That is 95%+ of all accidents where the aircraft was fitted with a fixed ELT, and includes not just GA, but some airline accidents as well.
That's not a very useful statistic. What you need to know is activation rate in accidents where it might make a difference i.e. with survivors. (The same type of accidents you use for your statistics for portables.) For the purposes of GA, we should consider GA accidents only - not airline. It's debatable whether accidents into water where it's guaranteed not to work should be included (35% of aircraft accidents in Australia end up in the water? Really? How does that happen?)
So of GA accidents in Australia on land with survivors, what is the activation rate?
I'll give you one where it didn't activate - VH-POJ near Horsham. It's unclear from the report whether the survivor could have activated a portable beacon if they had one - I suspect not, either due to injuries or unfamiliarity.
A quick read of some reports of ELT effectiveness suggests that a major factor is that the crash G is not in the right direction to activate it (aircraft slews sideways etc?), so maybe a helicopter style multi-axis g switch is what's required?
So the failure rate of an ELT ... is ... 95%+ on land.
...
By contrast, where there are survivors to operate a portable, the technical failure rate of portables to broadcast a signal successfully is near to zero as makes no difference.
Fixed ELT failure rate is 95%, portable is near enough to 0... sounds impressive (too good to be true!) but you are not comparing apples to apples there. I'm not sure whether you are being deliberately misleading, or not understanding what you are comparing?