Outcomes & methodology
We’ll publish numbers when they mean something.
Daybreak is new. There is no outcome data yet, and we’re not going to invent any. This page is the standard we’ll be held to once there is.
This page, today
- Numbers published0
- Metrics committed to4
- Update cadence≤ quarterly
- Last updated—
What we’ll publish
Recovery metrics, not engagement metrics.
A short list of things we think matter — and that we’ll be willing to defend, with methods, when we have enough data to publish them honestly.
- 01
Week-1 completion rate
How many users who start a habit-change plan complete the first seven days. Not a streak — just a count of who actually used the product, with the sample size shown.
- 02
Median time to first breakthrough
Self-reported, opt-in, in users’ own words. We publish the median and the spread, not a single best case.
- 03
Mood trend over four weeks
Anonymized, opted-in, with standard error. Direction, not drama. Reported as an aggregate, never tied to an individual.
- 04
Anonymized quotes
Only when consent is explicit, and only with the level of identifying detail the user chose — typically first initial and rough age.
When we’ll publish
When the numbers mean something.
We’re not committing to a specific user threshold here. The right sample size depends on the metric — some questions need a few hundred users, some need several thousand. We’ll always publish the sample size, the time window, and any caveats alongside the number itself.
We’ll update this page at most quarterly. Less often if nothing meaningful changed. The “Last updated” stamp at the top will become a real date the first time we publish.
What we won’t do
A short list of things we’ve ruled out.
- DAU, time-on-app, or session length — those are engagement metrics, not behavior-change metrics.
- Cherry-picked numbers from a narrow window or a flattering subset.
- A single user’s story used as a stand-in for population outcomes.
- Research-quality claims without an external partner who can run the study properly.
How the data is collected
Aggregate, opt-in, never tied to a journal.
The numbers we’ll publish come only from data the user has explicitly opted in to share for research. Journal entries and Dawn conversations are never included in any aggregate, opted-in or otherwise.
When you’ll see numbers here
Until then, this page stays a commitment.
Be notified when the first dataset is published — and when the methodology is open for review.