Let me give a shot too:

1.) There should be atleast one clustered index for a table which should be Narrow, ever-increasing, unique

2.) Try think of creating clustered indexes on the ones which are used in range queries and the ones which are used for ORDER by clause

3.) Frequently updated columns and non-unique columns are not a good choice

4.) Non- clustered indexes when needed but be careful about the number of indexes . more number of indexes more performance overhead on your table for DML operations

5.)Think about Covering indexes to cover all the columns retrieved in your query

6.) Order of Index columns pretty important as SQL server maintains indexes only on the lading edge of the indexed columns

7.) Give consideration about fill factor too when creating indexes, inappropriate fill factor may lead to index fragmentation