I read a bit more on it.
With decibels, 0 is full volume and about -80 is sufficiently quiet to be considered silent. Subtracting/adding 6db will half/double the volume.
So far so good. But using 20*log10(percentage) you get the same result:
Full volume (1) will give 0.
Half volume (0.5) will give -6.
Quarter volume (0.25) will give -12.
Either seems viable to setting volume and you can convert between the two and get expected results. The underling webaudio api uses percentages for volume.
Anyways, it’s merely a preference. People who have done a lot of audio processing seem to prefer decibels and a lay person like myself likes percentages.