Notes Aug 20, 2026 4 min read
Release day, hour by hour
What is worth looking at in the first two hours, the first day, the first week and day 14 of a release, and the three numbers to write down before you press the button.
Casey Morgan Co-founder, CEO
Release day used to mean refreshing three tabs until midnight. This is what we look at instead, and when, written from our own releases and from watching a few thousand of yours.
Before you press release
Write down three numbers from the last seven days: average daily downloads, average rating, and the rank of your two most important keywords. It takes a minute, and without it you will spend the evening trying to remember what normal looked like. Ascent pins these to the release for you when it sees a new version go live, but the habit is worth having either way.
Release in the morning, early in the week. A release at 4:00 PM on a Friday puts your first one-star review at 9:00 PM on a Friday.
Hour 0 to 2: nothing
The update takes time to reach devices. Automatic updates roll out over days, not minutes, so in the first two hours almost nobody has the new version. There is nothing to see and no number worth reading. Go for a walk.
Hour 2 to 24: reviews, and only reviews
Reviews are the first signal with any meaning, because the people who update by hand on day one are your most engaged users and they will tell you quickly if something broke.
What you are watching for is not the rating but repetition. One review that mentions sync is a person. Three in a day that mention sync is a bug. Jordan Reyes at Loafware saw exactly that pattern the morning after version 2.4 shipped: four reviews, all on iPad, three with the word “sync.” A fix was in review by lunch, and live 26 hours after the first complaint.
This is the one day we would tell anyone to turn on every alert they have.
Day 2: downloads against the same weekday
Yesterday’s downloads arrive in this morning’s email. Compare them with the same weekday last week, not with the day before. A release on a Tuesday will always beat the Monday before it, for reasons that have nothing to do with the release.
Unless you were featured or ran a promotion, do not expect much. An update by itself rarely moves downloads on day one. If downloads fell sharply, look at your product page before you look at your code: a screenshot that failed to upload is more common than you would think.
Day 3 to 5: ranks
If you changed your title, subtitle or keywords, this is when the store catches up. Ranks before day three are noise. A step that appears on day three or four and is still there on day six is real.
Day 7: trials and the paywall
A week in, the first trials that started on the new version are ending. If you changed onboarding or the paywall, this is the first morning the conversion figure can say anything about it, and it will be a small cohort. Read it as a direction, not a result.
Day 14: write it down
Two weeks after release, set the three numbers you wrote down at the start beside today’s. Then write one sentence about what the release did. “2.4 fixed the widget and the rating went from 4.3 back to 4.8.” “3.1 moved the subtitle and search downloads rose 44%.”
Most releases will earn a duller sentence than those. Keep the dull ones too. In a year you will have twenty of them, and together they are the only honest record of what moved your app, written by the one person who was there.
