A randomized complete block design (RBCD) is created in two stages.

The first stage consists in grouping the observational units according to one or more factors. We require that the number of units in each group be equal to the number of treatments. The number of groups should be equal to the number of replications.

In the second phase, we assign the units in each group to treatment groups at random.

A modification of this design would be if the number of units in each group is larger than the number of treatments (preferrably, a multiple). The units in each group can then be divided in subgroups. Each subgroup would be assigned to a treatment group as a whole.