In another thread, because he didn't have a scale, someone posted that he was using a volume measurement -- in this case, three level tablespoons, or approx 44.36 mL -- for 15 g of coffee for espresso. Since he was trying to optimize the grind for pulling shots, I thought there might be too much variability in the actual weight of his doses, and suggested he ought to consider getting a scale with a resolution of .1 g, just to get that variable under control. But then I wondered how much variability there was likely to be, so I did a little experiment.
I took nine samples of three level tablespoons -- trying for as much consistency as possible -- of the coffee I'm using right now (Scarlet City Warp Drive) and weighed each one in turn. Here are my results in grams:
So the average dose was 17.24g with a standard deviation of .464 g, and a maximum deviation of 1.3 g. I'll leave it to others more experienced than I to interpret the import of this, but I can say from experience with my lever machine, that a .4 g variance around 17g will necessitate adjusting the grind for optimal results.
FWIW, my take is that a decent scale is practically a necessity for dialing in the shot, and especially if one changes coffees from time to time.