Dr. Ryan Bell

I was analysing our checkout funnel in GA4 and saw a clear drop-off at the traveller details step. Progression there was significantly lower than earlier stages, which suggested a friction point.

When I looked into it, the issue was that we were asking users for quite detailed traveller information too early in the journey, before they'd fully committed to the purchase.

I made the decision to remove non-essential mandatory fields and simplify that step, while still ensuring we captured the required booking data later in the process.

There was a trade-off here - we were collecting less information upfront, which stakeholders were initially concerned about. But I focused on the commercial impact and the idea that progression was more valuable than early data capture.

After implementing the change, we saw a 20% improvement in progression through that step - from 8.9% to 10.8% - and a 43% increase in overall conversion from 5.1% to 7.4% over 30 days.

One of my primary KPIs was the volume of users visiting the product page and progressing through the funnel. I was looking to improve this, so I focused on the homepage search experience.

I had a redesigned version available - one that had an option allowing users to search without setting specific travel dates. I ran an A/B test when launching the feature, but initially it underperformed and the impact wasn't statistically significant - users weren't interacting with it enough.

Instead of dropping it, I looked deeper and realised the issue was visibility. The improved version was hidden behind a tab, so most users never saw it.

I adjusted the test to make the new version the default, and that led to a 26% increase in product views and a 20% improvement in progression into the funnel.

I identified a sustained drop in organic traffic to blog content over a 3-4 month period, significantly outside expected historical variance. When I broke it down, the decline was isolated to informational queries.

Given the timing, I attributed this primarily to AI Overviews and LLM-driven search behaviour - users were getting answers directly without needing to click through to read a whole article and find the answer.

At that point, the decision was whether to try to recover that traffic or adapt. Blog traffic had never been a strong direct conversion driver, so I chose not to double down on content production.

Instead, I reprioritised the roadmap toward making content more accessible and valuable within this new AI-search era. I implemented FAQ schema markup at scale, centralised review schema so it could surface in search results, and improved structured data consistency across the site.

There was some pushback due to competing priorities like a brand re-skin and ongoing platform work, but I prioritised this because it directly impacted visibility and commercial performance.

The outcome was that while traffic itself didn't recover - and I didn't expect it to - I improved visibility through rich search results and maintained overall conversion performance despite reduced top-of-funnel traffic.