As long as the injectors are clean and functional, data collection takes about one full day in the lab for me to get what I consider the basics of characterization. This means a complete flow curve at 13.5v for a set of eight injectors at a single pressure and a shorter series of tests at varying voltages in order to find the trendline of offset vs voltage and populate that for the same single pressure.
Adding multiple pressure breakpoints simply adds a complete repeat of the procedure, as does running more than 8 pieces. I typically test 4 injectors at a time to make sure that their flow doesn't outstrip pump capacity or rail distribution, so testing 8pcs means two sets of four. If we are trying to get data for a part number rather than just a single set of flow matched injectors, then we really should test more than 8pcs. There can be build tolerances of +/-6% for production injectors, more for aftermarket. Finding the real "average" behavior requires that we test more samples in order to make sure we get a statistically significant sample.
The going rate for engineering consultation is north of $100/hr (just like a good lawyer, accountant, doctor, etc...) so spending a full day or three on this quickly eclipses the typical cost most associate with "engine tuning." I'm not recommending that everyone has every single injector measured, as this would be prohibitively expensive. But it is certainly nice to be able to have the confidence to say that you really know exactly how much fuel is being delivered on each shot. To some race teams or injector manufacturers, this cost is literally a drop in the bucket compared to the rest of the operation. So for those guys, we try to make sure we have as much GOOD data as possible.