Eight rules I hold this app to
2026-08-23
This app exists to lower a tired parent's stress, not to become one more thing they get graded on. That reads like a slogan right up to the moment you have to decide whether a screen shows a countdown — and then it has to be a rule. These are the eight I hold. They extend the founding rule — a gauge, never a deadline — into something a code review can fail against, because good intentions in copy never hold a line for long.
The eight
1. A gauge, never a deadline. Every time on screen is the center of a gradient, not an appointment. A missed window gets calm amber and recoverable language, never alarm red.
2. Normalize before you advise. When behaviour is developmentally expected — no circadian rhythm in the first weeks, nap-count flicker during a transition, a scattered day — the app says so before it suggests anything. A parent who reads this is biology, on schedule stops hunting for what they did wrong. Advice can wait. Absolution cannot.
3. Age-honest voice. The app knows the child's age and never implies structure the child cannot have yet. Under two months it proposes no steering at all — there is no clock to steer by. While the clock is switching on, it stays loose about clock times. Ages count from the due date.
4. Describe the child, never grade the parent. Copy describes the child's state — building, not sleepy yet — never the parent's performance. Drift is habits run 23 minutes late, not you are late. The prediction league scores my algorithms against each other; the family is never a contestant. Banned outright: "you missed", "behind schedule", countdowns, any streak a parent can break.
5. The parent is the sensor, not the executor. The child's cues outrank every number on the screen, and the copy says so at the moments where it matters. The app proposes, the child disposes, the parent referees.
6. Uncertainty is weather, not failure. Thin data reads as too early to tell — that's expected, never as an empty red state. Bands show at their honest width. A suggestion to shift timing waits for a three-day pattern, so one odd day never becomes advice.
7. Small doses, glanceable, quiet. This is a five-second app. One reassuring line beats a paragraph, and the paragraph lives in the guide for whoever wants it at 2 a.m. No notifications, no urgency mechanics, nothing engineered to pull you back.
8. Every claim is traceable. Reassurance only works if it's true. Every developmental claim maps to a published source or to a result I measured myself. No invented certainty. No "studies show" without a study.
The one test that decides it
Before any sentence ships: if a parent reads this at 3 a.m., after the fourth waking, does it lower their shoulders or raise them?
If it raises them and the information is genuinely necessary — safety, a real decision — keep it short, concrete and free of judgment. If it raises them and the information is not necessary, cut it. Not soften it. Cut it.
Why this needs to be written down
The industry standard is the opposite of all eight, and that isn't an accident. Streaks, badges, red states and a push notification at the right hour are extraordinarily good at retention. They work by making you slightly anxious on a schedule, then selling you the relief. Every one of those mechanics is a solved problem; I know how to build them; they would measurably improve every number a product person would ask me for.
An app for people running on four hours of sleep has no business using a single one of them. Writing the rules down is how I make sure that stays true on the evening I am tired and a retention metric looks flat.