MATCHES: Driving conversion through a scalable CMS

In 2023, Matches experienced a 10% drop in monthly conversion rates over 3 months, especially in UK, our key market, which led to immediate revenue losses and weakened brand partnerships in the long run.

After looking into the data, I discovered that mobile users, accounting for nearly half of our traffic, and returning customers, who made up 65% of our base, were most affected.

At a high level, there are several internal and external factors that could contribute to this. Internally, the suboptimal mobile UX/UI experience and limited personalise content could hurt user engagement Externally, low demand due to recession, strong competition and a shift toward in-store shopping could also draw users away. Addressing these will be key to improving conversions.

To find out what we could do to optimise conversion rates efficiently,Firstly, I define the target segment as high-income millennial online shoppers who value personalized and curated shopping experiences—a key demographic for Matches.

Then, I collaborated with UX team and interviewed 2 target customers where I learned that users appreciated curated collections like holiday gift guides but were frustrated by irrelevant content.

This insight aligned with our internal data, which shows that seasonal campaigns drove 30% of annual revenue and converted 68% higher than the site average. And our personalised module like ‘shop similar’ achieved 12% higher conversions than other curated modules.

From this, our goal became clear: we needed a mobile-first, user-centered solution to make campaign content more engaging and relevant.

I also analyzed our own campaign process and benchmarked it against competitors like Net-a-Porter and Farfetch. From the table, we saw that Matches’ campaign experience was falling behind the industry as competitors were able to offer interactive, dynamic campaign pages with personalized recommendations, articles, and videos in just a few hours. In contrast, Matches CMS could only generate static listing pages by gender. A landing page similar to competitors would take over 4 weeks to custom-code and couldn’t be reused for future campaigns. This highlighted the urgency of building scalable CMS to bridge the gap and stay competitive in the luxury ecommerce market.

Next, I mapped the customer journey to capture the key pain points in the existing mobile experience for a gift campaign at Matches. Let’s take a look at Annie, a 30 year-old marketing manager based in London, a fashion lover and frequent online shopper who enjoys taking inspirations and discovering new brands online. She was looking for a perfect christmas gifts that would make her parents happy. She started to search for gift ideas online, compare prices and check user reviews. Since she had shopped with Matches before, she decided to also check potential gift options on the website. However, our gift listing presented key challenges such as frustrations to filter products by age group, complex navigation between genders on mobile, and lack of relevant products based on her purchase history, which drove her away to NAP and placed an order there.

From Annie’s customer journey, we discovered that limited support for product selection, poor navigation and static content at consideration phase were very likely to cause frustrations and drop-offs on campaign pages.

As such, I defined a product mission that aligns with Matches’ company mission. That is, redesigning the web experience to be mobile-first, frictionless, and personalised in high impact areas.

To achieve this, the tactic I proposed was to create a CMS that enabled editors to build personalized, modular campaign pages quickly and efficiently.

Based on the previous research and product strategy, it was clear that the user goal was to find the products they may be interested at Matches easily. On the business side, our success metrics were conversion rate uplift driven by supporting metrics like bounce rate, time spent on a page and page views per session.So I workshopped with developers and key stakeholders and prioritized solutions using a RICE framework where each solution has been scored based on user reach, potential impact on conversions, confidence in success, and the effort required, allowing us to prioritize those that will have the greatest impact.

As you can see, campaign page builder, targeting functionality and recommendation modules score the highest because they have relatively low efforts and high values These were our phase 1 initiatives that allowed us to focus on high impact campaigns first and then put more resources into page targeting and existing recommendation integration with an aim to replace the legacy homepage CMS with page builder eventually.

For phase 2, we plan to introduce automated copy suggestions and AI-powered content personalisation due to the significant relevance to current AI trend in technology.

Then, we would consider the smart scheduling and predictive editor analytics dashboard as means to better align editors’ curation with changing customer preferences.

Finally, other content types like videos, social media integration and gamifications have great potential. But they’ve been depriorisited due to lower user reaches.

To test whether creating a campaign page builder could drive higher campaign conversions as fast as we could, my plan was to develop an MVP in 1 month to catch up with the upcoming christmas gifting campaign in the UK.Here were the 3 key steps I’ve done:Step 1 was where I chatted with the editors and learned the biggest pain was the lengthy process to get dev resource. communicate requirements, and tweak the campaign page, which usually take around 4 weeks to deliver and only to be dusted after the campaign ended.

Step 2 was where I finalise the key features, scope and timeline with stakeholders, including a mobile-friendly UI, drag-and-drop interface for fast page creation, modular components for flexible layouts and google analytics integration to measure content performance.Step 3 was where I collaborated with designers, editors, developers and developed the agreed MVP in agile sprints. On the right of the slide, you could see the wireframes of the landing page and the CMS, which were used to gather stakeholders’ feedback and form alignments before implementation.

After building the MVP, we ran an A/B test comparing the new gift landing page to the old setup of a custom listing page.

The test ran for 2 weeks in Oct. And we split 40k campaign page visitors evenly between the control group, where users see a ‘gift for her’ listing page and the variation group, where users see a page covering multiple gift listings and relevant articles.

The results showed a 15.3% increase in conversion rates driven by easier navigation and more convincing gifting suggestions. Supporting metrics such as bounce rates dropped by 6%, and page views per session grew by 52%, suggesting stronger interests and user engagements.

These results validated the value of a campaign page builder for both users and the business and give us more confidence to roll it out to the whole UK market and other regional campaigns globally.

To ensure a smooth release & post-live supports in the UK market, I shared a detailed go-live plan and issue report process with editorial, marketing, engineering and analytics teams to ensure everyone in the working group is clear on who is doing what at what time and use slack as the main communication channels to give quick update and answer any questions.

Roughly speaking, at pre-launch period, I worked closely with editorial team to complete UAT and organised training sessions to onboard editors with the new CMS tool. I also involved marketing team to trigger targeted comms on socials and emails.

On the day of launch, I oversaw the go-live actions across the cross-functional teams, communicated update with stakeholders via slack channel, and monitored system performance.

After launch, I continued to monitor KPIs and customer feedback, addressed any bugs reported by editors quickly in slack, and prioritised improvements in the backlog for next iteration.

After rolling out the new gift guide in the UK for 6 weeks, we saw even more impactful results than a/b testing period thanks to the marketing comms and peak christmas season:

Compared to the christmas gifting campaign in 2022, conversion rates increased by 30% and order demand grew by 34%, adding almost 200,000 pounds in additional revenue.

User engagement has improved significantly with page views across gift listings increased by 177% and bounce rate reduced by 32%.We also received positive Trustpilot feedback where users were impressed by the amazing curation and beautiful product as a christmas gift.

This demonstrated how the CMS enhancements enable more targeted and inspiring content curation that entice customer to buy and in turn, drive conversions and revenue growth.

Reflecting on this project, here are some key takeaways I’ve learned: For me, the most challenging part of this project was to balance short-term needs with long-term scalability in a tight timeline. Without having a conversation with editors, it’s easy to go with the simplest tech solution by creating a hard-coded page for one time use because you want to be ‘agile’.

This project taught me the importance of prioritising high impact areas for maximum ROI, put yourself in both customers and editors’ shoes when the solutions involve operational works, and adopt an iterative approach to deliver quickly and refine over time.

This balanced approach helps my team achieve conversion growth as well as created a scalable, measurable CMS that can be reused for future campaigns and other landing pages across the site, allowing brand team to react to market trends and publish content landing pages in just a few hours going forward.