12 June 2026 · Niran Chaiyasit
The first week is a poor ruler for feature adoption
The week a feature ships mixes curiosity, staff checking, and the banner. It is a poor ruler for whether anyone came back.
A feature adoption curve that starts on the morning of launch will almost always look dramatic, and it will almost always lie a little. People tap because a banner told them to. Colleagues tap because they are checking the release. Support taps while sitting with a customer. None of that is false use. It is also not the use you will still have in week six.
When we open a Feature Adoption Brief at the studio, we draw the launch week, and then we set it aside as a marked stretch rather than as the ruler. The stretch we actually read is usually weeks two through eight, or through sixteen if the feature is seasonal. The denominator is not every account in the product. It is the people who could have seen the feature: the screen was in their build, the control was not hidden behind a flag they did not have, and they were not a staff account you can name.
That last cut matters more in Bangkok product teams than the slides suggest. A payments desk we worked with had a transfer shortcut that looked busy in week one and empty in week three. Once staff accounts and the support floor were lifted out, the remaining curve was smaller and almost flat, with a Friday pulse from people who already sent money every week. The feature was not a failure and it was not a triumph. It was a habit for a group the team had not named.
If you are preparing a table before you write to us, mark the launch week in the dates and list the internal domains or staff identifiers you know about. You do not need a perfect list. You need an honest one. We would rather drop a doubtful row than ink it into the denominator and spend the closing conversation arguing with our own plate.
We also ask what sat on top of the feature in that first week. A promotion, a push note, a contest, a visit from the founder on a community thread. Those are not shameful. They change what the curve means. A curve that only rises while the banner is up is a curve about the banner.
The plate that comes out of this is simple on purpose. One line for people who could see the feature, one line for people who tried it, the launch week held in a different ink so the eye does not treat it as the whole story. The reading then says, in sentences, who returned after the noise. If the sample after week one is too thin to support a return rate, we say that and show counts. We do not smooth a thin week into a confident slope.
If your review is next Monday and all you have is the launch week, bring it to a workshop morning rather than asking for a brief. A brief that pretends week one is a settled habit will only give you a handsome plate of the wrong thing.