We tried two different methods of making stockpile measurements and compared their results. Overall, we found that making fairly "tight" polygons around the stockpiles yielded more reliable results.
We made 20 attempts at measuring the same stockpile. As a rule of thumb, if we made the points close to the stockpile (a "tight" measurement) we'd get results close to 2,900 cubic meters. If we were much more generous with the room around the stockpile (a "loose" measurement) it would be closer to 3,000 cubic meters.
Example of a "tight" measurement:
Example of a "loose" measurement:
As you'll see from the results below, over the 20 measurements there was an average variation of 2.3% from the average measurements.
We recommend the following:
Decide how you are going to use the Volume tool (i.e. are you going to take "tight" or "loose" measurements), and then consistently use that method. You'll see from the below chart, that when separate averages for "tight" and "loose" are calculated, the average variation falls to 0.8% and 1.6%, respectively. What this means is that as long as you use a similar method across their stockpile measurements, even though you might get a slightly different result each time, the results will be be close to a 1% margin of error from the "actual" size.
We would recommend that a "tighter" method is better, although not too tight that the points are actually on the edge of the stockpile—this would cause the surface to be altered using that point and drastically skew the result (see below).
Especially when visually it is difficult to tell the shape of the stockpile, it can help to turn on the Elevation overlay from the menu before making the measurement, as it can show the stockpile more clearly.
The takeaway here is that as long as you've done your homework in the data-capture phase and are using a consistent approach to the measuring, then you can be confident in the results.