diff --git a/qurator/dinglehopper/character_error_rate.py b/qurator/dinglehopper/character_error_rate.py
index f63a15f..05cc931 100644
--- a/qurator/dinglehopper/character_error_rate.py
+++ b/qurator/dinglehopper/character_error_rate.py
@@ -1,21 +1,36 @@
 from __future__ import division
 
 import unicodedata
+from typing import Tuple
 
 from uniseg.graphemecluster import grapheme_clusters
 
 from qurator.dinglehopper.edit_distance import distance
 
 
-def character_error_rate(reference, compared):
-    d = distance(reference, compared)
-    if d == 0:
-        return 0
+def character_error_rate_n(reference, compared) -> Tuple[float, int]:
+    """
+    Compute character error rate.
 
+    :return: character error rate and length of the reference
+    """
+    d = distance(reference, compared)
     n = len(list(grapheme_clusters(unicodedata.normalize('NFC', reference))))
-    if n == 0:
-        return float('inf')
 
-    return d/n
+    if d == 0:
+        return 0, n
+    if n == 0:
+        return float('inf'), n
+    return d/n, n
 
     # XXX Should we really count newlines here?
+
+
+def character_error_rate(reference, compared) -> float:
+    """
+    Compute character error rate.
+
+    :return: character error rate
+    """
+    cer, _ = character_error_rate_n(reference, compared)
+    return cer
diff --git a/qurator/dinglehopper/cli.py b/qurator/dinglehopper/cli.py
index 7f9ea8f..08a36bb 100644
--- a/qurator/dinglehopper/cli.py
+++ b/qurator/dinglehopper/cli.py
@@ -57,8 +57,8 @@ def process(gt, ocr, report_prefix):
     gt_text = substitute_equivalences(gt_text)
     ocr_text = substitute_equivalences(ocr_text)
 
-    cer = character_error_rate(gt_text, ocr_text)
-    wer = word_error_rate(gt_text, ocr_text)
+    cer, n_characters = character_error_rate_n(gt_text, ocr_text)
+    wer, n_words = word_error_rate_n(gt_text, ocr_text)
 
     char_diff_report = gen_diff_report(gt_text, ocr_text, css_prefix='c', joiner='', none='·', align=align)
 
@@ -88,7 +88,8 @@ def process(gt, ocr, report_prefix):
         template = env.get_template(template_fn)
         template.stream(
             gt=gt, ocr=ocr,
-            cer=cer, wer=wer,
+            cer=cer, n_characters=n_characters,
+            wer=wer, n_words=n_words,
             char_diff_report=char_diff_report,
             word_diff_report=word_diff_report
         ).dump(out_fn)
diff --git a/qurator/dinglehopper/templates/report.json.j2 b/qurator/dinglehopper/templates/report.json.j2
index 62d3f77..62a242d 100644
--- a/qurator/dinglehopper/templates/report.json.j2
+++ b/qurator/dinglehopper/templates/report.json.j2
@@ -2,5 +2,7 @@
     "gt": "{{ gt }}",
     "ocr": "{{ ocr }}",
     "cer": {{ cer|json_float }},
-    "wer": {{ wer|json_float }}
+    "wer": {{ wer|json_float }},
+    "n_characters": {{ n_characters }},
+    "n_words": {{ n_words }}
 }
diff --git a/qurator/dinglehopper/word_error_rate.py b/qurator/dinglehopper/word_error_rate.py
index 2425200..7ed56e4 100644
--- a/qurator/dinglehopper/word_error_rate.py
+++ b/qurator/dinglehopper/word_error_rate.py
@@ -1,6 +1,7 @@
 from __future__ import division
 
 import unicodedata
+from typing import Tuple
 
 import uniseg.wordbreak
 
@@ -44,7 +45,7 @@ def words_normalized(s):
     return words(unicodedata.normalize('NFC', s))
 
 
-def word_error_rate(reference, compared):
+def word_error_rate_n(reference, compared) -> Tuple[float, int]:
     if isinstance(reference, str):
         reference_seq = list(words_normalized(reference))
         compared_seq = list(words_normalized(compared))
@@ -53,11 +54,15 @@ def word_error_rate(reference, compared):
         compared_seq = list(compared)
 
     d = levenshtein(reference_seq, compared_seq)
-    if d == 0:
-        return 0
-
     n = len(reference_seq)
+
+    if d == 0:
+        return 0, n
     if n == 0:
-        return float('inf')
+        return float('inf'), n
+    return d / n, n
+
 
-    return d / n
+def word_error_rate(reference, compared) -> float:
+    wer, _ = word_error_rate_n(reference, compared)
+    return wer