It depends!
If you have an external clock - that cable matters - a lot. This clock controls the timing of the internal d to a conversion process, so any jitter affects SQ. That cable is the most expensive one in the digital part of my system.
If you are using I2S - over an HDMI cable or RJ45 - it matters a lot, once again timing data is being transmitted over the cable.
If you are using Toslink - stop (except possibly to hook up your TV).
If you are using S/PDIF Coax - possibly. The cable must be "decent" - 75 ohm, shielded. Then it depends on the implementation of the S/PDIF in your DAC. If the DAC implements a good buffer, isolation and reclocking, then an OK cable will sound as good as a superb one. In "olden days" then the cable was critical as the DAC's clock was derived from the input data stream. If your DAC still does this then the cable matters a lot.
With USB - the cable needs to be good enough that the received data is correct, jitter is probably irrelevant as the data presented to the actual d to a conversion has been buffered and timing is solely dependent on the DACs clock.
But then there is the electrical noise that might be picked up by the cable. Then it depends on how well the interface was implemented, does it provide isolation so that electrical noise does not affect the DAC, unless the noise level is comparable to the signal level which could change the data? If there is any weakness in that aspect then the shielding or filtering of the cable matters.
My DAC is the Esoteric K-01XDSE wherein the engineers went to extreme lengths to isolate the d to a conversion process from the effects of jitter and noise on the input; I tried switching the Cardas USB cable for one that had come with a printer and could detect no change in any aspect of SQ given intense critical listening. In my layout the cable is in a pretty benign electrical environment, well separated from power leads etc. YMMV.