Comparison Table Presence

This parameter checks whether a COMPARISON or REVIEW page contains a real comparison table. friendly4AI scores 100 when it finds at least one qualifying <table> — two or more columns, two or more data rows, and two or more columns whose cells hold two or more distinct values. Anything less scores 0. Every other content type is skipped, so non-comparison pages are never penalised.

What does this parameter measure?

friendly4AI resolves the page's content type first. It only inspects tables on COMPARISON and REVIEW pages, where it scans each <table> element for a qualifying structure:

• at least two columns,
• at least two data rows, and
• at least two columns whose cells hold two or more distinct values.

The last test is the important one. It screens out layout tables and plain key-value tables, neither of which is a genuine comparison. On any other content type — article, tutorial, glossary, product, or unknown — the parameter does not apply and the page is skipped. No comparison claim, no penalty.

Why does a comparison table matter for AI-readiness?

Ask an AI engine "X vs Y" or "best tool for Z" and it wants side-by-side facts it can lift straight off the page. A well-formed comparison table gives it exactly that. Structured rows and columns extract far more reliably than the same facts scattered through paragraphs.

Make a comparison claim without a table and you push the work back onto the model: it has to reconstruct the comparison itself, and more often than not it skips yours and cites a competitor's table instead. On comparison and review pages, a qualifying table is one of the strongest signals you can ship for citation in comparison-style answers. It works alongside structured data and content-type schema match, which help engines classify the page correctly to begin with.

How is it scored?

On the pages it applies to, the v4.5 methodology scores this Content Structure parameter as binary:

• Pass (100): one or more qualifying comparison tables are found.
• Partial: a table exists but does not qualify — it is layout-only, a key-value pair, or short of the two-diverse-column and two-data-row thresholds — so it still scores 0 against the binary check. This band describes the realistic near miss. There is no native middle tier, because the underlying check is presence-or-absence.
• Fail (0): no qualifying table is present.

Pages outside the comparison and review content types are skipped entirely and left out of scoring.

How do you fix a low score?

• On a comparison or review page, add a real comparison table with at least two columns and at least two data rows.
• Check that at least two columns carry diverse values across rows. A column where every cell repeats does not count toward a qualifying table.
• Replace layout tables and key-value tables with a genuine side-by-side comparison of the options or features.
• Confirm the page reads as a comparison or review — clear vs, compare, or review signals — so the table gets scored instead of skipped.
• Re-scan and check the qualifyingTables, largestTableColumns, and contentTypeHint evidence fields.