Artificial intelligence has been seeping its way into conversion rate optimization (CRO) for years.
One of the earliest applications was chatbots. Amtrak made headlines when it reported a 30% increase in revenue from customers who interacted with its chatbot in 2016.
But today’s intersection of AI, machine learning, and CRO goes far beyond chatbots.
Here’s what’s going on:
AI is democratizing CRO methods like predictive modeling and multivariate A/B testing that once required huge amounts of resources, data, and traffic.
If CRO is like fixing a leaky bucket, adding artificial intelligence to the process is like having a machine test dozens of ways to patch each hole—while running simulations that help you design the bucket better in the first place.
AI has fundamentally changed conversion rate optimization. I’ve seen this firsthand with my proprietary AI-powered CRO process, which generates an average conversion lift of 193%.
The impact of artificial intelligence on CRO
In traditional CRO, the typical practice is to launch your campaign, gauge performance, and then run A/B tests to further optimize.
AI tools are changing every step of this process.
With tools like AI heatmaps and predictive analytics, you can test your hypotheses before your campaign even starts.
You can use AI copywriting tools to quickly come up with effective, conversion-focused copy.
Then there’s the testing itself. AI-powered A/B testing tools are faster, more flexible, and can automatically categorize visitors based on attributes like behavior and location—and then route them to a variant of the landing page that’s most likely to help them convert.
How AI conversion rate optimization works
Traditional conversion rate optimization focuses on strategy, research, data collection, analysis, testing, and optimization.
Let’s review each step to understand how AI fits into the CRO process.
Strategy
Strategy is driven by business goals, which isn’t something AI can help with. Your CRO strategy might be as simple as wanting to increase your conversion rate.
Some CRO strategies also focus on reducing customer acquisition costs, maximizing traffic, attracting better-fit customers, or increasing average order value.
Once you start getting tactical, that’s where AI can come into play.
Predictive modeling software like Obviously.ai predicts revenue outcomes for all those “what would happen if…?” questions common during high-level strategy sessions. Knowing the likely outcomes helps you decide which strategies to pursue.
Customer persona generation is another way AI supports CRO. Tools like Delve.ai speed up this process by automatically creating customer personas from Google Analytics data.
Once you’ve got a strategy in place, it’s time to go into research mode.
Research
First, make sure you understand why people are coming to your website. You won’t get far with an offer designed for the wrong audience. Analytics data and on-site metrics can help you here.
Second, make sure there aren’t any barriers stopping users from converting or causing them to bounce before even reading your copy.
There are many ways to tackle this, but all of them start with data. Watch your users’ behavior closely using heatmaps, session recordings, and autocapture software. Start recording as much data about your users as you can, including chat logs, support emails, and phone calls.
Ask users about their subjective experience with polls, surveys, and even one-on-one interviews to gather anecdotal information. Tools like Qualaroo can gather customer feedback with non-intrusive popups while customers are on a relevant page.
But what do you do with all this data? That’s where AI comes in.
Analysis
AI crunches data far better than humans.
Just look at Google Analytics, which is used by over 30 million websites. Google Analytics has a built-in AI-powered insights tool that surfaces trends in website performance so you can make sense of the huge amount of data being tracked.
More advanced analytics tools use a process called “autocapture” which logs every interaction on your website, helping you understand where conversions are being lost.
Digital experience intelligence platforms like FullStory capture everything happening and then process this mountain of data into actionable information.
Sentiment analysis tools like Keatext can analyze subjective customer feedback at scale, including surveys, reviews, chat logs, and social media, and put those into insights you can take action on.
Having an insight-generating analytics tool in place is crucial to identifying visitors who are leaving your site without taking action, and understanding what their obstacles are.
Predictive analytics
User feedback and interaction data used to be a prerequisite for building hypotheses for CRO campaigns.
But these days, tools like Attention Insight use AI to simulate how real humans would interact with your website. By training machine learning models with tens of thousands of image datasets from eye-tracking studies, AI heatmaps are 90% as accurate as studies involving humans—and they can be done in 60 seconds.
Predictive analytics tools go a step further and predict what actions individual users are likely to take. Tools like Pecan.ai have propensity modeling features that put each visitor into a different class according to their likelihood to convert. Propensity modeling is based on attributes like behavior, gender, income, employment status, and location, and is particularly powerful for upselling and cross-selling campaigns.
Personalization
Personalization has been a marketing buzzword for over a decade. But for years, the practical effects of personalization were mainly limited to small-scale uses like chatbots or inserting your first name into email copy (“Hey {name}, here’s 15% off”).
With the help of artificial intelligence, personalization is finally having its moment.
AI systems can display special offers to specific categories of users, and can even show users the landing page or dynamic headline that’s most likely to help them convert based on their behavior or demographics.
Kindly.ai nudges website users toward conversion with subtle, well-timed popups that contain personalized recommendations. These AI-driven nudges increase conversion rates by up to 12% and increase average order value by up to 25%.
Octane.ai recommends a personalized set of products based on customer quiz data. Using zero-party data like age, location, behavior, and product needs, Octane can increase revenue and average order value while building a closer relationship with customers.
AI-powered website search tools give customers personalized query suggestions and results based on location, time of day, past searches, and purchase history. Doing this populates a dynamically-generated search results page and helps them find what they’re looking for faster.
Personalization isn’t just powerful for ecommerce. Uberflip.com uses buying intent data and AI content recommendations to power online sales engagement, creating personalized portals for each prospect and pulling in the sales collateral and messaging that’s most likely to be effective.
Get the basics right
So far, we’ve talked about using data-driven insights and customer feedback to queue up actions to test in your CRO campaign.
But there are also time-tested conversion rate optimization best practices that are worth experimenting with in any campaign you do. Give yourself permission to put the analytics, personalization, and AI software to the side for a moment during your CRO process.
Once the AI tools are put away, conduct a good old-fashioned audit. Do a hands-on review of your site. Make sure it has easy-to-read text, a clear call to action, and a compelling benefit. Does the design flow well? Is the copy engaging?
Here are some basics to cover:
Design
Make your design easy to navigate. Don’t make users guess where to go next. Within a few seconds, users should be able to understand the point of the page, recognize the next step you’re asking them to take, and realize the benefits of taking action.
Stick to tried-and-true design layouts. UX research has shown that certain layouts (like an F-shaped layout) are more easily comprehended. Use your layout to lead the user’s eye to the most important elements on the page.
Make sure you remove unnecessary distractions from your landing pages. While it may seem intuitive that videos would increase user engagement, many CRO studies have shown simple images or text outperforming video. If it’s not directly helping the goal of conversion, remove it.
Check your website speed. Loading time can be a barrier to conversion. If your website isn’t loading fast enough, users may bounce before ever having a chance to read your copy and understand your offer.
Copy
Speak to your audience’s pain points. Use techniques like loss framing, gain framing, and price anchoring to properly frame your offer’s benefits.
Write skimmable copy—users only read 20-28% of the content on your page. Communicate clearly and concisely.
Your heatmaps will likely show you that most people read the headline, subheader, and call to action, and skim the rest of your page. Pay special attention to these three text elements. AI copywriting tools can help you test different copy variants for each of them.
Conversion
One of the cardinal sins of landing pages is targeting multiple conversions on the same page. Stick to a single conversion goal that matches your visitor’s intent.
Your call to action (CTA) is one of the most important parts of your page. Make sure it stands out and is visible above the fold. Use action-oriented language that makes the benefit of clicking the CTA clear.
Make it easy to convert. Use short forms or multi-step forms, rather than having users wade through long forms or complicated sign up processes.
One reason users don’t take action is fear and uncertainty. Do everything you can to alleviate their concerns.
Social proof like reviews and testimonials can have a powerful impact on conversions.
Using free trials and money-back guarantees lowers the perceived risk of taking action.
Build trust by prominently featuring awards, certifications, and customer photos.
Testing
By this point, your queue of CRO tactics to explore should be bursting. From predictive models to customer interviews to personalization, there’s a lot to cover.
But to validate which changes have the biggest impact, you’ll need to test.
Traditional A/B tests are slow and inefficient. On top of taking weeks to complete and requiring large amounts of data, they can only test a couple variants at a time and don’t provide any optimization benefits until the test is completed.
Modern A/B testing uses AI to allow you to test more ideas faster while increasing personalization. Tools like ABtesting.ai provide automated suggestions for your headline, copy, and call to action, reducing the time and effort needed to conduct tests.
AI-powered A/B tests require only small amounts of traffic, can test many variables at once, and automatically start showing the better-performing variants during the test. Different versions of a page can be shown to different people based on their behavior, location, and other attributes.
Once you choose a testing tool, it’s time to go through your list of changes and hypotheses and start experimenting.
Begin by making sure you get the fundamentals right—website speed, design, copy, and conversion elements. Take a look at your offer and make sure it’s being framed attractively.
You may not need to measure or A/B test every one of these factors, especially if they’re almost certain to have a positive impact on your conversion rate (as with website speed, for example).
Then, experiment. Review the insights generated from the customer data you’ve collected and adjust your website experience accordingly.
Use real heatmaps or AI heatmaps and figure out which elements of your site users are actually looking at.
Test elements of personalization like dynamic headlines, personalized search and product or article recommendations, and AI-powered nudges and popups.
Follow the data. Don’t get too attached to the changes you’re testing. Let the system decide which variant to show based on performance.
Optimizing with the power of AI
AI is revolutionizing the practice of conversion rate optimization.
Instead of waiting until months after launch to optimize your site, you can optimize during development by using AI to simulate how users are likely to behave on your site.
Instead of manually sorting through mountains of user data, you can use AI to crunch data from multiple sources and present insights.
Instead of queuing up multiple weeks-long A/B tests, you can change multiple variables within a single test—and you can trust your AI-powered optimization system to deliver the right version of the landing page to the right user.
From predictive modeling to autocapture software, the world of AI-powered CRO is changing fast. The only way to keep up? Keep testing, keep optimizing, and keep iterating.