diff --git a/src/rpft/cli.py b/src/rpft/cli.py
index 66fffef..d389afa 100644
--- a/src/rpft/cli.py
+++ b/src/rpft/cli.py
@@ -72,8 +72,11 @@ def create_parser():
         help=(
             "Tags to filter the content index sheet. A sequence of lists, with each "
             "list starting with an integer (tag position) followed by tags to include "
-            "for this position. Example: 1 foo bar 2 baz means: only include rows if "
-            "tags:1 is empty, foo or bar, and tags:2 is empty or baz"
+            "or exclude for this position. To exclude, precede by '!'.\n"
+            "Example: 1 foo bar 2 baz means: only include rows if "
+            "tags:1 is empty, foo or bar, and tags:2 is empty or baz\n"
+            "Example: 1 foo bar ! 2 !baz means: only include rows if "
+            "tags:1 is foo or bar (but not empty), and tags:2 is not baz"
         ),
     )
     return parser
diff --git a/src/rpft/parsers/creation/tagmatcher.py b/src/rpft/parsers/creation/tagmatcher.py
index 7e4d8bd..6110ffa 100644
--- a/src/rpft/parsers/creation/tagmatcher.py
+++ b/src/rpft/parsers/creation/tagmatcher.py
@@ -3,7 +3,8 @@
 
 class TagMatcher:
     def __init__(self, params=[]):
-        self.tag_patterns = collections.defaultdict(list)
+        self.include_patterns = collections.defaultdict(list)
+        self.exclude_patterns = collections.defaultdict(list)
         if not params:
             return
         current_index = None
@@ -20,11 +21,18 @@ def __init__(self, params=[]):
                         "Tags parameter must start with a "
                         "number indicating the tag position."
                     )
-                self.tag_patterns[current_index].append(param)
+                if param[0] == "!":
+                    self.exclude_patterns[current_index].append(param[1:])
+                else:
+                    self.include_patterns[current_index].append(param)
+                    # Empty string is accepted by default
+                    self.include_patterns[current_index].append("")
 
     def matches(self, tags):
         matches = True
         for i, tag in enumerate(tags):
-            if tag and i in self.tag_patterns and tag not in self.tag_patterns[i]:
+            if i in self.include_patterns and tag not in self.include_patterns[i]:
+                matches = False
+            if i in self.exclude_patterns and tag in self.exclude_patterns[i]:
                 matches = False
         return matches
diff --git a/tests/test_tagmatcher.py b/tests/test_tagmatcher.py
index 4e09ef3..04f2378 100644
--- a/tests/test_tagmatcher.py
+++ b/tests/test_tagmatcher.py
@@ -38,3 +38,33 @@ def test_two_tags(self):
         self.assertFalse(tm.matches(["something"]))
         self.assertFalse(tm.matches(["something", "foo"]))
         self.assertFalse(tm.matches(["something", "baz"]))
+
+    def test_exclude_tag(self):
+        tm = TagMatcher(["1", "!foo"])
+        self.assertTrue(tm.matches([]))
+        self.assertTrue(tm.matches([""]))
+        self.assertTrue(tm.matches(["bar"]))
+        self.assertFalse(tm.matches(["foo"]))
+
+    def test_excludetwo_tag(self):
+        tm = TagMatcher(["1", "!foo", "!bar"])
+        self.assertTrue(tm.matches([]))
+        self.assertTrue(tm.matches([""]))
+        self.assertTrue(tm.matches(["baz"]))
+        self.assertFalse(tm.matches(["bar"]))
+        self.assertFalse(tm.matches(["foo"]))
+
+    def test_excludeemptyonly_tag(self):
+        tm = TagMatcher(["1", "!"])
+        self.assertTrue(tm.matches([]))
+        self.assertTrue(tm.matches(["bar"]))
+        self.assertTrue(tm.matches(["foo"]))
+        self.assertFalse(tm.matches([""]))
+
+    def test_excludeempty_tag(self):
+        tm = TagMatcher(["1", "!", "foo", "bar"])
+        self.assertTrue(tm.matches([]))
+        self.assertTrue(tm.matches(["bar"]))
+        self.assertTrue(tm.matches(["foo"]))
+        self.assertFalse(tm.matches(["baz"]))
+        self.assertFalse(tm.matches([""]))