A/B testing

The safe method for long-term scaling of online shops

Online stores face intense competition and in order to survive they need to be smart about how they reach their target audience. With the rise of e-commerce, businesses need to ensure their website is user-friendly, visually appealing, and optimized to convert visitors into customers. An effective way to maximize conversion rates is through A/B testing.

Many only know A/B testing from Facebook ads or newsletter subject lines. However, A/B testing shows its real strength in the online shop. Because here A/B tests are an incredibly powerful tool to optimize websites and directly increase sales. A/B testing compares two versions of a website to see which performs better.

This allows online stores to identify which changes to their site result in the highest conversion rate, whether it's changing the color of a call-to-action (CTA) button or placing a product image. A/B testing is a powerful tool for online stores because it provides data-backed insights into what works and what doesn't, allowing them to make informed decisions about their website design and marketing strategies.

One always talks about "A/B" testing, but in the end there can be any number of versions that are tested against each other. Then, in addition to variants A and B, there are also variants C, D, E, F and many more. In multivariate testing, a particularly complex variant of A/B testing, there can also be hundreds of variants that are tested against each other. This way you can quickly find out which version brings more sales and profit. A/B tests are nothing really new. But advances in technology have made it easier than ever to conduct experiments and generate meaningful insights.

The importance of A/B testing for online shops

A/B testing is vital for online stores as it helps increase conversion rates and increase sales. According to a survey by ConversionXL, A/B testing can lead to an average increase in conversion rate of 14.79 %. This is a significant improvement and can lead to significant growth for online stores. Studies also show that a significantly higher conversion rate increase is possible if testing is not carried out purely in-house, but professionals with years of experience take over the optimization or are integrated into the testing team.

A/B testing not only improves conversion rates, but also offers online stores valuable insights into their target audience. By testing different elements of their website, online stores can determine what appeals to their target audience and what doesn't. This information is invaluable to online stores, allowing them to make data-backed decisions about their website design and marketing strategies.

A/B testing also helps online stores identify opportunities for improvement. If an online store tests the placement of a product image, among other things, and finds that a particular placement results in a higher conversion rate, they can make that change to their website and continue to improve their conversion rate over time.

Successfully design A/B tests for online shops

Successful A/B testing requires a few very specific things:

  • One hypothesis: Make an assumption about how changes to the website will affect customer behavior.
  • sufficient amount of data: Because A/B testing is based on statistics, it's important that you have enough data points to draw reliable conclusions.
  • test duration: Depending on the size of the variation in question and the accuracy you want, allow enough time for your test.
  • Analysis & interpretation: Results are only useful if they are evaluated correctly so that meaningful insights can be derived from them.

Define a clear hypothesis

Before an online store can start A/B testing, it needs to come up with a clear hypothesis. A hypothesis is a statement that explains what the online store believes will happen if a change is made. For example, an online store might hypothesize that changing the color of the CTA (“Add to Cart”) button from green to red will result in a higher conversion rate.

Choose the right items to test

Once an online store has defined its hypothesis, it needs to choose the right elements for the test. Some of the most tested elements on an online store website include the following:

  • headers
  • headlines
  • sliders
  • product images
  • (call to action) buttons
  • lyrics
  • form fields
  • Check out

Even small details such as color and font size can have an impact on sales figures - and can of course be tested. By testing different elements against each other, you can find out which version works best for your business and online store. Because of course there are many best practices that generally ensure more sales in the online shop, but when an online shop is really running, it is no longer the best practices that ensure more sales, but the individual optimization of the online shop for them specific target group with their specific requirements and wishes.

The right amount of data

Sufficient data volume is particularly relevant. If the sample size is relatively small, incorrect conclusions can be drawn from the test results. You always have to make sure that enough visitors see the test variants during an A/B test so that meaningful conclusions can be drawn from the results. Exactly how many "enough" visitors are depends on several factors (in particular the current conversion rate and the extent of the changes in the test variants), which can be calculated in advance.

As a "house number" one can assume that a shop with around 2,500 visitors/day is suitable for A/B testing.

Choose the right location for the A/B test

In online shops in particular, tests are therefore often not carried out on individual pages (URLs), but on entire templates. The advantage is obvious: more traffic and thus faster and more meaningful results. Such templates are usually page types, like

  • product detail page,
  • category pages,
  • distribution pages,
  • blog articles/SEO pages,
  • landing pages,
  • checkout process.

Of course you can also test on other sites. The home page and the order confirmation page often have a high potential here. However, meaningful results can only be expected if enough visitors take part in the test.

The right A/B testing tool

Once an online store has selected the items to test, they need to set up A/B testing. The basis for this is an A/B testing tool that is integrated into the online shop. For a long time, A/B testing was extremely easy because Google offered an extremely powerful A/B testing tool for free. With "Google Optimize" one could carry out practically all varieties of A/B testing in all situations. And you had the enormous advantage that there was deep integration with Google Analytics (Universal Analytics). But Google has since discontinued the free version of its tool, leaving online stores without an easy-to-use A/B testing tool and looking for an alternative.

Luckily, there are alternatives to Google Optimize that are specifically designed for ecommerce. Most of these are not free, but some providers offer free entry-level versions of their tools. Because one thing is clear: once you start A/B testing, you never want to stop. Some examples of this are VWO (Visual Website Optimize), Webtrends Optimize and Kameleoon.

It is important to consider data protection laws when choosing an A/B testing tool. Because A/B testing goes much further than web analysis. Not only is visitor behavior recorded, but experiments are carried out to determine which changes which visitors react to.

This means that the data protection laws of each individual country must be taken into account. Therefore, an A/B testing tool should provide secure storage and lawful use of personal data. It is important to ensure that the tool used provides a secure environment in which visitor data is treated in accordance with the rules of local and international law. A good foundation is having a privacy policy that is transparent and easy for your site visitors to understand. In addition, privacy settings in the tool itself should be easy to adjust.

No burden on the IT department

The beauty of A/B testing is that it takes place in the user's browser without changing the shop. This is all done by the A/B testing tool, which is usually integrated into the shop with a single line of JavaScript - and sometimes even more easily via a tag manager. The browser then loads the A/B testing tool and executes the JavaScript it provides in the browser, showing the changes. A cookie ensures that in the event of a conversion, this is assigned to the correct test variant. With A/B testing, we are therefore completely independent of all technical problems and the resources of the IT department.

Conducting A/B tests

One often hears that when setting up A/B testing, it is important to ensure that the two versions of your website are as similar as possible, except for the element being tested. You can call this the “pure doctrine” of A/B testing. In essence, this is also correct - if you want to make scientific evaluations of each individual element of the shop. However, if you are testing mainly with the goal of increasing the conversion rate, it often makes sense to test larger concepts in order to quickly achieve a significant increase in the conversion rate.

This is because each and every A/B test takes a sufficient amount of time to produce results that can be trusted, ie the results are accurate and reliable. How long an A/B test should run depends on the traffic on the website, but more specifically on how large the measured change in conversion rate is in the test. Therefore, the versions must be designed in such a way that they contain a significant change (“minimal detectable effect”), so that the change in user behavior is very clear. At a minimum, however, tests should run for a whole week so that fluctuations in the test results due to different behavior on weekdays are compensated for.

Analyze the results

Once the A/B test has run long enough, it's time to analyze the results. Online stores should compare the conversion rates for the two versions of their website to determine which version performed better. A wide variety of segments must be taken into account and analyzed in more detail. It must be ensured that the results apply to all segments - or one finds the segments that behaved significantly differently and in which one can now test adapted hypotheses.

All of this counts towards static significance, which is ultimately the basis for making truly informed decisions. Statistical significance refers to the likelihood that the results of a test are not due to chance. It's important for online stores to ensure that the results of their A/B tests are statistically significant before making any decisions based on the results.

Make informed decisions

Based on the results of the A/B test, online stores should make informed decisions about what changes to make to their website. If a particular change is found to result in a higher conversion rate, then that change should be made permanent in the store. This does not have to be done directly as an adjustment in the template, but can first be implemented as personalization in the A/B testing tool. Only when a large number of necessary adjustments have been determined can you then collectively pass them on to development and have the template adjusted.

The cycle of A/B testing

It's important to note that A/B testing is an ongoing process. Online stores need to constantly test different elements of their website to ensure they are always making the most of their website's potential to convert visitors into customers. The optimization is becoming more and more detailed. If you test rough changes at the beginning, such as hiding entire teaser elements for all users, later optimization is aimed at specific target groups or segments (mobile, desktop, traffic sources, entry pages, etc.). The transition from A/B testing to personalization is then smooth and the shop automatically becomes massively more user-friendly.

A/B testing as a competitive advantage

A/B testing is a powerful tool for online stores that can help maximize conversion rates and increase sales. By following the steps outlined above, online stores can turn A/B testing into a success and continue to improve their website and marketing strategies over time.

Online stores that invest in A/B testing can gain a competitive edge by making data-backed decisions about their website design and marketing strategies. By constantly testing and improving their website, online stores can ensure that they always offer their target audience the best possible experience and convert more visitors into customers.

It's important to approach A/B testing with an open mind and a willingness to try new things. While it's always a good idea to come up with a hypothesis, online stores should be prepared for the results to challenge their assumptions and provide new insights about their target audience.

In summary, A/B testing is a simple and effective way for online stores to optimize their website and marketing strategies and of course to increase conversion rate.

It's important to remember that A/B testing isn't a one-off solution, but rather an ongoing process of experimentation and improvement. Online stores should always be open to trying new things, analyzing the results and making informed decisions based on data. By investing time and resources in A/B testing, online stores can gain a significant competitive advantage and be successful in the long run.

In the fast-paced world of e-commerce, it is imperative for online stores to stay ahead of the curve and keep up with the latest trends and best practices. A/B testing is one of the most effective ways for online stores to stay ahead of the competition and grow their business. So, what are you waiting for? Start A/B testing today and see the positive impact it can have on conversion rates and the overall success of your online store.