Cargill Empyreal75
A/B Testing

A/B Testing to improve your online marketing

Cargill Empyreal75

Visitas: 77

By: Sarah Cornelisse*

What gets you to open the emails that land in your inbox or to click on an advertisement that appears in your social media feed? Perhaps you are drawn to humor or to images of practical applications for products. Now, consider how you develop your own online marketing.

What gets you to open the emails that land in your inbox or to click on an advertisement that appears in your social media feed? Perhaps you are drawn to humor or to images of practical applications for products. Now, consider how you develop your own online marketing.

For instance, if there are yellow call-to-action (CTA) buttons on your website how did you select that color? You may have relied on generalized statistics or personal preferences when initially developing your business website or email newsletters.

As you generate and collect data on conversion, bounce, or open rates, or begin to spend marketing dollars on social media advertising, the value in optimizing all aspects of your online presence becomes apparent. You want your marketing budget to be spent wisely and effectively and this is where A/B Testing can assist.

A/B Testing

The principle of A/B Testing is similar to that of partial budgeting; it allows for the comparison in performance and effectiveness from making one change.

As stated by Go (2022), “The premise behind A/B Testing is comparing the performance of two or more variations to identify how different marketing strategies or activities perform with an audience. It’s designed to improve user experience (UX) and increase conversion rate optimization (CRO).”

“To illustrate, imagine that you are running a Facebook ad campaign and it is not generating the number of click-throughs that you would like. For the ad, you initially used an image of the product by itself, but you now speculate that an image of someone using the product would be more visually appealing, prompting greater a number of Facebook users to click on the ad.”

This type of situation is ideal for performing A/B Testing. There are numerous elements of your online presence that you can experiment with through A/B Testing. A selection of elements is outlined below and categorized by marketing tool.


✓ Call-to-action (CTA) buttons: size, color, placement, and text.

✓ Social media sharing buttons: placement, size.

✓ Headings: color, text, and size.

✓ Content: format (text, image, video, audio), text font size, style, placement, and structure.

✓ Video/Images: different composition and placement.

✓ Product descriptions: inclusion of reviews, format.

✓ Forms: amount and type of information requested.

Social Media

✓ Use of Image/video: text only vs. with image or video, image vs. GIF, people in image, video length, number of images.

✓ Hashtags: hashtag use vs. not, multiple vs. one, placement within the post.

✓ Post text: length, style (quote, statistic, statement, question), emoji use, punctuation, tone (casual, formal, passive, active).

Email/Email newsletters

✓ Email subject lines: questions, statements, emoji use.

✓ “From” name.

✓ Send time.

✓ Content: length, font size, style, placement, and structure/layout.

✓ Video/Images: composition, placement.

✓ Personalization.

✓ Preview text.

✓ Social media sharing buttons: placement, size.

Performing A/B Testing offers numerous benefits including:

✓ Quick results.

✓ Wide applicability to features of your online/digital presence.

✓ Reduced risk from making untested changes.

✓ Improved user engagement and experience (improving trust and customer retention).

✓ Enhanced content.

✓ Reduced bounce rates.

✓ Greater conversion rates.

✓ Higher conversion rate values.

✓ Reduced cart abandonment.

✓ Increased sales.

✓ A/B testing allows you to make data-driven decisions rather than decisions based on gut feelings and personal preferences.

Stated succinctly, “A/B Testing provides practical, tangible evidence of your most impactful marketing techniques” (Grayson, n.d.).

While there are a number of benefits to A/B Testing, there are also a few challenges to be aware of. As noted by Singh, V., Nanavati, B., Kar, A.K. et al. (2022), these include:

A/B Testing only works for specific goals. If the goal is difficult to measure, or unmeasurable, A/B Testing is not effective. As stated previously, A/B Testing is data-driven therefore you must have a variable that you can measure and track.

Long-term relevance. You must ask yourself if 6 or 12 months down the road whether the A/B Test would generate the same results.

Continuous testing. These ties to the previous challenge. For long-term relevancy, marketing should be tested regularly. Unfortunately, many A/B Tests are once-and-done exercises.

Time and resource-heavy. Since you are only testing one variable at a time, if you have numerous variables to test it could take quite a while to perform all of the desired A/B Tests. The slow nature of this process can see testing-related expenses accumulate quickly.

Suggested processes for performing an effective A/B Test include the following steps which closely align with the scientific method (Figure 1).

A/B Testing


1 Identify the problem/opportunities for improvement.

2. Set/know goals.

3. Develop a theory/hypothesis.

4. Create variation( s).

5. Run test(s).

6. Analyze results.

7. Adjust (or leave alone).

8. Monitor and repeat.

As with any process, there are some best practices that will aid in the effectiveness of A/B Testing. These include:

Know your goals. Have a clear vision for what it is that you want to accomplish in your business and online marketing. For instance, with social media is your goal brand awareness or sales maximization? What is it that you want consumers to do, or how do you want them to respond?

Have a clear question that you want to test. You need to be able to clearly articulate your hypothesis, how it will be measured, and how the outcome will help you better achieve your stated goals.

Test only one variable at a time. If you make more than one change, you can’t be certain which change the audience is responding to.

Prioritize testing variables to those that will have the greatest impact. To try to identify these, focus on bottle DIGITAL AND SOCIAL MARKETING BYTES necks or pain points that your consumers are experiencing. Monitoring analytic data for your website, social media, and emails will allow you to identify aspects that are proving to be hurdles with your consumer audience.

Allow sufficient time for your test(s) to run. While A/B Testing offers the advantage of providing quick results, you must still run tests for an adequate length of time to allow for variations in consumer activity. For example, running a test on an aspect of your social media between Monday and Friday only when you know that a large percentage of your audience is active on the weekends will result in lower statistical significance.

Get ideas from others. As with many aspects of business planning and operation, you can benefit from asking others (customers, social media audience, employees, etc.) for their input on what could be tested.

In addition to the benefits and challenges that have been previously covered, there are some additional factors to consider when performing A/B Testing. The first is the cost of implementation. Most A/B Tests can typically be easily created, and a change implemented if test results indicate to do so.

However, if either development of the test variation or change implementation requires changes to website code or significant additional work on the part of a hired marketing consultant, you should determine whether the benefit from performing the test will outweigh the additional costs.

“A second consideration is the demographics of your audience. A study looking at eye tracking and A/B Tests for a social media marketing campaign discovered that men and women were attracted to different types of components (Nichifor, 2021).”

It is reasonable to believe that other demographics, such as the age of the consumer may affect the effectiveness of marketing characteristics that you consider for A/B Testing. Due to the potential for demographic differences, try to gather and include demographic data along with the data you are collecting to measure A/B Test performance.

A/B Testing

Finally, different tools exist to use when performing A/B Testing. Social media platforms (Facebook, Instagram, etc.) offer A/B Testing functionality when creating advertisements. For websites, two options are Google Optimize and Optimizely.

“Most email marketing services (MailChimp, Constant Contact, etc.) also offer the ability to perform A/B Tests. For business owners or marketing managers interested in optimizing their online marketing, A/B Testing is a relatively easy-to-implement method to use.”

“The goal of every A/B Test is to find what variant drives more engagement, and then implement that change” (Wingerberg, 2021). It may take time to discover the most effective variant and variants may perform differently with different audiences.

Therefore, it is important to perform A/B Tests regularly, collect demographic data on your audience in addition to data for measuring your hypothesis, revisit tests that you previously performed, and stay abreast of marketing trends that influence consumers.

Where trade names appear, no discrimination is intended, and no endorsement by Penn State Extension is implied.

 Sarah Cornelisse

References cited by the author available under previous request to our editorial team.
*Sarah Cornelisse is a Senior Extension Associate of agricultural entrepreneurship and business management at Penn State University in the Department of Agricultural Economics, Sociology and Education.
Sarah has expertise in direct marketing, valueadded dairy entrepreneurship and marketing, the use of digital and social media for agricultural farm and food business marketing, and business and marketing planning and decision making.
Originally from New York State, she has a B.A in mathematics from the State University of New York at Geneseo, and M.S. degrees in Agricultural Economics and Animal Science, both from Penn State University.
Correspondence email:
Editor’s note: references cited by the author within the text are available under previous request to our editorial team.

Cargill Empyreal75

Leave a comment

Tu dirección de correo electrónico no será publicada. Los campos obligatorios están marcados con *