Firstly, what sort of cardinality do you have on AddrType.

The sample queries all appear to state a value of 'ME', so is that one of many AddrType values?

My thinking here is the clustering sequence being switched from CustNo, AddrType, to the other way round of AddrType, CustNo. I.E. Keeping all the ME's relatively closer for accessing with fewer reads.

Secondly, as Grant rightly says, are you really using all these indexes or think their presence is helping?

Thirdly is it the CustomerReport table you are later rebuilding the indexes upon, or the RETDCUST_CUSTADDR table? Again the order of things is not totally clear.

Personally, I would see how it runs with significantly less indexing. If as I said above AddrType='ME' is a mere percentage of the 1.2Million rows the clustering order can be a big factor on your perceived application performance end-to-end.