There are a lot of variables that can effect the efficiency numbers. Oxygen content of the air will effect the efficiency of the burns. Humidity. Ambient temperature. Surface you put the stove on. Height of the burner/pot. Etc... 10% is within tolerance. I usually give these numbers a spread, min, max and average. I call 15% a significant deviation to account for the crudity of most tests. For example, depending on weather you have a burner on high, medium or low, but there is no set definition for these terms. People eyeball the medium setting. High is the maximum for a stove, but this can vary between 4500BTU and 11000BTU. Low can sit there forever without boiling a liter of water, yeilding inconclusive results. Some stoves will not run evenly on lower settings (Simmerlite for example.) And, how low is low? Too low will drive *up* fuel consumption.
I usually try for a 10-12 minute burn for a half liter. This usually produces the best fuel efficiency. Longer means more heat is radiated out. A 20 minute burn for one liter is barely tolerable, though. If you want a 10 minute burn, you will use more fuel. This is assuming all else remains the same: pot, lid, distance, heat screen, etc. For 2 Liters, the numbers can get worse. At least with the Joule, it will make a good cold weather stove(<20F) for a larger group(>4 people.) I think this is where Ryan is headed with this. For me, myself and I, well, we have no use for it. But it is interesting to read about these little toys.