Longer is better?

There is unlikely to be a difference. But theoretically really short ~ 6” can cause reflections… but unlikely. 3’ is great to give you flexibility. It is the speaker amp connection that is really good short. 

Can’t say for sure but some very knowledgeable folks here in the past as well once asserted short digital wires are more prone to reflections than longer ones.  I respected those folks technical acumen enough to always lean a bit longer when in doubt. 

USB - Curious Cable makes a 200mm (just under 8 inches)  "Regen Link", which is a short USB cable originally designed for the connection between an UpTone Audio USB REGEN and a DAC.  I have a similarly short USB cable from Wireworld.  Their top level Platinum Starlight® 8 USB 2.0 can be purchased in a 12 inch length. Curious Cables’ Rob Woodland says this about USB cable length:

Does usb length matter?

First  up, usb is different from spdif 75 Ohm digital cables.  Remember the  argument about 1.5m being the ideal length for spdif cables?  This  doesn’t apply to usb cables. 

When it comes to usb, shorter is better i.e. any length under 2 metres is ideal. 

Some  audiophiles with high resolution systems may notice a slight difference  between a 200mm and 2 metre Curious cable.  However – this is very  slight.  Some may say miniscule!  

With S/PDIF cables, there are experts who believe a (minimum or optimal) length of 1.5M should be used to prevent reflections that can result from "imperfect" (RCA) connectors.  At a length of 1.5m (or greater), the time required for a reflection to travel the cable’s length three times is sufficiently long to prevent it from distorting the timing of the next "bit," and may therefore result in lower jitter than with a short cable.

In the world of high-end audio, where many enthusiasts insist "everything matters", I do not understand why there is not more concern with manufacturers continuing to use RCA connectors on their high-end DACs, when BNC connectors are considered more accurate.  A BNC connector’s true 75-ohm impedance and secure bayonet lock improve shielding, impedance matching, and signal integrity, which reduces reflections and noise.  Would a minimum digital cable length even be necessary for cables with BNC connectors?

The minimum length doesn’t seem to be as big a deal for 110-ohm AES/EBU cables using XLR connectors. 

@mitch2   It is the same bit that is affected.  Beginning of transition ("knee") reflection from the end of the cable comes back and affects the shape of the same transition at source.  Reflection comes from any characteristic impedance boundary.  Signal travels at about 5ns/m making total propagation time 2x1.5mx5ns/m=15ns.  Reflection will deform originating transition after midpoint (threshold) since typical transition is approx 25ns (won't affect 12.5ns threshold point).   It is guessing game since transitions can be longer or shorter while signal propagation depends on cable dielectric.  Perhaps 2m would be even safer.

As follows is what AI had to say on the question at hand.  Interesting. 

Shorter USB cables are generally better for connecting a streamer to a DAC in audio setups — they minimize electromagnetic interference (EMI/RFI pickup), voltage drop (especially if bus-powered), and overall signal risks.

•  Longer cables increase potential problems — more chance of noise, power issues, or even dropouts beyond ~5m (USB spec limit), with no real benefits in typical modern asynchronous DACs.

•  The “reflections” argument for longer cables comes from a niche audiophile claim: very short cables (<1–1.5m in some old discussions) might not let signal reflections (from impedance mismatches) decay fully before the next data bit, potentially worsening jitter in poorly matched systems.

•  In practice, this claim is weak and not widely supported — modern asynchronous USB DACs reclock and buffer data with their own precise clock, largely eliminating cable-induced jitter/reflection effects; measurements (e.g., Audio Science Review, Archimago) show no audible/measurable audio differences from cable length variations when the link is stable.

•  Consensus from engineering & objective tests — shorter, well-made, properly shielded USB cables (0.5–2m) are reliably better or neutral; deliberately going longer to “fix reflections” is mostly cable-marketing hype, not physics-backed evidence.

•  Bottom line — Your retailer is likely stretching (or inventing) a reason to sell a longer/premium cable. Stick with a short, quality one for best results.