diff --git a/PROJECT.org b/PROJECT.org
index 08f9d2f..3c894b8 100644
--- a/PROJECT.org
+++ b/PROJECT.org
@@ -489,6 +489,36 @@ to GMT to save it in the database.
 
 It would be nice to not duplicate comments that exist on a task when it's first scrobbled.
 
+** TODO [#B] Consolidate albums in the same musicbrainz_releasegroup_id :music:albums:metadata:
+
+*** Description
+
+When we look up albums, we should check if one already exists with the same
+musicbrainz_releasegroup_id and prefer that one, rather than creating a new
+album.
+
+Also, we should create a data migration to clean up albums with matching
+musicbrainz_releasegroup_id fields by consolidating tracks around the first
+album found.
+
+** TODO [#A] Clean up metadata on music tracks :music:tracks:metadata:musicbrainz:
+
+*** Description 
+
+There's a over 3K tracks without a musicbrainz_id and almost 30K without a
+base_run_time_seconds. We should have a clean up script that can run through the
+ones missing musicbrainz_ids and see about getting metadata from musicbrainz and
+for the 30K without base_run_time_seconds, check with their musicbrainz_ids to
+see if we can look up the track length, and if not, tag the track with
+"missing-metadata" tag.
+
+Also, if the tracks without musicbrainz_ids actually are not in the MB database,
+we should tag those with "not-in-musicbrainz"
+
+And a big part of this work will probably involve checking for "Various Artists"
+tracks where the track erroneously got set with a generic artist. In those
+cases, we should try just looking up by track title.
+
 ** DONE [#B] Make artists_m2m field source of artist truth for albums :music:bug:albums:
 :PROPERTIES:
 :ID:       05a24455-0e71-45ef-ac6e-c1b7b843047d