Clonal variation, wherein a range of specific productivities of secreted proteins are observed from supposedly identical transformants, is an accepted aspect of working with Pichia pastoris. It means that a significant number of transformants need to be tested to obtain a representative sample, and in commercial protein production, companies regularly screen thousands of transformants to select for the highest secretor. Here, we have undertaken a detailed investigation of this phenomenon by characterising clones transformed with the human serum albumin gene. The titers of nine clones, each containing a single copy of the human serum albumin gene (identified by qPCR), were measured and the clones grouped into three categories, namely, high-, mid- and low-level secretors. Transcriptomic analysis, using microarrays, showed that no regulatory patterns consistently correlated with titer, suggesting that the causes of clonal variation are varied. However, a number of physiological changes appeared to underlie the differences in titer, suggesting there is more than one biochemical signature for a high-secreting strain. An anomalous low-secreting strain displaying high transcript levels that appeared to be nutritionally starved further emphasises the complicated nature of clonal variation.Electronic supplementary materialThe online version of this article (doi:10.1007/s00253-017-8317-2) contains supplementary material, which is available to authorized users.