Local Coordinate System
What proportion of your surveys are on a local coordinate system?
If you mean local arbitrary system not tied to WGS84: less than 2%
If you mean local low distortion projection that is geodetically referenced: greater than 98%
If you mean both: 100%
Off the top of my head I can't think of the last time I set up a project on random / local coordinate system. I have been brought in to work on projects that appear to be local (without proper meta data it is hard to say). Just this morning, the practice of tying into an established system paid off big time. At the end of October I did a little design survey for about 600 feet of track realignment in a mill. The engineer was supposed to design it the next day so I could turn around and lay it out. I set two control points and tied them to the ORGN. The plans were not completed until yesterday. I went today to find that one of my two points had been the victim of a utility relocation crew at the end of last week. Being able to set a new point saved my bacon.
By local coordinate system I mean 5000/5000 (or similar) with bearings based on some classical method such as an old survey line, compass, sunshot, or just assumed. Not geodetically referenced.
For the record I would wager that well over 90% of surveys done in the Portland/Vancouver area are on local systems.
Maybe 5% of the jobs I do are done on local systems. They tend to be small and simple jobs done using conventional equipment where the scope is very limited, or they're based on legacy work that was done on such a system where it's just easier and less hassle to stay within the confines of what had been done previously. These days 95% of my work is done in State Plane and tied to the CORS network. I'm just finishing up a project utilizing ties from four previous surveys done over the last decade. Likely saved me several days of work and my client never argues when I save them money by recycling data they've already paid for once.