This is very normal behaviour for reception in out-of-footprint areas. Many explanations for this phenomenon such as ionospheric/atmospheric absorption etc. have been presented too. My best bet though, is that these variations are due to varying attitude of the satellite.
The official footprints are based on how the satellite signal "illuminates" the earth when aimed at "boresight" ie. the decided central point of it's beam on the earth. Since all stability in orbit is artificially produced, either by gyros or spinning the entire satellite body, combined with a virtually frictionless environment we end up with a situation where every movement will continue forever unless changed by another force.
To keep the satellite at the right attitude (pointing it's beam at boresight) small trusters in different directions are fired periodically as needed. In order not to consume the entire supply of thruster fuel in a short time some slack in the attitude control must be accepted. As a result the satellite attitude will slowly wiggle forth and back over the direction we call boresight. At one end of the wiggeling cycle the signal beam may favour an out-of footprint area only to disappear there when the cycle is opposite. Remeber that signal strength tend to roll off steeply at the edges of a footprint so small changes can make all the difference. In the intended coverage area (within the footprint) users will hardly notice this effect since the signal strength there is not affected by small variations in satellite attitude.
The lifetime of a geostationary satellite is by far most limited by how much truster fuel it carries at launch and how wisely it is spent over time.
Ize
! . 
