IANAP (I am not a physicist) but my guess is that it's an acceleration thing.
As the engine burns it uses fuel. Therefore mass goes down and acceleration increases during the course of the burn. This means that each extra second of burn changes the velocity of the spacecraft by a steadily greater amount. This in turn increases the effect of an error if the engine burn is too long or if the engine's thrust differs from expected by a serious amount.
The first burn is as short as is needed to place the spacecraft into a stable orbit, even if it's elliptical. Then, once they know exactly what the shape of the orbit is, they do a second burn to adjust it into what they actually want.
However, happy for the physicists to explain where I'm wrong.