Is this expensive Clustered index seek good or bad? If not how could I improve it
In terms of the plan itself, one thing worth mentioning is that each of the seeks has a residual predicate (a test applied to rows that meet the seek criteria). If these predicates can be incorporated into the index keys instead, you might find an improvement. Not that there's much improvement to look for in a 2ms execution, but the general point is valid.