We Got the Denomination Filter Wrong. Here's What We Learned.

The assumption we shipped with

When we built the denomination filter for build 1, we thought we understood the space. We'd spoken to dozens of church staff and event organisers. We knew that Catholics want to find Catholic events. Pentecostals want Pentecostal events. Anglicans want Anglican events. So we built what felt logical: a simple dropdown. Select your denomination. See your events. Done.

What we'd missed, sitting in the design studio, was how messy the real world is. A joint event hosted by a Baptist church and an independent evangelical group would get tagged as Baptist in our system because that's where it was listed from. But an attendee from the independent church wouldn't find it, because the filter was too narrow. Meanwhile, someone looking broadly for non-denominational events would miss everything Anglican, which isn't always what they wanted.

The Leeds administrator's message was the first crack in that logic. It wasn't rude. It wasn't even a complaint, really. Just a gentle 'this doesn't work for how we actually look for things.' We started getting similar messages over the next fortnight.

Why we'd got it backwards

The real issue was that we'd flipped the intention. We thought the denomination filter was about accuracy. Make sure people only see 'their' events. Keep the directory pure. What we actually needed was discovery. Help people find what matters to them, even if the boundaries weren't where we'd drawn them.

A visitor to Gathrd might be Church of England but open to attending a worship conference run by a charismatic network. They might be Catholic but interested in a joint prayer vigil with evangelical churches. Or they might simply not know what denomination they wanted until they browsed and saw what was nearby.

Our filter was gatekeeping instead of enabling. We'd built Gathrd on the promise that every faith event, one place. That means a faith-only directory without nightclubs mixed in. It doesn't mean we should make it harder for people to stumble across something beautiful outside their usual circle.

The fix, and what it cost us

We rebuilt the filter for build 2. Instead of a strict denomination selector, it became layered. You could filter by primary denomination if you wanted to, but we also added denomination tags for co-hosted events and a 'view similar' suggestion system. If you filtered for Baptist events but a joint Baptist-Pentecostal prayer meeting was nearby, we'd still surface it, but with clarity about who was actually hosting.

The rebuild took two weeks. We had to change how we stored event data, rewrite the search indexing, and test across iOS, Android, and web. It wasn't a visual tweak. It was structural. And it meant pushing back a feature launch we'd promised to a group of ministry leaders in Birmingham.

That conversation was harder than the code. We told them the truth: we got something fundamental wrong, and we're fixing it properly rather than quickly. They appreciated the honesty. They also said, 'Can you also let us tag our events for age group?' Which we then had to add.

What it taught us about building for real communities

Building for churches isn't like building for generic event-goers. Church communities are overlapping, interdependent, and often intentionally blurred at the edges. A choir might draw from three denominations. A prayer network might span four counties and five different church traditions. A youth conference isn't just for one tradition; it's for teenagers whose parents trust that particular event.

The denomination filter taught us that we have to think in terms of what people are actually looking for, not what we think they should be looking for. We're not trying to be a directory that says 'keep everyone sorted into boxes.' We're trying to say 'here's every faith event near you, and here's what you need to know about it so you can decide.'

That's a harder UX problem than dropdown filters. It's one we're still solving, honestly. But it's the right problem to solve if you want to serve church communities rather than just process bookings.

The bigger lesson about shipping

There's a pressure, especially early on, to ship everything as slick. Make it feel finished. Make it feel like you know what you're doing. What we learned is that getting real feedback from real church staff, even feedback that stings a bit, is worth more than any amount of internal testing. The Leeds administrator who messaged us wasn't trying to be difficult. She was doing her job. And her job helped us do ours better.

We've kept the Leeds administrator's original message in our Slack. Not as a complaint, but as a reminder. Whenever someone suggests a change to how events are filtered, discovered, or tagged, we ask: does this serve the person doing the actual work, or does it just serve our mental model of how the directory should work?