What do you do when you need to create some data but you haven't built out the UI for that data yet? A seeder is a great approach, but is it always the right one?

In the latest episode of the No Compromises podcast, we dive into a real project where starting with the most complex feature made test data management painful. Instead of exploding the complexity of our seeders, we built a minimal UI to manage test data.

We also talk about some other unexpected benefits, and talk through the trade-offs and why detours like this should feel uncomfortable (and be tightly scoped).

• (00:00) - Starting deep exposes messy user permutations
• (02:45) - Seeder explosion vs. a minimal UI
• (03:45) - Reframing the “detour” after using it
• (05:30) - Why the mini-UI helped: faster iteration, fewer seed resets
• (07:45) - Dogfooding + tester debugging benefits
• (08:00) - Guardrails: detours should feel uneasy and stay tight
• (09:00) - Silly bit