It's about feasability. If you want to see if a reaction will happen you can chose your own two half-cells and measure the p.d. across them. This will be faster than setting up each reactant with a hydrogen half cell.
However, if you want to see is 30 reactions will happen: have three reactants A. B and C and react them each with 10 others, it is much easier to find their relative electrode potentials (setting the hydrogen cell to zero), which would be 13 experiments, than measuring the electrode potential for each experiment, which would be 30 experiments.