I would say that the “misleading” part is that even though the quadrupole moment becomes a better approximation for l<<d, the actual size of the of the quadrupole moment becomes much smaller. This should come to no surprise, since if l decreases, the nucleus can be more approximated like a point, and the need to describe any deviations from a spherical shape becomes negligible. Luckily, you would need a rather ridiculously large nucleus to reach an quadrupole moment that seriously diverges from the analytical solution.