They are expected to provide micro millimeter consistent accuracy, when the end user is cutting with a dull axe.
How many samples are a few pounds dumped into a plastic bucket, stirred and shaken and then enough pulled out to fill the little paper sack provided? Then this may represent a whole field or one hector, 2.5 acre grid, or a zone.
I have beat the world to death talking about the foot by foot variation in real fertility values. So if the sampler can not return to the exact same ft² each time GPS is a little better than eyeball but not much. Then the grower will want to find out why the results from the lab is inconsistent. What we are paying for is for 10% normal variation from zero. Talk to the sewage folks who must test the ground they spread sludge on. They take multiple samples, and pay 10 X what we pay. This is on a fairly small patch of ground.

Old farmer brown may have pulled samples for 10 years himself, sent a pubescent kid to sample for 5 years, had the fertilizer dealer sample for another 5 years and then has a commercial consulting firm pull the samples for two years and all the time complaining about inconsistent results from the lab.

The labs are set up to grind up a standard sample sack full of dirt, not a 5 lb sack full. For my probe, 4 probes nicely fit in one sack. I close the sack and label it #1. Then I put another 4 probes in another sack, sample # 2. I seldom send less than 3 separate sacks of soil samples per field. Of the 15 to 20 separate samples I do send, I know which ones go together. I look at those for irregularities. If there are none I average the results and figure my program. Not perfict but reasonably accruate.