As with any optimization process, it’s important to approach your personalization strategy as an iterative series of experiments based on a data-driven, informed hypotheses.
So, let’s talk about the importance of an informed hypothesis for personalization and how to formulate one effectively.
(p.s. If the idea of experimentation as the core of your strategy sounds like something hard to sell within your organization, check out our upcoming blog about creating a culture of experimentation!)
Hypothesize to Personalize
There’s a very simple reason why any personalization program needs an informed hypothesis.
A clear hypothesis, informed by customer data and a strategy focused on impacting key business goals, sets your personalization program up for success. Without a hypothesis, you risk wasting resources and investment on a strategy that doesn't serve the customer or your organization in a meaningful way - and you lose a valuable tool to help you understand the efficacy and impact of your personalization campaigns.
All About The Goals
In a previous blog post, we talked about the importance of goal mapping BEFORE you begin a personalization program. If you haven’t read it yet and/or you’re not sure if your company goals are mapped correctly - I strongly suggest you go back and read the blog.
Once you have your business goals identified and prioritized, you can start to make informed guesses as to how you can optimize customer journeys to meet that business goal.
The best way to think about this is as a problem, “What is keeping your customer from converting?”
Using existing data and insights, identify where personalization might improve the experience that most impacts the conversion you’ve decided to focus on.
Once you have your data-based problem defined, start to think about what solution you can test for on your site. While there are typically a number of different variables you could test for remember, only pick one variable at a time for an effective hypothesis.
Here are some ideas for a simple CTA personalization experiment:
- Test different language for your CTA messaging
- What is the unique selling point for the segment you are looking at?
- Test the design of your CTA
- Can you target imagery by segment?
- Test the placement of your CTA
- Adjust the placement of your CTA based on where your segment is in the conversion funnel
Also, make sure you have site behavior data that backs up your proposed solution.
The final component is to determine your expected result. This includes the action of the customer you plan to influence and will define your personalization metric.
When you put all these components together, it should look something like this:
Mobile users abandon their shopping cart 60% more than desktop users
Add a CTA reminder to check out on their return visit
Increased completed purchases from abandoned carts
In this example, your personalization metric is: % increase of completed purchases from returning visitors that previously abandoned their cart.
With these three components defined, you are ready to formulate your hypothesis.
You may see slightly different ways of formulating a hypothesis, but in general they all include the following:
If (cause) then (effect) because (rationale)
Let's use the same example from above to formulate a hypothesis:
If we show a reminder CTA to our mobile users on their return visit, then we will see an increase in completed purchases from abandoned carts, because our mobile user click tracking data would demonstrate they take action at the top of the screen on return visits.
It’s important to remember the outcome of your hypothesis may not be definitive, but it will always provide you with more data to inform you about customer behavior and expectations.
Crucially it will also inform additional hypotheses and further testing.
With your hypothesis established, you can quickly explain what your personalization experiment is testing. You are now ready to implement your personalization strategy.
Stay tuned for our next blog to review our implementation checklist!