Jason,
I have executed Reindex on all the indexes. Currently, fragmentation is < .02 or = 0 on every index.
I included the fields suggested in the non clustered Row Index lookup (RID). This did not affect performace.
Luis,
I am a VB programmer who dabbles in SQL when I need to. I know I can get better performance by returning the 600 records from GMFPick and doing the subsequent lookups. But I am trying to understand what I am going wrong here with SQL.
I have attached the Actual SQLPlan for the CTE.
Please let me know if other information would be helpful.