Most launch reviews are written from a sample that arrived a week or two before the embargo lifted. That window is set by the manufacturer, and it shapes what a review can honestly claim.
The embargo window sets the schedule
A manufacturer ships review units on one date and allows publication on another. Everything a reviewer can say has to be established between those two fixed points.
That period is rarely generous. Hardware often arrives days before launch, and the reviewer is running tests, shooting photographs and writing the piece at the same time.
The verdict is therefore formed under time pressure rather than through extended use. Reviewers usually say so somewhere in the text, but the caveat travels far less widely than the score does.
Some properties need months, not days
Battery capacity loss, hinge wear, thermal paste ageing and the slow accumulation of software cruft are not observable in a week. They emerge only over sustained ownership.
A reviewer can measure a battery's capacity on day one and can stress a device until it throttles. Neither test predicts how the same part behaves after eighteen months.
This is why the most useful sentences in a launch review are the ones about design, controls and measurable performance, rather than the ones about durability.
Firmware moves under the reviewer
Review units frequently run pre-release software that changes before retail units ship. A day-one patch can alter camera processing, power management or thermal limits after the review is published.
That makes some findings genuinely perishable. A stutter or a battery result recorded on pre-release firmware may not describe the product a buyer receives.
Good reviewers note the build they tested. Fewer readers check it, and search results keep the original text in circulation long after the software has moved on.
The sample size is one
A reviewer receives a single unit, and manufacturing variation across a production run is real. Panel uniformity, coil whine and fan noise differ between individual devices.
One unit cannot separate a fault in that specific device from a fault in the design. A reviewer who reports a defect risks describing bad luck as a pattern.
Reviewers who handle many units, or who compare notes with others testing the same product, are better placed to tell the two apart.
What the format is actually good at
Launch reviews are strong on things that are visible immediately: build quality, screen behaviour, control layout, port selection, software design and how the device compares with its predecessor.
They are weak on everything that accrues. Longevity, repair experience, support duration and how the manufacturer behaves after the news cycle passes all sit outside the window.
Reading a launch review as a description of the first week, rather than a prediction of the next five years, matches the format to what it can support.