I typically use a Phantom 4 Pro for all of my stockpile volume work. I capture images with Pix4D Capture and process them with DroneDeploy. I had a mishap with the P4Pro and had to use a Mavic 2 Pro for a couple of flights. I was able to replace the P4Pro with a V2 and flew a site that has been flown a total of 5 times, including today. The previous 2 flights were with the M2Pro. The measurements made with the P4Pro and M2Pro have been very consistent. Anyway, today I flew the site with the P4Pro V2 and used DroneDeploy to capture the images. When the map finished processing, I copied polygons from the previous flight and had to adjust the X and Y position slightly to get them to match the piles, which is no big deal. What was a big deal, however, is that all of the volumes increased from the previous flight by an average of 2%. The problem is that no material has been added to any of the piles, only removed. I checked the length of the truck scale and the length measurement difference from high to low is only .04% of the average of 3 measurements. My question is whether the flight app or type of drone can affect the volume outcome that drastically.
Are you using ground control points? If not then I would say that while a different drone can produce different results the more likely cause is probably the difference in flight controllers. Even if all the specs were the same the way the drone flies even something small like how DD and Pix4D corner differently can affect repeatable accuracy. I would at least find one monument on the site with a static elevation that you can calibrate the maps to vertically. DroneDeploy tries to align them but that is just a rough adjustment.