My name is Ivan Tsekov, and I have specialized in performance marketing for more than eight years now. I’m here to share my experience implementing marketing automation and tools to achieve success in the projects I’m working on.
In this article, I’m going to show you the process I use for:
- analyzing the bounce rate (key user engagement metric) of my competitors
- conducting a series of A/B testing experiments inspired by the good practices of my competitors
- applying the results to my projects
What tools are we going to need?
For this strategy, we’re going to need two tools:
- A/B testing tool – my choice in this case will be Convert
- Competitor analysis tool – we’re going to use Spycrow
Process
We’ll use Spycrow’s website engagement report to detect which of our competitors’ websites are performing better than ours in terms of bounce rate.
After that, we’ll analyze their websites and create at least three hypotheses about why their websites have lower bounce rates. Then, we’ll use an A/B testing tool to accept or reject the hypothesis.
1. Generate competitor reports
We can generate the website engagement report for our competitors by following this process:
- Add your email and request a Spycrow report
- Wait for the Website Engagement report to arrive in your inbox
- Unlock the report (it’s completely free)
Here’s what the “Website Engagement by Company” segment of the reports looks like:
Now, we can start with the analysis.
2. Analyze the competitors’ websites
Let’s assume that our business is Asana (the first website in the report above). We can see that our bounce rate is lower than the industry average. That’s good, the lower the bounce rate is, the better.
However, Trello (our biggest competitor) has a significantly lower bounce rate, which means they’re engaging users more efficiently.
Let’s see how that is happening. In this case, we’re going to look only at the homepages of the two websites, but you can replicate the same process with any other main page type that you know is receiving significant traffic.
Here’s how the homepage of Asana looks like:
Now, let’s look at the better-performing competitor Trello:
There are a few main above-the-fold elements that we should look at for any landing page:
- Main headline
- Call-to-action buttons and elements
- Hero image/video
It’s important to mention that our analysis below assumes that Trello’s elements perform better than Asana’s. However, that’s not necessarily the case. It’s just an insight generated based on the competitor data we have from Spycrow. That’s why we’re going to A/B test to confirm whether that’s true or not. It might be partially true (headline performs better, but CTA doesn’t).
Headline
The headline of Trello is much clearer about what the tool is about. When people search for a task management tool, they don’t care as much about AI. Yes, it might be a nice additional feature, but it’s not the main USP. We can see that Trello also mentioned AI, but in a much more elegant way, in the banner below the main menu.
If we were in Asana’s place, we would like to test a more clear headline that reveals the main value or USP of our business.
Call-to-action
There are two major differences here:
- Trello has a field for inputting your email before clicking the signup button. This allows them to send follow-up emails even if you don’t complete the registration process.
- All CTA buttons contain “Free” in their copy
Hero image/video
Here, we also have a major difference – Asana has a video that’s only visible in half (bad UX in itself), while Trello has a static image and a small secondary CTA in case you want to see their video.
We would definitely want to test removing/hiding that video and using the above-the-fold real estate better.
3. Create a list of hypothesis
Now, it’s time to create the list of hypotheses we want to test. I’m using the following formula to create a hypothesis:
“If we change [variable], then [target KPI] will improve because [rationale].”
Let’s look at the 3 examples we’ve listed above and create a hypothesis for each of them.
Hypothesis #1
If we change the headline to [new headline], the bounce rate will improve because the headline provides a better understanding of the product’s purpose.
Hypothesis #2
If we change the main CTA button to “Get Started for Free”, the bounce rate will decrease because users are more likely to stay because the product has a free version.
Hypothesis #3
If we hide the video behind a “Watch video” button, the bounce rate will decrease because there will be more real estate for the rest of the content on the landing page that might be useful to the users, and they’ll be more likely to stay on the page.
4. Setup A/B testing
The setup before starting the experiment will require the following steps:
- Create an account in Convert.com
- Create a project for your website in the account
- Implement the code snippet on your website – you can learn all about it in their guide here.
- Test if everything works correctly
It’s important to ensure that your website does not flicker when it loads due to the new code. This can be achieved by ensuring that the snippet loads before the rest of the JS on your website.
Once you have the code snippet installed on your website, you can proceed to the next step.
5. Create A/B testing experiments
The experiment’s setup will be very easy, considering that Convert has a visual editor that allows you to change website elements without the need for a developer.
Steps:
- Log in to the platform
- Select your project
- Click on “Create a new experiment”
- Add the URL of the homepage or the page you want to test on
- Change the element from your hypothesis by clicking on the element and updating the content. Warning: Changing more complex elements, such as hiding videos behind a button, might require the help of a developer
- Save the variation
- Start the experiment
Important: Test only one hypothesis per experiment, and do not run multiple experiments on the same page simultaneously.
You can see more detailed information about the steps provided by Convert here.
6. Analyze and apply results
After you start the experiment, you’ll gain access to the report that shows the progress – how many visitors have seen each version, the uplift in conversion rate, and which version performs better.
In this case, our main goal will be to see if the bounce rate is being reduced by showing our variation. However, I would recommend more significant conversions, such as sign-ups. This will provide you with even more insights.
Once the experiment gathers enough data (the confidence level reaches at least 95%), then you’ll have a confirmation on whether your hypothesis is confirmed or rejected.
How much time it will take? It’s based on the number of visitors to the specific page and the current bounce rate of the page. The bigger the uplift in conversion rate, the sooner your experiment will gain statistical significance (95% confidence level).
This strategy guide won’t be a deep dive into analyzing your results, but I highly recommend checking Convert’s guide because it’s crucial that you read the experiment’s results correctly.
Conclusion
This strategy combines competitor intelligence data analysis, generates insights, and converts them into actionable strategy—a series of A/B testing experiments that will significantly improve your business’s performance.
It’s considered an advanced strategy due to the complexity of A/B experiments and the implementation process.
