On a basically stock car, the mail order tunes will be within probably 90% or better of a dynotune in most cases - and I'm talking about SCT tunes. To get the absolute most out of a car a dyotune is the way to go. But a mail order SCT tune on a basically stock car isn't that far behind.
On blower cars and heavily modded cars, we will NOT do a mail order tune unless the owner will supply us with a lot of data after we send out the first 'round' of tunes for the vehicle. This means that the owner either needs a wideband or access to one and also needs to datalog with the XCal 2 and send us the results. In fact, you can read about how we do it on our website in the tech section (see sig for link) - we send out a list of parameters we need to have logged - then after we get the data we email the customer a new file(s). The other thing that's really important when doing any tune is that adaptive learning has to be shut off in the initial tune until things are dialed in or you end up chasing your tail. We turn it off in initial tunes, then once it's all dialed in, we turn it back on and email the final file. Doing things this way makes it a LOT easier to dial in a tune via mail order. Still not as good as a dynotune, but when someone isn't close to a good tuner or wants to spend a little less, it works much better than a 'blind' mail order tune. 'Canned' tunes usually have spotty results because you can dynotune 10 blower/turbo cars of the same year with similar mods and usually none of them will be identical as far as the tune goes - small variations in the combo sometimes make big differences.
Don