Analytics showed a pattern common to large content hubs: high arrival, low engagement, significant early abandonment. Scroll depth data showed the majority of users were not getting past the hero.

Search query analysis consistently showed that people arrived with intent expressed in one of three frames: how to attend (event format), what they care about (topic or product), who they are (audience role). That hierarchy became the conceptual foundation for much of the catalog development.

I conducted an audit of all the existing pages, mapping where content element currency existed, their positions, and, engagement. I worked with the client team to understand the current content hierarchy, and found that page designs often reflected business priorities over user learning agendas. A more surprising takeaway, though, was that many navigation and filter options were ordered according to engagement metrics without regard to search data, broader context, or usability principles. For example, listing countries in order of engagement made it exceedingly difficult for anyone outside the dominant markets to find a local event. They would have to read through a list of 30+ countries to find the correct filter. The audit also revealed that the taxonomy and naming often reflected internal organizational terms that held little meaning for users. This meant building in organizational principles and naming frameworks into my recommendations. 

My work included hundreds of changes to the events hub, but I’ll focus here on content strategy for the primary search pages (eg, home, catalog) versus secondary pages (eg, campaigns, event detail) or copywriting.

Homepage

For most pages and especially the homepage, I needed to explore different levels of curation versus automation in order to meet the business requirements of promoting priority events but also ensuring continuous display of upcoming events with minimal effort. At the same time, the page needed to feel personalized, even for a new visitor who may not yet have an account. 

Given the engagement metrics around search, there was strong client desire to prominently feature it on the homepage. While I didn’t disagree, I felt the data told a more complicated story: search worked well for users who arrived with a specific query in mind; it consistently overwhelmed users who were in a browsing mode, looking to see what was available within broader parameters.

To solve for the latter use case, I had the idea for a “light personalization” frame that offered the ability to select from multiple, popular dimensions. Submitting their choices led users to a pre-filtered catalog page where they could further refine their results. I wrote the module as a conversational prompt, mimicking the AI chat I originally recommended, but could not integrate within our timeframe. This solution worked because it educated users on search parameters in a stepped approach before arriving at the catalog page with its lists of filters. It also allowed for personalization without relying on account data. 

Below that frame, the rest of the page was structured the page in tiers according to audience attraction – global and local. Flagship events appeared first, not only because they’re the most popular, but because they carry brand authority and scale to immediately signal industry dominance. Plus, recognizable event names would reduce uncertainty for newer users. 

Below flagship, events were curated by type and category, followed by an automated feed of upcoming local events. These were both design to demonstrate breadth of events but also provide faster pathways in. Toward the bottom were modules for complementary content, including additional learning and community opportunities outside of events to maintain engagement with users who may not have found an event to meet their needs.

Event catalog

Separating the event catalog from the event cards, my focus for the former was filter clarity and hierarchy. For clarity, I evaluated and then rewrote filter labels against the vocabulary from search query data to understand where the internal language was getting lost on users. The difference between a filter someone understands and one skipped was often a single word. Once I determined which terms were most meaningful, I revised the labels for plain vs. internal language and then developed a framework for future naming.

I recommended that filters be broad enough to produce at least ten results. For example, filtering for all the individual countries events were hosted in meant some only returned one or two results, contradicting the desired impression of a robust catalog. In that case, I recommended filtering by regions (eg, EU), especially in markets where travel between countries may be easier. The client, however, originally chose not to adopt that approach, but after successful user testing, later implemented.

I also made changes to ensure filters clearly narrowed down results. For example, events were often tagged with multiple or even all industries because organizers considered the event relevant to all. Rather than list 15 industry verticals – each returning similar results – I collapsed the dimension into a single “Industry focused” filter. It not only helped reduced the overall number of filter groupings, but elevated events with a specific industry focus.

The order of the filters was another major issue — one that generated the most internal pushback. The existing filter lists were ordered by engagement: the most-clicked options appeared first. On the surface, that sounds logical. In practice, it meant that a list of 30+ countries was ordered by market popularity, making it genuinely difficult for anyone else to find a local event. A competitive analysis I conducted showed alphabetical ordering was standard practice across comparable platforms. As an established mental model, that made sense. Still, that evidence helped make the case.

Event cards

The original event cards were attempting to do too much with little visual differentiation. Each card displayed the event title, a full date and time with timezone, format, language, event type, a truncated description paragraph with a “Show more” link, and a blue “Registration and details” CTA button. The result was a wall of near-identical blocks of text with no clear hierarchy telling a user where to look first.

A key point of tension was between completeness and scannability. The card was trying to give users all the information for making a registration decision, but, in doing so, prevented them from making any decisions. Users scanning a full page of results need to be able to eliminate certain cards quickly.

My solution was to strip the card back to only what was essential for the scanning decision, and let the filters the user had already selected carry the rest. If someone had filtered for free digital events in English, there was no need to repeat those details on every card they saw. Removing that redundancy gave the remaining content space to breathe and created room for actual visual hierarchy.

The other main tension was balancing information consistency with event-specific flexibility. A rigid card template would create visual and structural coherence but would flatten meaningful differences between a free 45-minute webinar and a multi-day, in-person conference. The solution was to define a fixed information hierarchy so every card followed the same sequence of fields, but then left some flexibility in fields used and language within them. At the same time, we stripped away some of the visuals (eg, icons, logos) that were creating unnecessary cognitive load.

Outcomes

While results are still being gathered and have not yet been shared with the agency team, I can speak to early success against an internal metric. The content system successfully accommodates edge cases and business constraints, and has scaled across teams and markets without requiring additional resources or work.

I will add that this project reinforced that the most critical content decisions on a platform of this scale are often the most overlooked. I also learned something about the value of constraint. Having no user interviews or research forced me to be more thoughtful and rigorous about what behavioral evidence could be drawn from the web analytics and search data. It’s easy to over-interpret user intent from a few stated preferences. Search queries and click distribution gave us insight into user patterns at scale.