Files that include a lot of information, such as bitmap graphics, audio or video files, cannot be compressed much with lossless compression because there is so little data to index.
Lossy compression works differently, it data that is not needed, either because a drop in is acceptable or the difference cannot be detected by the human or ear.
With lossy compression you do not get the data back when it the compressed data is decompressed. This sort of compression would therefore NOT be suitable for anything that needed to be reproduced such as software, databases or any file that contained .

Lossy compression on bitmap images:

This works best when there are lots of coloured pixels together. In such cases the compression algorithm will be able to calculate the average colour of a large number of blocks of pixels and store all those blocks as one colour, greatly reducing the file size. JPG or JPEG image files are an example of lossy compression.

Lossy compression on audio files:

This involves removing data from the original file so there is a drop in quality. However, to minimise this the algorithms that carry out the compression put more emphasis on removing data that the human ear detect or is less sensitive to.
A typical CD music file can be compressed to less than 1/10 of the size by compressing it using the file format.
data can be compressed even more than music data and still be acceptable. Such voice compression is used extensively for conversations over the Internet to reduce the required.