diff --git a/website/docs/usage/rule-based-matching.md b/website/docs/usage/rule-based-matching.md
index 13d3dcd32..ee901e3fd 100644
--- a/website/docs/usage/rule-based-matching.md
+++ b/website/docs/usage/rule-based-matching.md
@@ -405,7 +405,7 @@ class BadHTMLMerger(object):
         for match_id, start, end in matches:
             spans.append(doc[start:end])
         with doc.retokenize() as retokenizer:
-            for span in hashtags:
+            for span in spans:
                 retokenizer.merge(span)
                 for token in span:
                     token._.bad_html = True  # Mark token as bad HTML
@@ -679,7 +679,7 @@ for match_id, start, end in matches:
     if doc.vocab.strings[match_id] == "HASHTAG":
         hashtags.append(doc[start:end])
 with doc.retokenize() as retokenizer:
-    for span in spans:
+    for span in hashtags:
         retokenizer.merge(span)
         for token in span:
             token._.is_hashtag = True