Get in touch!

we respect your privacy and wont share your information. By entering information in this form you accept CXperts' privacy policy
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Close Window

The Five Levels of Conversion Rate Optimization (CRO)

By Steven Shyne

Much has been said about the “digital maturity model” in recent years. While this discussion can make some good points, we at CXperts found it to be an overbroad topic for many individuals and brands to be practical and useful. 

In taking a narrower approach than the expansive world of “digital,” we decided it’s worth a look at the much more narrow subject of website experimentation and how more basic tactics can lead to more advanced ones and so on. That said, know that these tactics are additive: your experimentation program won’t ever leave A/B Testing behind; rather, you’ll add to it, increasing in complexity and sophistication.

Before we dive in it’s worth pointing out that any firm engaged in - or considering - digital experimentation is in the right mindset to provide an ever-improving user experience and possibly even outshine the competition. So wherever your brand is in the process, it’s never a bad idea to start - or advance - web and digital experimentation to inform small improvements that can compound into big results over time. 

Here are the core tactics that help define the experimentation maturity roadmap, plus some real-life examples of winning tests from current and previous clients:

A/B Testing

Also known as “A/B/n Testing," A/B Testing is a process of scientifically testing different digital components in a controlled manner by creating one or more variations of one particular element and having users randomly but equally exposed to those variations, measuring what, if any, differences those variations have on user behavior. A/B tests are great because you can test and tweak individual components – like buttons, copy, layout – or whole pages, or even other digital experiences, like email CTAs. A/B Testing is the basic building block of many other Conversion Optimization tools, like Multivariate Testing, Multi-Page Testing, and Dynamic Website Personalization. This is the one to start with and to master.

Example: Your ecommerce site cart gets a lot of traffic but not a lot of continuation. Your UX team suggests there aren’t a lot of confidence signals on the page, and you want to improve that by adding a badge as a trust signal. You work with design to create a few variations, and deploy a handful of versions to see which badge has the most positive impact on conversion. You're controlling for one change only – the trust badge – and seeing which variation performs the best.

Split Page Test (Split URL Test)

Similar to A/B/n testing, Split Page Tests (also known as Split URL Tests) are macro-level, where you evenly split traffic between or among two or several different page versions and measure the overall impact to the shared page goals. This can be a great way to pit two designs against one another in a winner-take-all showdown. Or if you want to confidently burn a page down and start over, do it strategically with a Split Page Test. Key point: whenever testing multiple pages we become less sure of exactly what elements are contributing to the changes in behavior, just that there are changes and we’ve observed them. 

Example: your B2B paid media landing pages are meant to drive demo requests, and your team is sick and tired of the existing, so-so design. You have two designers that have each created new comps, and rather than having your team pick favorites, you build and deploy both, and through Split Page Testing you let your audience determine which page is more effective at driving demos.

Multivariate Test

Multivariate Tests (MVT) are like A/B Tests, in that they are a tactic to scientifically test different digital components in a controlled manner, but in MVTs you create several variations of several elements on a page as a way to understand the co-unit interaction (or interaction patterns across multiple, tested elements), and measure all the different combinations of changes. Multivariate tests take a lot more energy and planning to build and more time and traffic to run, but are a worthwhile investment to measure and understand the combined effects of different experiences. 

Example: Your ecommerce site has all the basic components that users would expect (headline, details, photo, quantity, price and add to cart button), but through usability research you’ve uncovered it’s rather unintuitive. Instead of taking a guess, you build an MVT test that switches things around, trying to find the best possible grouping to improve add to cart events.  You're testing all possible combinations of different changes to see which of them is the best performing.

Multi-Page Test

Multi-Page Tests are also like an A/B Test (see the theme here?), but on a macro level: instead of individual components on one page, app or email, Multi-Page Tests are a way of testing entirely different user flows across several pages in a controlled, scientific manner. Or build a baseline for understanding personalized user flows and testing out new versions of critical paths, like signing up or checking out flow. Depending on the granularity of the changes, we may or may not be able to infer what changes exactly are contributing to the changes in behavior, but depending on the goals of the experiment this may or may not be a concern.

Example: Your B2B company is feeling pressured by outside competitors to display sample pricing throughout the site, but real-world pricing is difficult to estimate accurately and can also nudge potential prospects into bargain hunting mode. Through multi-page testing you can strip away mentions of pricing throughout the primary user flow and measure impact to leads generated. This helps you determine whether your website or your sales team are the best choice to handle the price conversation.

Dynamic Website Personalization

A website UX tactic wherein web content is automatically and usually without notice personalized to increase relevance in a user's website experience. This is typically done implicitly (without the user actively providing feedback on how the site should be personalized) and can range widely, from region-based content, tailored product recommendations, or even inserting users’ names or other personalized information into the web experience. The hardest part of Dynamic Website Personalization is often determining where to start.

Example: Your ecommerce site sells the same products but for different applications. Rather than building a different product detail page for each use case, you can use Dynamic Website Personalization to swap out content based on the category page that users came from, making a more relevant experience users and better conversion rate for the company.

Within each of these tactics, there are additional details to sort out when deploying tests, but the above should give you enough information to know how, why and when we use certain testing approaches.

If you read the above and are still not sure how to start, we can help figure that out for you at: