← Back to assessments

SH010 – Spec Label Normalization

Overview

Checks specification tables for consistent labeling: same concept should use same casing and terminology (e.g. Weight vs weight, or weight vs mass).

Probability that AI systems use this signal: Crawlers and generative AI pipelines consume structured and visible page signals for training and inference. Assessment-specific signals (e.g. schema, canonicals, trust cues) affect how likely a page is indexed and surfaced. The probability that this assessment's signal influences AI behavior is high when the page is in a product or compliance context.

Impact on geo compliance: Passing this assessment supports geo compliance by ensuring machine-readable and visible content meet standards that reduce the risk of wrong locale, pricing, or trust in AI-generated answers. Failing can lead to non-compliant or misleading surfacing.

What We Check

  • **_find_specification_tables:** tables where _is_specification_table: class/id/text contains specification/spec/details/features/technical/dimensions/weight/capacity/storage/battery, or table has ≥2 th and ≥2 td. No spec tables: pass ("not applicable").
  • **_analyze_spec_labels:** all th text from spec tables. **Case:** group by lowercase; if multiple different casings for same label → "Inconsistent case for 'x': a, b". **Common inconsistencies:** weight/Weight/WEIGHT/mass/Mass; dimensions/size; storage/capacity; battery/power — if >1 variation found in labels, issue. −15 per issue.

Pass / Fail and Score

  • **PASSED** if score >= 60.

How to Fix When It Fails

  • Standardize specification labels (consistent capitalization and terms across tables).

Common Issues

  • Same attribute with different casing (e.g. "Weight" vs "weight"); mixed synonyms (e.g. weight vs mass, storage vs capacity).

Dependencies

None.

How to Verify

  • Review all spec table th labels for consistency.

Additional Resources

  • Accessibility: consistent labels; i18n/l10n