This paper is the third part of three papers on sampling by the number of particles, focusing on analytical variability. The objective is to propose a target variability of waste and contaminated soil analyses (extraction and quantification), that can be used for calculation of the size of a representative sample. Data of intra- and inter-laboratory variability are presented. As the variability of the quantification step (after extraction) is limited in waste and soil analyses to about 0.01, the analytical variability stems from three main sources: (i) non-homogeneous test portions; (ii) for partial extraction methods, variable extraction rate, due to presence of options in the method or insufficient time for equilibrium (leaching or percolation test, biotests); and (iii) ill-defined solid/liquid separation (leaching or percolation tests), critical since there are colloids and nanoparticles in the leachates, representing from 0 to 100% of the element fraction in the leachate. Counter-intuitively, the centrifugation (annex E of EN 12457) series before the 450 nm-filtration delivers leachates more concentrated in particles (median size 150 nm, 1 sample) and statistically more concentrated in elements (+13%, 27 samples, 287 paired data). Without centrifugation, the filter cake that builds up on the membrane is an additional filter. A target intra-laboratory variability of CVr = 0.10 (10%) and inter-laboratory variability of CVR = 0.20 (20%) is proposed for all analytical methods. The methods with higher CVr and CVR should be revisited to not jeopardise the sampling and characterisation efforts of waste and soil, particularly for valorisation in the circular economy.