Electronics Giant Unlocks Surprising Wins with Segmentation

Sony optimizes to drive 20% more visitors through the configuration checkout funnel

As a leading global electronics company, Sony knows providing the right experience for online shoppers is critical to business. The company uses many different marketing tactics to personalize the customer journey to increase purchases on its online store.

Evelien Geerens is a web merchandiser at Sony who uses Optimizely to increase conversions across the company’s web merchandising platforms. During a recent site audit, Evelien discovered that banner ads promoting customizable laptops on Sony’s homepage and product pages were underperforming – clickthroughs and purchases from the banners were extremely low. Evelien set out to fix this conversion problem with A/B testing.

CHALLENGE #1: Low Banner Ad Conversion

The banner ads in question presented two different calls-to-action – one advertising the customizable Sony Vaio notebooks and one promoting a recent sitewide offer for a free memory upgrade.

Evelien hypothesized that displaying dual CTA messages was confusing and overwhelming to site visitors, making them less likely to click through on the banner. However, her team’s usability research suggested otherwise. In Sony’s UX tests, Evelien learned that some site visitors were actually put off by the customizable laptop option. These visitors believed the customization process would be too timeintensive and not worth the outcome.

Faced with contradicting opinions, Evelien turned to Optimizely to let data make the final decision.


Using Optimizely, Evelien built an A/B/C test to test two different variations of the banner against the original. With Optimizely’s targeting feature, Evelien targeted the experiment only to site visitors from the Netherlands and the United Kingdom – two of Sony’s largest markets for the customizable laptop campaign. One variation focused entirely on the customizable laptops, while the other focused only on the sitewide offer.

Testing helps you make decisions based on objective results, not subjective guesses. It’s a good way to be in touch with how the customers react to the content of your website.

— Evelien Geerens, Web Merchandiser, Sony

Evelien’s goal was to measure how each banner impacted the number of visitors who clicked through the banner and then eventually entered the checkout funnel. She measured this by setting two goals – a click goal on the banner itself and a pageview goal to measure each time site visitors reached shopping cart landing pages after clicking through the banners.

The original banner displayed messaging around the personalization campaign and the promotion.

Variation 1 focused entirely on the idea of customization. The banner included a “Vaio by you” logo, and a call-to-action to create and customize a laptop. The banner did not reference the sitewide offer.

Variation 2 focused on the promotional aspect, showing specific messaging to create urgency around the offer. This variation made no reference to personalization.


Variation 1, the banner focused on customization, saw a 6% increase in banner click-throughs and an uplift of 21.3% in visitors that reached the shopping cart compared to the original banner.

Variation 2, the banner focused on the sitewide promotion, only increased click-throughs by 1.8% and actually performed worse than the original in terms of shopping cart views, decreasing conversions by 2.9%.

Post-Test Segmentation

As the data came in, Evelien used Optimizely’s segmentation feature to dig deeper – breaking down the test results by key user segments to understand where there might be gaps or outlying data. Segmentation paints a more complete picture of test results and helps testers like Evelien pinpoint areas to test next.

First, Evelien segmented based on geolocation – the Netherlands versus the United Kingdom – to ensure both key markets were performing comparably throughout the test. The results were similar, indicating that the customization banner experience was a strong choice to increase conversions in both markets.

Next, she segmented the results based on device type – mobile versus desktop. In these segments, she uncovered surprising results around the banner click-through rates. The trend toward customization was reversed for mobile visitors, who were more interested in clicking on the promotional banner. Variation 2, the banner that focused solely on the sitewide promotion, increased mobile click-through conversions by 21% compared to the original, while variation 1, the banner focused on customization, only increased click-throughs by only 16%.

Based on this insight, Evelien’s team is currently developing a series of followup tests targeting different promotional options to Sony’s mobile audience to maximize revenue from visitors on all devices.

CHALLENGE #2: High Bounce Rates in the Checkout Funnel

With a clear understanding of how to best drive mobile and desktop visitors into the checkout funnel, Evelien ran several followup tests to learn what would keep them there and compel them to make a final purchase.

Her first test focused on customers traveling through the customized laptop checkout process. Using analytics to gather metrics from each step in the funnel, Evelien found out that after the second step – the configuration of customizable components for each laptop – 39% of all visitors abandoned the buying process. She hoped to find a way to keep these users from bouncing.


Evelien believed that shorter product descriptions would make it easier and faster for customers to make an informed choice on the customizable components page. This, in turn, would drive more visitors to continue through the checkout funnel.


Using Optimizely, Evelien set up an A/B test, pitting the original funnel page against a variation with shorter, more digestible product descriptions. In addition to this change, the variation page changed the term “configuration” to “components”, added top seller tags to indicate popular choices, and highlighted promotions on the page. To track purchases from each variation, Evelien set a pageview goal on the order acknowledgment page – where customers land after making a successful purchase.

The original funnel page (Left) displayed long product descriptions, and did not highlight top-sellers or promotions. The variation funnel page displayed shorter product descriptions. It also changed the term “configuration” to “components”, included top seller tags, and highlighted promotions.


The variation page outperformed the original – with 20.6% more visitors reaching the order acknowledgement page. This meant a measurable increase in revenue for Sony.


Dive Deeper Into Test Results with Segmentation

No two site visitors are the same. Understanding how specific visitor segments interact with your site will help you provide a more personalized experience to maximize conversions. Evelien segmented her results based on device type. Segmentation is also available based on a number of other parameters, such as browser type, campaign, referral source, custom segments, and more.

Challenge Assumptions

Sony’s qualitative user research suggested that customers were not interested in the laptop customization campaign but Evelien’s instinct told her otherwise. Testing the options against each other provided clear, quantitative data to back up Evelien’s hypothesis and convince her team to keep the campaign running.

Sometimes Less is More. Reduce Choices to Increase Conversions

Evelien’s tests on Sony’s webstore revealed that paring down options and information throughout the checkout funnel can actually provide a more clear path for site visitors. Test different options to understand how much information your site visitors prefer.


Increase in Banner Ad Clicks


More Shoppers Engage with Cart


Increase in Checkout Conversions

By changing our website based on test results, we can better meet the online expectations of our customers and, at the same time, increase conversions.
Evelien Geerens

Web Merchandiser, Sony

Get Started with Optimizely

It's your turn. How will you experiment everywhere?