Should I consider going for a longer stem at a certain point?
For the changes you are looking at, probably not, however at some point you may need to change your stem, but surprisingly you typically need large changes in stack before you really need to worry about stem length. Note that as you drop the stem, your reach automatically gets a little bit longer due to the head tube angle. The question then becomes how much can you drop the stack before you need to change the stem length. The actual answer surprised me.
Over the years I took notes on what I found to be comfortable stack and reach dimensions under a variety of riding conditions (Figure 1a). Not surprisingly, under easier riding conditions I tended to prefer a higher stack and shorter reach dimension. When I compared this to the fit dimensions produced by various stem lengths if I only altered spacer stack height (Note, this is only a exercise, most bikes only allow 30 mm of adjustment in stack height), I found one stem length covered a wide range of stack heights (Figure 1b). Moving from the tallest position, I would need to switch to a slightly longer stem fairly soon, but after than the longer stem would adequately cover a wide range of stack heights. As I dropped the stack even further my preferred reach shortened a bit relative to how the fit dimensions changed by reducing the stack, requiring a switch back to the original stem.
This exercise assumed a head tube angle of 72 degrees, a very common angle for endurance road bikes. This makes me suspect this is why road bike geometry has centered around this head tube angle (it is possible to calm steering response down or speed it it with other geometry tweaks so you have a little freedom in your choice of head tube angles).
Finally, as an aside some may ask why my stack height preferences varied so much. I was wondering the same thing so I plotted it against average power readings and was quite surprised at strength of power as a predictor (Figure 2). In short, under higher efforts you support more of your upper body through core activation, and we often use our upper body weight as a counter balance to pedal forces.
Figure 1. Preferred stack to reach dimensions under a variety of riding conditions (A) and (B) fit dimensions of provided by two stem lengths when stack height is changed by spacers. Solid blue line indicates polynomial regression line, with shading indicating the 95% confidence region.
Figure 2. Preferred stack dimensions under a variety of riding conditions and associated average power. Solid blue line indicates polynomial regression line, with shading indicating the 95% confidence region.