Do you have any data examples you can share that you have collected with it ?
To reduce the amount of points you can use the subsample tool and select a more appropriate point spacing.
Cloud compare also can reduce the amount of figures in a coordinate (if you points are georeferenced to a datum, if not then this won't save much unless it's a very large area), for example if all of your points fall with say a 1000m X 1000m box and are referenced to to OSGB36 it could remove the first common digits from the Eastings and northings, if often prompts this when you load in a dataset but you can write out a shortened file if this is desirable.
Unfortunately not yet, I have literally taken it out for a 2 hour test run the evening before lock down started and so I have only produced the quick test which I have put on this thread.
The actual data output consists of 2 files, a survey file and a point cloud file. The survey file which is in a format for walls (its American!) contains the normal survey data (survey stations, distances, azimuth, inclination etc) this file layout could be altered for a European survey program if required. The second is just the point cloud data file. The survey file allows you to check the survey file and adjust any loop closures prior to adding the point cloud. The point cloud is then put into a point cloud program (I have used cloud compare because it is free!) this will then produce the 3D map. I haven?t yet found the best method for sharing these point clouds due to the file size (the example I have attached above is only a small section of passage and consists of 1.6million data points)