diff --git a/pipeline/beam_etl/helpers/parse_xml.py b/pipeline/beam_etl/helpers/parse_xml.py
index 266c4bf..1cab2d8 100644
--- a/pipeline/beam_etl/helpers/parse_xml.py
+++ b/pipeline/beam_etl/helpers/parse_xml.py
@@ -6,6 +6,16 @@ it requires more effort to parse.
 
 First I parse the xml into a dictionary. Then, I standardize the found keys into
 some keys I expect.
+
+Notes:
+Looks like "dimenions" can be "Dimensions (Overall)", "Dimensions" or other
+things like "Assembled Dimensions" or "Piece X Dimensions". But this latter two
+options are incomplete (lack the height), harder to parse and rare enough that
+I'll just drop them
+
+Package Quantity and Number of Pieces are never found together. I will assume
+they refer to the same thing.
+
 """
 
 import logging
@@ -13,6 +23,17 @@ import xml.etree.ElementTree as ET
 
 from typing import Dict
 
+FIELDS_MAPPING = {
+    "Material": "materials",
+    "Package Quantity": "packaging",
+    "Number of Pieces": "packaging",
+    "Dimensions (Overall)": "dimensions",
+    "Dimensions": "dimensions",
+    "Weight": "weight",
+    "TCIN": "tcin",
+    "Origin": "origin",
+}
+
 
 def iter_parse(root: ET.Element) -> Dict[str, str]:
     """Recursively parse the XML tree into a dictionary Each key/value pair is
@@ -39,17 +60,6 @@ def parse_raw_specs(raw_specs: str) -> Dict[str, str]:
     This involves first recursively parsing the XML tree and then renaming
     the key values"""
 
-    fields_mapping = {
-        "Material": "materials",
-        "Package Quantity": "packaging",
-        "Number of Pieces": "packaging",
-        "Dimensions (Overall)": "dimensions",
-        "Dimensions": "dimensions",
-        "Weight": "weight",
-        "TCIN": "tcin",
-        "Origin": "origin",
-    }
-
     try:
         xml_root = ET.fromstring(raw_specs)
     except ET.ParseError:
@@ -58,8 +68,8 @@ def parse_raw_specs(raw_specs: str) -> Dict[str, str]:
 
     parsed = iter_parse(xml_root)
     specs_dict = {
-        fields_mapping[key]: value
+        FIELDS_MAPPING[key]: value
         for key, value in parsed.items()
-        if key in fields_mapping
+        if key in FIELDS_MAPPING
     }
     return specs_dict