The idea of transcoding h.264 files (lossy by definition and nature) to yet another lossy format, even h.265, is not a good idea and never recommended unless you really don't have a choice and I can't imagine a situation where that kind of transcode is required aside from you or whoever else wanting smaller files. The preferred and recommended solution would be to get the original source material and then do a single generation encode from source to h.265 aka HEVC.
h.265 HEVC (on paper) offers similar or slightly better visual quality at roughly half the bitrate of h.264 in most situations which results in files that can be up to about 50% smaller (based on just the video stream size, the audio stream content is rarely considered in such calculations). So if you had a 1080p clip that was 1 hour long and came in at about 2GB in size (video and audio streams in the container) in h.264 format as encoded directly from source, you could theoretically end up with a ~1GB file (video and audio streams in the container) in h.265 format also encoded directly from source.
In a transcode from h.264 to h.265 the result might be even smaller because of the lossy source material so again, using original source material for the single generation encode (meaning Source > h.265 one time) is going to result in the best quality possible.
Also take into consideration that "defaults" for h.265 encoding don't always end up with the best results so, some hand-tuning of the possible encoding options is going to be highly relevant as well. To get the very best quality with respect to smaller file sizes as the end result you're going to have to do some test encodes to learn how h.265 works (and how it differs from h.264 encoding in those respects).
h.265 is now coming on pretty strong and with a little bit of GPU assist in the latest generation Intel processors means it won't take as long to encode as it has in the past, obviously, but those encoding options will be the final arbiter in terms of the quality of your resulting encodes so, it's a learning process all over again just like it was for using h.264.