A green background with squiggle designs
A/B Testing: What We’re Learning
Small design choices can change what users do. Our A/B tests help us understand which choices make important actions easier to find—and which changes do not meaningfully affect behavior.

AB Testing Experiments

A/B testing in plain language

We compare two versions and watch what users do.

An A/B test shows different versions of the same webpage or web element to similar groups of visitors. One group might see the page as it currently exists, while another sees a version with one specific change—such as a button being removed, moved or styled differently.

We then compare how visitors interact with each version. Do they click? Do they find another path? Does the change make an important action easier or harder to complete?

The goal is simple: make web decisions based on how real users behave, not assumptions about what we think they might prefer. Over time, these tests help us better understand what works across ohio.edu and turn those findings into practical guidance for creating clearer, more effective webpages.

A/B Test #1: Do hero buttons help users take action?

We tested what happened when important CTA buttons—such as Apply, Visit, Admissions and Request Info—were removed from the hero area. Everything else on the page stayed available.

  •  

    Control: Hero Call-to-Action Buttons Present

    Users can immediately see the pages's most important next steps.

  •  

    Variant: Hero Call-to-Action Button Removed

    The same actions may still exist in navigation, body links, cards or the footer—but users must search on page for them.

  • Clear Paths Lead to More Action

    The pattern was consistent: when the hero CTA disappeared, users clicked other locations more often—but those alternate paths recovered only about 27% of the conversions that were lost. This reinforces the importance of making your most important actions easy to find.

Results from A/B Test #1:

−71%
conversion-rate loss when hero CTAs were removed from page #1 in this test
−60%
composite conversion-rate loss across key actions from page #2 in this test
−90%
composite conversion-rate loss across Apply and Request Info actions from page #3 in this test

What this means for web editors

Make the most important next step easy to see. A link existing somewhere on a page is not the same as a link being easy to find. If an action matters to the user journey, give it clear visual priority.

  • Do

    • Identify the primary action for the page.
    • Place high-priority actions prominently, especially on high-intent pages.
    • Use clear CTA labels such as Apply Now, Schedule a Visit or Request Info.
    • Repeat an important action in logical places when users may be ready at different points in the page.
    • Use navigation, cards and body links as supporting paths—not replacements for the primary CTA.

  • Don't

    • Assume users will “find their way” to an important link.
    • Remove a primary hero CTA only because the same link exists elsewhere.
    • Use vague labels such as Click Here when a specific action is available.
    • Treat every page the same—start with the user’s intent and the page goal.

A/B Test #2: Does changing promo box color increase clicks?

On the /apply page, we compared different background-color treatments for the Undergraduate Application Methods promo boxes.

  •  

    Control: Trillium White

    2.34% Conversion Rate — No meaningful improvement even though higher conversion rate

  •  

    Variant Mixed Colors

    2.19% Conversion Rate — No meaningful improvement

  •  

    Teal Variant

    2.30% Conversion Rate — No meaningful improvement

  • The takeaway from A/B Test #2:

    color alone did not meaningfully change user behavior (conversion rate). On a task-focused page, clear labels, hierarchy and easy-to-understand choices are more useful places to focus than decorative color changes.

Four questions to ask before you publish.

What is the goal? Name the one or two actions that matter most on this page.
Can users see it? Important actions should be visible and easy to recognize without searching.
Is the label clear? Tell users exactly what happens next: Apply, Visit, Request Info, Register.
Is anything competing? Reduce visual noise so secondary content does not overpower the primary task.

Want the details about these tests?

The guidance above is intentionally simple. Expand these sections if you want more context about how the tests were interpreted.

How were the hero CTA tests evaluated?

We compared a control experience with the standard hero CTA button(s) to a variant where a target hero CTA was removed. Other navigation and link locations remained available. The confirmed high-traffic pages showed statistically significant declines in key actions when hero CTAs were removed.

Why do we say only 27% was recovered?

Users did shift some clicks to navigation, body links, cards, featured media and the footer after the hero CTAs were removed. In other words, some users who could no longer find the action in the hero searched elsewhere on the page and eventually found another way to reach the same destination.

However, those additional clicks replaced only about 27% of the conversions that were lost when the hero CTA was removed. The remaining clicks were not made up elsewhere on the page.

This is what we mean by “recovered.” Users did try alternate paths, but those paths were not nearly as effective as providing the action prominently in the hero. The finding reinforces that simply having an important link somewhere on a webpage does not mean users will find it. For high-priority actions, clear and prominent placement can make it much easier for users to take the next step.

Does every page need a hero button?

No. The lesson is not “put a button on every hero.” The lesson is to make high-priority actions prominent when the page has a clear next step. For low-traffic or informational pages, page goals and user needs should still guide the design.

What we're testing next?

Higher-impact tests may include CTA wording, information hierarchy, ordering of choices, microcopy clarity, spacing and other ways of emphasizing important content.

  • Putting These Findings Into Practice

    Curious about how your own CTAs are performing? UCM has an Analytics Dashboard available for all ohio.edu websites, giving web editors access to user behavior data for their own sites. You can see where users click, how they navigate through your site, which content they engage with, and where they may be leaving.

    Have a specific question or want to dig deeper? Schedule a personalized consultation with UCM. We can analyze additional Google Analytics data, help you understand how users are interacting with your content, and identify actionable opportunities to improve your webpages.

    To request a consultation, use the Project Request Form and select Web → Web Analytics/Dashboard Support Request.

    Schedule a Consultations

Web Analytics Report Training Instructions

Additional Analytics Resources