When the compression is large enough to be relevant to whatever you’re calculating or discussing. Part of my job is literally to develop models that, for instance, predict the compressibility of water. For pretty much all practical purposes I can attest that the isothermal compressibility of water is completely negligible.