My hosting company told me they do not use SSD because they did not see a
significant change in performance over their 15k SAS HDD drives.
They are drunk, idiots, on drugs or simply used a benchmark making little sense or being too specific (like a web server that will not see a large difference - compared to a databsae server for example). A lot depends on the usage scenario - but it makes a generic answer like the one given to you meaningless.
I would like to know your opinions if the SSD drives would show noticeable
performance gains and if it warrants 300 per month?
You tell us. Basically - are you IO limited or not? Any OS worth a cent or nothing (i.e. open source) can measure IO load. If you are IO limited then yes ,the gain will be brutal - we talk of a factor of 100 times more IO as a guideline for a decent SSD (and that is a low number) a 15k SAS disc is measured at IIRC 450 IOPS... 45K IOPS are not exactly stellar for a SSD, and you can go up beyond 200k.
Virtualization - yes.
Databases - possibly. But not sure your tiny setup (100k products are not exactly a large database) hits here. Can be you serve it all out of memory anyway.
Basically a semi competent admin has to make an evaluation - it is possibly the benefit will be extremely, it is possible the benefit will be hardly something to measure. Totally depends where your bottleneck is now.
I replaced a RAID of 10k SAS HDD in a Raid 10 (8 of them) with a Raid 50 of 6 SSD and the throughput jumped. But then this is a 3tb database that at times does heavy analysis - with just 48gb memory as a buffer. Use cases matter extremely here.
OTOH I also use SSD generally as boot discs on workstations on computers - it does make a difference during patch days and a small SSD is not exactly that expensive (and yes, you likely do get ripped off).