After building Dream.Me, we stopped treating scope creep like a surprise. It’s not a bug, it’s a byproduct of unclear priorities.

The problem wasn’t the work itself, it was the drift. We built a 40-screen onboarding flow because it looked smart on paper. We rewrote our data layer mid-cycle because CloudKit hit a hard wall. We reworked the App Store pitch five times before we even knew what users truly cared about.

Now, every Zephyr Studios app starts with one question: what’s the one behavior we’re here to support?

Brush Together, our toothbrush tracker, only exists to help households track brushing without micromanaging it. PromptScrub exists to strip malicious characters from GPT prompts so users can hit send with peace of mind. Held exists to create space for real emotional processing. Each one has a single point of clarity.

That’s how we scope now. Features can be cool, but if they don’t serve the core MVP behavior, they go in the backlog. Not into the build.

We still use timelines, but now we scope to the timeline instead of scoping first and adjusting the schedule later. We also build for testing in phases, not just release. If we can’t ship a version that brings real signal in two weeks, it’s too big.

Every project will drift if you let it. What matters is how you constrain it from the start.

— The Zephyr Studios Team