Our previously reported ‘blue solution’ oxygen-evolution catalyst consists of an isomeric mixture of coordination complexes containing the (pyalk)IrIV–O–IrIV(pyalk) core unit and is thus entirely molecular but only when formed from organometallic precursors such as Cp*Ir(pyalk)Cl or Ir(pyalk)(CO)2 (pyalk = (2-pyridyl)-2-propanolate). We now show that attempts to form it from such obvious coordination precursors as Na[Ir(pyalk)Cl4] or Na[IrCl2(pyalk)(O2CPh)2], under a variety of conditions, always fail in our hands, leading to a mixture of molecular ‘blue solution’ species and IrOx nanoparticles, rather than the purely homogeneous catalyst formed from organometallic precursors. The loss of the pyalk ligand during the oxidative activation is associated with the nanoparticle generation. External chelating ligands also failed to stop the nanoparticle formation. This work implies the paradoxical conclusion that organometallic complexes are effective as catalyst precursors, even when coordination complexes are the catalytically active species, since the inner-sphere organometallic ligands, although ultimately lost or degraded, may nevertheless have a stabilizing effect sufficient to suppress undesirable nanoparticle production pathways in the activation of the catalyst precursor. A key aspect of the present study is that organometallic complexes in general may be useful catalyst precursors even if the organometallic ligands are lost in the activation process.