Background
Optimising an ecommerce store takes time, expertise and the right tools. Learning which tool is right for you can be the first stumbling block for many brands when considering growth and user experience (UX) improvements.
Whilst there are rules of best practice that can be applied, the nuances of UX and conversion will always vary because each store is unique.
Swanky has helped multiple brands increase key ecommerce metrics by testing and optimising their Shopify stores. Get in touch with our solutions team to learn how they, along with Webtrends Optimize, could support your brand’s next chapter of growth.
In this post, I will explore why a luxury fashion brand chose Webtrends Optimize for AB testing and optimisation. I’ll demonstrate what this looked like in practice, with details of some of the tests that have been implemented and the results achieved.
Why we selected Webtrends Optimize
Firstly, a quick look at why Webtrends Optimize was a good fit for the client. There were three main points:
1. Scalable packages
When looking for an AB testing tool, the client had a number of desired features in mind. One of these was a tool which could scale with the brand - and Webtrends Optimize does just this. Unlike other testing tools, Webtrends Optimize doesn’t block certain functionality within the platform. Instead, it with scalable packages for stores that grow.
Beyond experimentation, Webtrends Optimize offers several supplementary features for brands to enhance UX, including product recommendations, social proof messaging and data-driven merchandising solutions.
One of these additional features that the client leveraged is the use of recommendations. As a luxury fashion ecommerce brand that is committed to curating thoughtful and relevant product suggestions, each product page should present meaningful recommendations to the customer, sharing other styles they might be interested in.
Ultimately, the aim is to emulate the experience of being in a brick-and-mortar store, getting support from a stylist.
Whilst the client wanted to maintain control of this to ensure product curation is kept to a high standard, the Webtrends Optimize recommendation engine has allowed the brand to leverage automation to improve how suggestions are displayed on-site.
For example, if a product from the client’s selection is out of stock, Webtrends Optimize can fill that recommendation slot with a secondary option. The platform identifies this alternative product using an algorithm designed to determine the best option based on the priority of conditions outlined by the client.
2. Understanding the customer journey
The client was eager to understand their on-site customer journey in more detail and use this data to create a prioritised AB testing strategy.
A data-focused solution is central to achieving this; something that is exemplified in Webtrends Optimize’s baseline analysis tool. This tool was a particularly appealing aspect of Webtrends Optimize's offering for the client, as it can assist in behaviour tracking and analytics.
3. Creating different experiences for different users
Lastly, the client was seeking a tool that enabled them to control the on-site experience of different user groups. For example, they wanted to serve unique experiences for users who have logged in, users who have added a product to their bag, and users in different regions/countries.
Building on the learnings from the baseline tool and pairing this with insight from other analytics tools in their tech stack, Webtrends Optimize would allow the client to identify opportunities to create more personalised experiences.
The experiments and their results
Below I outline a series of AB tests we implemented for the client using Webtrends Optimize. The learnings from each test fed into the next, emphasising the benefits of ecommerce testing beyond ‘winning’.
1. Testing product tile layout on mobile collection pages
The layout of imagery should always be optimised for mobile screens, as it’s true for most stores that the majority of users are viewing it on a mobile device. When it comes to the specifics of page layout and content configuration, testing is key for gaining valuable insights.
With this in mind, we conducted a test to learn how best to present product imagery on the client’s collection pages for mobile devices.
In the test’s control, the default collection page view showed four images (or product tiles) at a time.
In the variant, we altered the page’s functionality so that the single image view was the default option on mobile. This allowed users to see each product image in much more detail, with the ability to swipe to see photos from different angles; an action that users are familiar with on other websites and apps.
The Results
The variant produced a statistically significant increase in conversion rate of 31.52%. We calculated that if the test were to be pushed live to 100% of users, it would generate six-figures of additional revenue across a 12-month period.
Based on this information, we declared the variant to be the winning execution. This test highlighted the benefits in allowing users to see the collection page imagery on a larger scale.
Learning from this, we started to hypothesise where else we could optimise content on the client’s site, with the aim of better showcasing the intricacy of their products.
2. Testing placement of video on product pages
This led us to our next test. Video content is known to engage customers for longer, with 84% of marketers noting that video marketing increases dwell time on their website.
Unfortunately, it isn’t as simple as replacing imagery with video and seeing an uptick in conversion rate. This is because there are other factors, such as site speed, at play.
Most of the client’s product pages already contained video content that presented the beautiful detail of their product. However, this content was often confined to the final slot in the image carousel.
We were interested to learn whether prioritising the video content on these pages - by placing videos in the first slot in the carousel - would affect how users interacted with the page.
This built upon our findings in the previous test which showed that promoting visual content more prominently on the collection page led to an increase in performance.
The Results
Analysis revealed that whilst transaction conversion rate increased for the variant, ATC (Add To Cart) rate actually reduced. This suggested to us that while fewer visitors added items to their cart after interacting with the variant compared to the control, those that did add products were more likely to follow through and complete their purchase because of the positive impact of the video content.
We could therefore not declare the variant an overall winner without further exploration around the drop in ATC rate.
When we looked into this further, we identified that the poor ATC rate was most evident amongst mobile users, which indicated a problem with the experience on mobile devices. This led us to hypothesise that perhaps poorer connection speeds on mobile devices (due to these often relying on data rather than WiFi) were causing videos to load slowly and not play smoothly, thereby creating a poor product page UX for these customers.
3. Testing thumbnails & video CTAs on product pages
The next test we implemented continued our exploration of video use on the client’s Shopify store and considered how we could reduce the burden of video content.
The control remained the same as in the test outlined above.
In the variant, rather than having the video load through as default, we included a video CTA - a play button - to take users to a video in the main viewport of the page. This allowed us to reduce the burden of loading the video for users with poorer connections, but still increase the prominence of the video functionality on the product page.
We also redesigned the thumbnail format to make all content easier to access for users and reduce the amount of space images took up on desktop. These changes made the content easier to navigate and view.
The Results
In this test, the variant saw a conversion rate increase of 17.54% compared to the control, which achieved statistical significance. ATC rate also increased for the variant.
Based on these results, we declared the variant the winner.
With this variant being rolled out to 100% of traffic, we predicted a 12.7% uplift in revenue per user. This version of the product page was later pushed live.
4. Product Recommendations
The client wanted to use Webtrends Optimize’s product recommendations tool to automate and filter out their out-of-stock items, as they were currently spending a lot of time manually updating this process.
We discussed the problem with Webtrends Optimize and they created a bespoke product recommendations engine to enable this.
The way the tool works is by allowing the client’s team to set rules around the attributes of recommended products if the original desired product is out-of-stock. The tool ensures that the chosen elements of the product are a match with the out-of-stock item.
For example, you could set a rule that any recommended product must be the same length as the original item, or the same colour.
You could also set negotiable attributes, for example, the price. This can be configured so that all other elements of the product recommendations match the out-of-stock item, but the price can vary.
The Results
Prior to using the tool, the client would spend about 40 hours a month on manually updating their stock.
The utilisation of the tool has meant their shoppers can still benefit from a ‘human intuition’ approach, whilst saving the client valuable time that can be used elsewhere.
Webtrends Optimize created the entire back end of the tool for us, ensuring the audits were all up to date and nothing was showing by chance. The client also had easy access to the tool too so they can also check anything they need to within the process.
What Next?
For any client working with Swanky’s Growth Accelerator team, we are always thinking about what’s next in the optimisation process. In the case of this client, further Webtrends Optimize powered testing is in the pipeline to build on the learnings detailed above and help better communicate the USPs of this digital-first brand.
Another key area of focus is building trust in the client’s brand, as well as in the quality of its products. We are therefore considering strategies to boost visitors’ confidence in making purchases and to improve the service customers receive.
Whilst this case study has focused on test results for the UK market, we are also looking to adapt content to the international markets that the client is operating in.
This requires us to understand the ways users shop in different areas of the world and implement more localised strategies - something that Webtrends Optimize will be able to support.
