Is it ok to average ranks?

Dear users,
I have a following problem, I have a matrix of results of testing of some chemical substances on a set of various experimental objects. I would like to somehow express general sensitivity of the objects to the chemicals applied experimental objects. The sensitivities expressed as concentration reducing half of the biomass.
I was thinking about doing the following:
1. rank all experimental objects by their sensitivities to each compound
2. taking average of ranks for individual objects
3. ranking the resulting values.
So the lowest rank of average ranks should get the object that showed on average highest sensitivity to the compounds.

Is it ok? I dont feel very comfortable about averaging ranks, but with medians there are problem that many compounds are toxic for only few objects so their in most cases you are ranking objects insensitive to the lowest concentration tested so their are getting middle rank. Taking median insteaed of average in step 2 means that you all the resulting numbers are the same.

Thank you for your opinions.