If you want the quick and dirty answer: because the coordinates in one SRID are not the same as the coordinates in the other SRID. Basically what a coordinate system is, is an approximation of the shape of the earth. Because there are a lot of different goals and purposes to represent the earth, there are a lot of representations out there, and depending on what you need and where you are working these coordinate systems can be useful or are completely rubbish (for example: if I would use the coordinate system of the Netherlands to display map data in the US, it will get wildly distorted and is basically useless). So essentially what you are doing is moving your data from one representation to another, but without changing the coordinates. That would be similar to projecting your map on a cube, then replacing the cube with a piramid (ok, that's a bit of an exxageration, but it works as an example I hope ). The long answer gets a bit theoretical, so I'll just post a few links.
To be fair, moving from one geodetic coordinate system to another will probably not have much of an impact, especially not if the scale you are working on is fairly small. If you change from a geodetic to a projected coordinate system, your data effectively becomes rubbish, so that is definitely a point where you need to transform. I think a reasonably decent starting point for more information on this subject is actually the Oracle Documentation itself: http://docs.oracle.com/cd/E11882_01/appdev.112/e11830/sdo_cs_concepts.htm#SPATL050
Wikipedia also has a decent explanation:Coordinate system - Wikipedia, the free encyclopedia
I would suggest you read up on this, to get a bit more background information, and if you have any questions left: well, we're mostly around here at one time or another.
Hope this clears things up a bit,