This isn't Newton's law of cooling applied to a single average temperature — it's a genuine 2D axisymmetric finite-volume simulation. The can or bottle is divided into a ring-and-layer grid; each cell exchanges heat with its neighbours by conduction (boosted by an empirical correlation that stands in for internal buoyant mixing) and with the environment by convection (natural or forced, using standard Churchill correlations) and radiation (Stefan–Boltzmann). A lightweight buoyancy check lets warmer, lighter fluid rise past colder, denser fluid — the same effect that makes a fridge-cold drink's core stay warmer than its surface for a surprisingly long time.
A 355 mL can starting at 22°C in a 4°C still fridge takes roughly 3 hours to approach fridge temperature — matching the common "give it a few hours" advice almost exactly. Switch to a fan-forced fridge and an 8°C target, and the same can gets there in about an hour. Drop it in an ice-water bath instead, and the target falls in a few minutes — ice water is dramatically more effective than air because water conducts heat away far faster than a breeze ever can.
Water's convective heat-transfer coefficient is roughly two orders of magnitude higher than air's, even still air against forced air. That's why the classic party trick — ice water plus salt, or just ice water alone — beats any fridge for speed, even a fan-forced one.
Heat has to conduct (and buoyantly mix) its way from the centre out to the wall before it can leave. Right after you put a warm drink in the cold, the surface responds almost immediately while the core lags well behind — which is exactly why a "cold to the touch" can isn't always cold all the way through yet.
Conduction moves heat between adjacent fluid cells inside the drink. Convection is heat leaving through the container wall into moving air or water outside. Radiation is the (usually small) contribution from the container's surface glowing in the infrared toward its cooler surroundings — included here because it's rarely negligible for a still fridge.
A bottle's neck has far less surface area relative to its volume than a can's cylindrical body, so heat has fewer places to escape from near the top. The simulation models the actual tapered geometry — body, shoulder, and neck — rather than treating every container as a plain cylinder.