The short answer is that NASA didn't intend for Skylab's orbit to decay as rapidly as it did, but when it became apparent that the station was descending faster than expected, they attempted to use what means they had to make it decay in a more controlled manner.

Skylab was launched in May 1973 and hosted three successive resident crews, with the last one returning to Earth in February 1974. At that time, the station's propellant and consumables were diminished but not yet exhausted, with a later analysis estimating that enough oxygen and nitrogen remained on board to support an additional stay of about ninety days. Because the station might still prove useful, immediately after the landing of the final crew controllers oriented Skylab vertically with respect to Earth and turned off all power and communications.

Prior to returning to Earth, Skylab's final crew had used the fuel of their Apollo spacecraft to boost the station's orbit up to 455 x 433 kilometers (283 x 269 miles) in an effort to extend its lifetime as long as possible. This was deemed necessary because of the phenomenon of orbital decay: although the boundary of "space" is officially regarded as lying at 100 kilometers (62 miles) above sea level, the Earth's atmosphere does not simply stop at that altitude. The extremely tenuous outer layer of the atmosphere, known as the exosphere, extends thousands of kilometers out from the planet until it merges with the interplanetary medium. As a consequence, any object of any size in low Earth orbit is gradually slowed by the drag it experiences colliding with these trace atmospheric particles, resulting in its orbit gradually spiraling down into the lower atmosphere. Furthermore, the density of the atmosphere at a given altitude is not constant, and it can increase or decrease based on the expansion and contraction of the upper atmosphere due to solar heating.

All of this means that while the lifetime of a given orbit is predictable to some degree, there is always an unavoidable level of uncertainty in these predictions. At the time power was turned off to Skylab, the station was predicted to reenter the atmosphere and be destroyed in early 1983. This date for fortuitous for those who wanted to further utilize Skylab, because the Space Shuttle (then under development) was expected to be operational by the late 1970s and could return to the station for a rescue mission. These plans called for a shuttle to attach a module to the station containing fuel and rocket engines that would boost Skylab to a higher orbit, allowing it to be returned to service.

Unfortunately, this is where that uncertainty in orbital prediction comes into play. NASA knew that the upper atmosphere would inflate going into the 1979 solar maximum, but did not predict that it would inflate as much as it did. Skylab's orbit was decaying faster than expected, and it eventually became clear that without intervention the station would reenter well before the Shuttle (experiencing its own development delays) would be ready for flight. It was therefore decided to reactivate the station in early 1978 with the goal of keeping it flying long enough for the Shuttle to visit it; or, failing that, to control its orbital decay in such a way that its reentry would minimize the risk to human life, since large fragments of wreckage were expected to survive and hit the surface.

This control could be accomplished by changing the orientation of the station. I mentioned above that when it was switched off in 1974, Skylab was oriented vertically with respect to the Earth. This orientation was advantageous because it could be maintained without any power or fuel, but it also presented the most surface area to the tenuous atmosphere, essentially making the orbit decay at the fastest possible rate.

When Skylab was switched back on in 1978, NASA flight controllers almost immediately oriented the station so that its long axis was parallel to its direction of motion and the Earth's surface, like an arrow. This reduced the atmospheric drag and extended the orbit's lifetime. It was still a delicate balancing act, because the attitude could only be controlled by the station's gyroscopes or maneuvering thrusters, neither of which were limitless resources. One of Skylab's three gyroscopes had already failed while a crew was onboard and another was showing signs of failing, and the maneuvering thrusters had only limited fuel remaining. The "arrow" orientation was only meant to buy time until Skylab's eventual fate became clear.

By 1979, it was evident that there was no way that the Shuttle would be ready in time to visit Skylab, and that the station was definitely going to reenter the atmosphere later that year. With that in mind, the plan became to minimize the risk to the population from large space station fragments impacting the ground. As the Earth turned beneath Skylab's orbit, the station passed through "bad" orbits and "good" orbits. "Bad" orbits were those where the station passed above population centers or highly developed areas, while "good" orbits spent much of their time over ocean or sparsely-populated areas. By observing Skylab's accelerating decay and changing its orientation over time, flight controllers could either accelerate or slow the rate of orbital decay to aim for one of the "good" orbits.

This strategy mostly succeeded. Skylab entered the atmosphere on July 11, 1979, on a "good" orbit just as intended. However, the station held together a few minutes longer than predicted before breaking up. Instead of all of the surviving fragments impacting the Indian Ocean, some fell on western Australia. No one was injured, but several pieces were recovered and (in one memorable case) a resident of Esperance collected a fragment that had landed near his home and quickly flew to California to claim a $10,000 prize that the San Francisco Examiner newspaper had offered for the first fragment delivered to them.