Lots of different things. Particularly at QN.

Wind shifts or gusts can disrupt the aircraft’s approach profile and trigger a ROPS warning (overrun prevention), prompting a mandatory go-around. This system is particularly sophisticated and conservative in protecting landing performance margins. Although ROW/ROPS doesnt play too much of a factor, the only ROW/ROPS warning I've ever seen in ZQN is "if wet, too short" which you can continue with of the runway is dry, Ive also had a couple of "braking, MAX braking" warning after landing. This normally occurs when you land too softly and the autobrake takes a few seconds to kick in until the spoilers deploy and you get full "weight on wheels", but that is all once you actually on the ground.

The GS Mini and auto thrust has got a lot to answer for on the airbus, especially on the NEO's as those engines can be a little slow to spool up and they use a slightly different wind formula. G/S mini certainly put us on the back foot sometimes compared to the boeings. They land with manual thrust so is you see the speed bleeding off (or increasing) you can respond straight away without waiting for a computer to work it out.

One of the most common scenarios in Queenstown (especially landing on 23) is that the southerly circles round the arrowtown basin and gives you a tailwind when you come past lake Hayes and past Morven, the GS mini then sets the thrust in an attempt to give you a constant speed across the ground. Then, as you cross the fence, you pop out from the relative shelter in the lee of the remarkables, into the full southerly (from the left), the wind goes from a moderate tailwind, into a moderate head/cross wind. That gives you a sudden and significant increase in airspeed and the thrust drops to idle.

Having the thrust at idle is a company SOP based go round, some guys are happy to delay the go round to see if things trend back, but landing with thrust idle is a very bad idea as it takes a LONG time for engine thrust to increase from idle if you need more energy for the flare or go around. Beyond that, it’s not just about visibility. The basin can produce mechanical turbulence and localised windshear on short final, so one aircraft might hit a brief lull and get in, while the next encounters a gust peak or directional swing. Crosswind gust spread can also be a factor - steady wind might be within limits, but the gusts can push control inputs or drift close to tolerances. Timing really matters when the wind is cycling around the valley. Stable approach gates play a role too, but equally it can come down to wind component changes near minimums, tailwind limits, or simply an approach that doesn’t feel comfortably within margins. Different operators also have slightly different tolerances and policies, which can explain why one flight lands and another goes around in seemingly similar conditions.

More on ROPs here, it's quite fascinating. https://safetyfirst.airbus.com/furth...unway-overrun/
More on approach gates - https://skybrary.aero/tutorials/stabilised-approach
More on G/S mini https://safetyfirst.airbus.com/contr...h-and-landing/