Whenever I try anything new as I simulate a differential amplifier, I find I have to adjust a lot of things in the long-tail pair, until I believe it's performing true differential action. (Until I succeed it could be something else like a mixer or adder.) The output needs to make sense since it is supposed to subtract one waveform from the other. Common mode should not appear in the output.
The aim is to find a 'good operating point' combining supply voltage, input amplitudes, current levels, device sensitivities, etc.