1) Setting severity scale for the effects :
This is most common challenge , if we are setting a low severity for a real effect/criteria , it might impact the business critically and can give an adverse effect when occured.
2) Missing out the correct level :
When the effect of failure has to be considered by the team in manufacturing or assembly level itself, but missed at these stages and identified in the later levels the impact of the effect could be worse and might end up with higher severity rating.
So looking at the effect/criteria a little more closer and also in detail