diff --git a/boofilsic/settings.py b/boofilsic/settings.py
index 0fe28d46..ddf122d0 100644
--- a/boofilsic/settings.py
+++ b/boofilsic/settings.py
@@ -70,6 +70,8 @@ env = environ.FileAwareEnv(
     NEODB_DOWNLOADER_CACHE_TIMEOUT=(int, 300),
     # Number of retries of downloader, when site is using RetryDownloader
     NEODB_DOWNLOADER_RETRIES=(int, 3),
+    # Number of marks required for an item to be included in discover
+    NEODB_MIN_MARKS_FOR_DISCOVER=(int, 1),
     # Disable cron jobs
     NEODB_DISABLE_CRON=(bool, False),
     NEODB_DISABLE_CRON_JOBS=(list, []),
@@ -180,6 +182,8 @@ INVITE_ONLY = env("NEODB_INVITE_ONLY")
 # If you are running a development server, set this to True to disable this behavior
 DISABLE_DEFAULT_RELAY = env("NEODB_DISABLE_DEFAULT_RELAY", default=DEBUG)
 
+MIN_MARKS_FOR_DISCOVER = env("NEODB_MIN_MARKS_FOR_DISCOVER")
+
 MASTODON_ALLOWED_SITES = env("NEODB_LOGIN_MASTODON_WHITELIST")
 
 # Allow user to create account with email (and link to Mastodon account later)
diff --git a/catalog/jobs/discover.py b/catalog/jobs/discover.py
index ddf11d28..9041f6cd 100644
--- a/catalog/jobs/discover.py
+++ b/catalog/jobs/discover.py
@@ -10,14 +10,14 @@ from common.models import BaseJob, JobManager
 from journal.models import Comment, ShelfMember, q_item_in_category
 
 MAX_ITEMS_PER_PERIOD = 12
-MIN_MARKS = 1
+MIN_MARKS = settings.MIN_MARKS_FOR_DISCOVER
 MAX_DAYS_FOR_PERIOD = 96
 MIN_DAYS_FOR_PERIOD = 6
 
 
 @JobManager.register
 class DiscoverGenerator(BaseJob):
-    interval = timedelta(hours=2)
+    interval = timedelta(hours=1)
 
     def get_popular_marked_item_ids(self, category, days, exisiting_ids):
         item_ids = [
diff --git a/compose.yml b/compose.yml
index 46f5bcdb..3d860aac 100644
--- a/compose.yml
+++ b/compose.yml
@@ -34,6 +34,7 @@ x-shared:
       NEODB_DISABLE_DEFAULT_RELAY:
       NEODB_DISABLE_CRON:
       NEODB_DISABLE_CRON_JOBS:
+      NEODB_MIN_MARKS_FOR_DISCOVER:
       NEODB_SENTRY_DSN:
       TAKAHE_SENTRY_DSN:
       NEODB_DB_URL: ${NEODB_DB_URL:-postgres://neodb:aubergine@neodb-db/neodb}