Based on How to Create a Marketing Growth Strategy - CXL by Tom Whatley
Business benefits
Design your experiment. Start with a hypothesis, including an independent and dependent variable.
For example, you’re a marketer for Canva. You’ve mapped your customer journey and the associated metrics, analyzed your data, and you see an opportunity to improve your conversion rate. You notice that trial conversions of Canva Pro aren’t converting well.
- Your assumption could be that users cancel because they don’t understand the full suite of available benefits.
- Your hypothesis could be that increasing the number of educational emails throughout the user’s trial period would increase the conversion rates of trial users by 10%.
- The independent variable, or what you’re changing to better understand the outcome, is the number of emails you send throughout the trial period.
- The dependent variable, or what you’re measuring as you make changes to the independent variable, is the percentage of new customers who convert at the end of their trial period.
Knowing the value you’ll provide to your customers also helps you determine how you’ll design the emails. It’s not about the number of emails, but the value they provide inside. This is where you’ll carefully design your experiment to showcase the premium membership.
Ship your experiment by starting with A/B testing.
A/B testing is the simplest way to have a control group to compare against your independent variable.
For example, using the same Canva example, your first move would be sending an email about the premium membership benefits compared to not sending an email. Then you’d measure the lift to understand the value of sending that email.
If there is some lift in your conversion rate, start experimenting with other variables to try and beat the original lift level. Next, you might look at the copy in the email and test whether language about price or language about features performs better.
Set another A/B test and keep repeating that process while benefiting from the successful experiments.
Analyze your experiment by measuring the results and key metrics.
Was your hypothesis correct? What else did you learn about your customers throughout the A/B testing process?
Following the Canva example, measure metrics like what percentage of users converted to a paid membership, as well as email performance metrics like open rate, click-through rate, and unsubscribes.
Ask questions that get the wheels turning about future experiments you can run:
- Did a certain email have a significant bump in opens?
- What trends can you track relating open rate to time of day or week, or subject line copy?
- What was user behavior following them clicking through to your site?
- Did they bounce immediately or browse your content and features?
Automate and scale, or take your learnings and redesign some or all elements of your experiment.
In the case of Canva, if a certain email outlining the value of the premium membership boosts the conversion rate by 10%, automating that email to go out to all free trial customers should add 10% more value into your business. Automating that successful email allows you to focus on the next experiment.
No matter the outcome of your experiment, you’re learning and collecting data about your customer behavior. This information should fuel your future experiments and campaigns.
Last edited by @hesh_fekry 2023-11-14T10:13:49Z