Gail already mentioned some of the caveats of the missing-index DMVs.
Here are a couple more:
(1) Because the data in the DMVs are kept in memory, they are at risk of being flushed from the cache if there is memory pressure. Therefore, the DMVs need to be polled regularly and stored in a database for later analysis. This is something Gail has mentioned before in one of the posts here.
(2) The DMVs are no substitute for a strategy of regular index defragmentation and statistics updating. The optimizer may make the wrong choices when it constructs the 'best' execution plans if indexes are defragmented and/or the stats is not up-to-date and accurate. This can cause "wrong" data to be deposited to the DMVs and therefore erroneous information.
I absolutely love the DMVs. They have opened a much-needed window into the internals of the database engine and have made DBA work all the more interesting. I use them all the time.
We just need to keep these limitations in mind as well.
SQL Server 2016 Columnstore Index Enhancements - System Views for Disk-Based Tables[/url]
Persisting SQL Server Index-Usage Statistics with MERGE[/url]
Turbocharge Your Database Maintenance With Service Broker: Part 2[/url]