I see two great things about this tool, and not exactly for what it was designed:

1) Fix bandwidth problems caused by API designer laziness.

2) Detect API designer laziness and programmer laziness.

The API designer laziness merits description. Essentially, if you as an API host server are sending back tons of redundant data, you are doing a disservice to your user by not passing references to this data in the first place. The results in your API should either include an entities section or links to retrieve more data. (Note the scale of repeated data I am talking about here are subtrees, not just key/value)

The programmer laziness is when an he just sends back JSON.stringify(someVeryLargeObject), instead of making sure only a minimal object is sent back.

So we can detect this laziness by running this tool and looking for 4x compression gains!