email of the drip sequence, send a short survey.
This can include demographic questions (I like open ended stuff like “describe yourself.” This can give you more emotional and rich insights on who your subscribers are).
It can also include preference questions. For example, at CXL we asked what products people would buy from us if we were to make them. That insight led to the CXL Institute.
For post-signup surveys, it’s all about customer satisfaction, user experience issue identification, and segmentation. Your best bet is to learn how delighted people are at a given point in your product experience, what improvements there could be, and how you can segment these people to potentially aid personalization as well as acquisition efforts.
And for specific optimization campaigns, well, you’ve got to think through the questions and survey design for yourself.
For example, if I’m optimizing a paid landing page I’ll have different questions than if I’m designing a signup flow. For the paid landing page, I’d want to make sure to survey people who signed up via that page, and my questions would attempt to tease out desires, hesitations, doubts, and expectations from that audience.
At the end of the day, the power of surveys lies in the questions and design. It’s an art and a science, but there are articles that can help you with inspiration.
Tools to help you with customer surveys include:
6. Website analytics
While most of the tools on this list cater towards the qualitative side of things, you can’t discount the quantitative for identifying issues in the first place. And when it comes to quantitative website analytics tools, Google Analytics is king.
Obviously, the platform contains more than just messaging insights, but one thing I like to do when optimizing a site is find underperforming pages. You can do this by going to Behavior > Site Content > Landing Pages and using the comparison table view. This shows you which pages are above and below site averages when it comes to your given metric (typically conversion rate or something like bounce rate):
Google Analytics can also help you identify how many people are clicking and engaging with specific buttons and elements on your site, which can help you prioritize those areas for optimization.
However, when it comes to messaging testing, website analytics can’t tell you the “why” or give you solutions when something isn’t working. You’ll still need qualitative tools like user testing and surveys to identify that stuff.
7. Social listening
In copywriting, being clever can help, but you know what is generally most effective: using the words your prospective customers use.
Much of what we’ve already covered in this list has been about ways to gather that voice of customer research. But one of the most underrated ways to do this is to find where your prospects are already hanging out and see how they speak in those places.
This could be Reddit or Twitter, or it could also be on review sites like Amazon or G2. In these places, especially social media, people are speaking more honestly and informally. Sometimes when you actually conduct a test, you get a Hawthorne Effect, where the respondent will alter their answers due to their knowledge of it being a study.
People on Reddit, however, are less filtered. So you can find how they actually speak about your product or product category.
There are many ways to gather this form of messaging data depending on your industry, but here’s a quick tip from Joanna Wiebe on how to do so on Amazon.
8. A/B testing
Finally, we can’t rule out the gold standard of research: the controlled experiment.
A/B testing isn’t going to be possible for everyone; for example, our agency doesn’t have enough traffic to run experiments.
And even if you do have enough traffic, sometimes it’s still hard to pinpoint exact parts of the copy that are and aren’t resonating.
But if you have the traffic and are able to isolate elements of the page, like the headline, then A/B testing can give you one of the best ways to validate better or worse performing variants.
The way I’d look at it is the qualitative tools above can help you identify components of your copy that need fixing as well as potential solutions. These get filtered into hypotheses or ideas to implement. Then, if you have the ability, you can A/B test them and quantify at a high level if the new version is *actually* better.
There are tons of A/B testing tools to do this now, including most of the landing page builders out there:
Copywriting is a mysterious art, but some of the clouds are clearing on the process.
Then you’ve got tools like Google Analytics to monitor performance and analyze opportunity areas.
Then you’ve got a whole area, now, of tools to help you progressively fix and improve copy, from message and user testing tools (Wynter, TryMyUI) to survey tools (HotJar, Typeform) to A/B testing tools (Optimize, Convert)
Now more than ever, it’s possible to measure and improve your messaging through technology and process.
Alex Birkett is a Co-Founder at Omniscient Digital. He moved to Austin, Texas after graduating from the University of Wisconsin (go Badgers!). When he’s not optimizing websites, he’s probably reading books, lifting weights, or karaoke jammin’.