I shoot primarily E6 film, and haven't sent any to a lab in years. Here's my take on it:
If you can control temp and times, temp with a Jobo or equivalent and times with a cheap multi-timer, then it's a fairly easy thing processing your own E6. Chemicals can go down the drain (I use the Kodak Single-use E6 kits). Setup of the Jobo and chem mixing takes about an hour, which includes temperature stabilization times (I mix the chems in warm water to hasten the temp stabilization), then each batch requires about 50 minutes of effort (from loading the tank to hanging the fiml to dry).
Each kit costs about $45, and I find that I can process 18 sheets at a time in the 2840 tank using three 2509n reels. Each batch uses one liter of chems (the Jobo CPA/CPP had 1 L storage bottles). I can process two batches with each liter fo chems, so the 5L E6 kit will process 180 sheets. That works out to $0.25 a sheet. And if I carefully extend the first developer time, I can easily process a third tank with each liter of chems, lowering the cost-per-sheet to $.17. The $800 I spent for a used Jobo CPA and the extra $200 in additional drums and reels paid for itself after 500 sheets (about 1.5 years worth of shooting, presuming I have sheets processed locally for $2.25 each). Pretty good economics for me.
In two years I have ruined a total of 12 sheets (all on the same day, probably due to some fixer getting into one of the developers, but I never tracked it down).
Using the R-3 chems is also working out very economically for me. I'll be able to process over 2000 16x20 prints with the pro 12.5/25/12.5 gallon R3 kit, which costs $350. Fuji Type-35 paper costs either $1.50 per 16x20 paper-backed sheet or $5.00 for the polyester backing. You can compare this with what your local lab charges for a 16x20.
Now, my experience with home color porcessing is based on my research in enzyme kinetics, where time and temperature control is paramount, so I found that getting consitent development was no problem at all. If you've never dealt with tightly-controlled processes, or have trouble following directions exactly, you may find that consistent work is a little harder to achieve.
Consult if you wish two articles I've written on the subject: