In their excellent paper, Bekas & Michev1 present a sober assessment of the inherent weakness of the Mental Health Clustering Tool and ICD-10 coding. Although clustering has already been used for many years in acute care, what is suitable for acute care is not necessary applicable to psychiatry. We are expected to cluster patients with similar symptoms, needs and disabilities in 21 clusters which are used as the basis for financial funding. However, subjectivity in psychiatry is a fact and it does not really matter how many tools and scales we implement to change this. The chance of subjectivity may be reduced but never eliminated. Diagnosis and formulations vary between clinicians within the same profession and even between members of the same team. One can identify quite easily a sizeable number of patients with an ever-changing diagnosis over a number of admissions. It follows that clustering is not a static tag but a changeable process that ought to be regularly updated. I agree wholeheartedly with Bekas & Michev that the final arbiter should be clinical judgement. It is not uncommon practice for clinicians such as myself to override the cluster concluded by other members of the team, relying on and trusting my clinical judgement.