Great point for other people who are tring this!
I faced this dilemma when calculating the amount to be donated to CATF. My colleagues raised that we perhaps should use the ‘average cost’ to offset a ton of CO2 (assumed as $10) for the calculation. I was fine with the approach, of course, but since it was mainly one partner in the company that did the offset I did not want to multiply the sum by a factor of 5, in which case he may not have been willing to just pay for it himself and instead raise it to all the partners where it could have been blocked (unfortunately I think one of the partners would block such contribution).
I may raise this when calculating and offsetting the emmissions for 2019.
PS: I always used the upper boundaries for the $/CO2 estimate as well as any other aspects of the calculation, even adding a 25% of the calculated CO2 as ‘unnacounted elements’ (some known, some unknown).