I expect there are other operations as well. It sounds like you have a lookup operation going on. There's a nonclustered index on a heap table. The nonclustered index has enough data to assist in filtering the query, but it doesn't have all the data needed by the query. So an additional lookup operation, the RID Lookup, is going against the heap to retrieve the other necessary columns. To "fix the query" is hard to say without seeing the query, the execution plan, or your structure. I can give you some general advice. Most tables should have a clustered index. If the index being used in this query is the most frequent path to the data, it's possibly a candidate for being a clustered index. A clustered index stores all the data, so it won't need an additional lookup operation. Another possible solution is to use the INCLUDE operation to add additional columns to the leaf level of the nonclustered index. That will eliminate the lookup operation too. Which of these is a good solution for you? I just don't know without more information.