If personalization isn’t a major part of your marketing and sales strategy…
You’re living in the past, my friend. In 2019, personalization isn’t a nice to have — it’s a must-have for any data-driven marketer.
According to marketers, the top three advantages of personalization are:
- Increased visitor engagement (55%)
- Improved customer experience (55%)
- Increased conversion rates (51%)
Personalization takes A/B testing to the next level. Rather than running experiments across all of your traffic, a well-designed personalization experiment can provide the most relevant experience to visitors.
And it helps you achieve the best results from each visit, every time. In this article, we’ll talk through 3 common ways you can incorporate A/B tests and personalization.
Use audience segments for A/B tests
Traditional A/B tests provide you with a one-size-fits-all result.
Let’s say you test the layout of landing page A versus landing page B.
The results of the A/B test are: Landing page A gets a 60%signup rate; while landing page B gets a 40% signup rate.
On average, page A performs better, so you use that page going forward.
But here’s where we run into a problem.
You could be missing out on a bunch of conversions from those who prefer a certain landing page. What do I mean?
Just because landing page A wins in an A/B test, doesn’t mean that it’s the global winner. There are many other behaviors at play. For instance, an E-commerce customer could prefer landing page B while a SaaS customer prefers landing page A.
That’s why it’s better to split your audience into segments in order to carry out more personalized A/B tests. Segmentation is the process of dividing your traffic based on different characteristics. From the get-go, it can seem more daunting than A/B testing, but the results can be monumental.
So how do you segment for personalization? There are a number of ways you can segment your audience.
Email Marketing Consultant, Jordie van Rijn, outlines four “Pillars of Segmentation” in the chart below:
Types of segmentation range from basic demographic segmentation, such as using data points such as age and gender, to more advanced differentials such as intent and life cycle stage.
Use the data you have at your disposal to group users into segments for A/B testing. By looking at what information you already collect on visitors (and what you can potentially enrich via 3rd party data tools), you can start thinking about where to personalize your site.
Next, choose your most important segments and test variables for each segment.
You may discover, for example, that women who like yoga respond better to a different landing page layout than women who like cycling.
Thus, you will identify a preferred result for each segment that you can adapt to provide a better user experience or get more conversions.
Rule-based personalization experiments
Most marketers (68%) take a rule-based approach to personalization.
This means that they peel off another layer, delving deeper into their audience segments.
Rule-based personalization relies on behavioral or contextual cues.
A/B tests are only shown to users that meet a certain rule.
For example, a rule might apply to users who have spent less than $500 at your online store.
For these users, you might test which discount or offer is best to show them, offer A or offer B.
And if your goal was to increase the average order value of this segment, this would be a supremely useful test.
To carry out this kind of test, you need to set up rules within an automated personalization tool.
You will discover the best-performing result for each rule-based segment.
The future: predictive personalization
The only thing better than personalizing for segments is personalizing for each and every individual. It’s the most difficult thing to accomplish, but with great difficulty comes magnificent results.
After analyzing the performance of 330,000 CTAs, HubSpot found that smart CTAs convert 202% better than basic CTAs.
Smart CTAs are tailored to individuals.
They adapt according to demographics or behavioral cues such as name, company, browsing behavior, etc.
To fully personalize for an individual, you’ll need to implement predictive personalization using machine learning. How does that work?
A machine continually monitors your site and optimizes the experience for each individual visitor. It learns about your visitor (and/or customer) as they use your site — and it adjusts the experience in real-time to adjust for their unique needs.
This is extremely useful as visitor preferences and behaviors change over time.
So, even if you A/B test using rule-based personalization, you cannot guarantee that the result will be the optimal result over an extended period of time. What works today won’t necessarily work in a month or even a year.
There are seasonal, lifecycle, cultural, and world-view changes that impact user behavior in the long-run.
So naturally, if you want to keep up with these trends for your brand, you’d need to use AI-based predictive personalization.
Digital bank, Chime has implemented predictive personalization with much success.
They tested 21 different ideas and 216 different versions of their homepage over a three month period.
Chime was able to work out aspects such as which version of their homepage worked best on mobile:
They also experimented with which version worked best in certain geographical areas:
The overall result was a 79% lift in new accounts.
That’s pretty awesome, huh?
It just goes to show that predictive personalization can increase performance tremendously.
A/B testing and personalization are not separate entities.
They can be combined to provide a better, more relevant experience to visitors. Which means you benefit from increased engagement, more conversions and more goodwill from your customers
A/B testing segments of your audience is simple and effective.
Rule-based personalization goes further by basing tests on behavioral and contextual cues.
And the most advanced method is predictive personalization using AI.
So why not try implementing some basic personalization tactics next time you carry out an A/B test?
About the Author
Emil Kristensen is the CMO and co-founder of Sleeknote: a company that helps e-commerce brands engage their site visitors—without hurting the user experience.