A launch review and a long-term test of the same car often reach different conclusions. The reason is that most of what makes a car good or bad to own only becomes visible with time.
A first drive samples a narrow slice
Launch events use selected routes in favourable weather, with cars prepared to a high standard and often specified in their most flattering trim.
The tester has a few hours, which is enough to assess ride, handling, powertrain response and ergonomics, and very little else.
Nothing in that format can reveal how the car behaves after a year, because no mechanism of wear or degradation has had the opportunity to operate.
Irritations need repetition to register
A control that takes one extra press is unnoticeable on a single drive and becomes genuinely annoying when performed twice a day for six months.
The same applies to a boot lip that catches luggage, a door bin too shallow for a bottle, or a driving position that is comfortable for an hour and not for four.
These are the details owners cite most often when asked what they dislike, and they are almost invisible to a short test.
Software behaviour emerges over time
Infotainment systems and driver assistance features are tested for a long period because their failures are intermittent rather than constant.
A navigation system that occasionally loses its position, or a lane assist that misreads a particular type of road marking, will not necessarily misbehave during a launch drive.
Over-the-air updates add another dimension, since the car a reviewer assessed at launch may behave differently several months later without any hardware change.
Running costs only appear with mileage
Tyre and brake wear rates, real consumption across a full year of weather, and the cost and frequency of scheduled servicing all require distance to establish.
Depreciation, which is usually the largest single cost of ownership, cannot be observed at all in a short test and is only estimated at launch.
A long-term test that records these figures gives a buyer something closer to the true cost than any list price comparison can.
What a long-term test still cannot tell you
One car is a sample of one. A fault in a long-term vehicle may be a pattern across the model or may be unique to that example.
Six months rarely reaches the age at which the more expensive failures occur, so genuine long-term reliability remains a matter of owner data rather than press fleets.
Press cars also receive prompt dealer attention, which means the ownership experience around a fault is not represented at all.