I think first thing you should do is compare your compression algorithm with other algorithms (zip, rar, 7zip, etc). Good compression of few files does not means that compression algorithm is good. There are standard benchmark datasets available for this.
This[
^] is one example. Once you have established that compression algorithm is indeed good then you have few options:
If you don't mind sharing the algorithm with others then you can publish in a good peer review journal. If you don't want to share the algorithm then you might try to get it patented.
In either case you can sell the software.
-Saurabh