Heading and Schema Fix: A 5-Check Audit

Short answer: one H1, clean levels, and valid FAQ schema cleared every branch. The service page shipped two H1 tags, jumped from H1 to H3 twice, and carried no structured data. We fixed the outline with the 5-check heading audit and added FAQ markup with the Schema Generator. All branches passed on re-audit.

No client is behind this study. No traffic numbers exist here. The dataset is our own tooling: the heading checker with its 5 focused checks, the schema generator with 7 supported types, and the 45-check page analyzer. Before and after values below are an illustrated fixture, clearly marked, so you can repeat the same pass on your own pages.

Key takeaways

• One page gets one H1. Two H1 tags split the topic signal. The fix keeps the page title as the single H1.
• Levels must step by one. H1 to H3 jumps confuse readers and parsers alike. Fill the missing H2 and the outline reads clean.
• Schema must validate. One bad block can suppress the whole rich result. Test every block before you ship it.
• Headings feed AI answers. Clear H2 questions with short direct answers are easy to quote and easy to cite.
• Re-audit the same page. The verdict is the same tool on the same URL, not a memory of what you changed.

Why headings and schema fail together

Both describe page structure. Headings describe it for readers. Schema describes it for machines. When both are wrong, neither audience gets a clean outline. Our heading tool runs 5 focused checks: exactly one H1, no skipped levels, a logical sequence, no empty headings, and no duplicate heading text.

A page outline is a table of contents: one title, ordered sections, and nothing the reader must guess.

The service page in this study sold water-tank cleaning for housing societies. The template printed the brand line as an H1 and the service title as a second H1. Two section blocks used H3 with no parent H2. The FAQ section below answered real questions but shipped as plain text, so search engines saw questions with no structured answers.

The method we used, step by step

We ran the same five steps any site owner can repeat in about thirty minutes.

Step one: audit the headings. We pasted the page HTML into the Heading Hierarchy Checker. It runs its 5 focused checks and reports each one. Result: fail on single H1, fail on level skips, pass on empty headings, pass on duplicates, fail on sequence. Three fails out of five.

Step two: score the whole page. We ran the URL through the SEO Analyzer. Its 45 checks confirmed the same story from the other side: the content group flagged the duplicate H1 and the head group found no structured data worth verifying.

Step three: rebuild the outline. We demoted the brand line to plain text and kept the service title as the single H1. We converted the two stray H3 blocks into H2 sections with H3 steps beneath them. No words changed, only levels. The heading hierarchy guide shows the same single-H1 rule applied across page types.

Step four: add FAQ schema. We built an FAQPage block in the schema generator for the four questions already on the page. Each answer stayed between 45 and 70 words, complete on its own. We validated the block, pasted it once, and re-tested until clean. The schema markup guide lists the required fields per type so nothing ships half-formed.

Step five: re-run both audits. Same HTML, same URL, same tools. Heading checks: 5 passes out of 5. Page audit: the heading and schema branches moved to pass, and the other checks held steady.

Before and after: the illustrated fixture

Every number below is illustrated, not measured from a client site.

Before. H1 count: two. Level skips: two jumps from H1 straight to H3. Empty headings: none. Duplicate heading text: none. Schema blocks: zero, so the schema branch reported nothing to verify. Heading result: two passes out of 5 checks.

After. H1 count: one. Level skips: zero, with each H3 nested under a real H2. Empty and duplicate headings: still none. Schema blocks: one valid FAQPage block with four questions. Heading result: 5 passes out of 5 checks. Page audit: heading and schema branches pass, remaining checks unchanged.

What did not change: copy, images, links, and speed. Only levels and markup moved. That isolation lets the before and after mean something. One variable changed, the same tools re-ran, and the report shows exactly which branches flipped.

How to measure the real outcome in Search Console

Structure fixes show up in two places. First, eligibility: valid FAQ markup makes the page eligible for rich results it could not earn before. Second, matching: a clean outline helps the page match more specific queries. Our MCP server exposes 23 Search Console tools, including per-query clicks and impressions by page.

Pull four weeks before the fix and four weeks after. Watch impressions first: new long-tail queries are the earliest sign the outline reads better. Watch clicks next: rich-result eligibility should lift the share of those impressions. Track the page against its own past and note the date of the fix, since recrawl timing decides when the window starts.

Do not claim a rich-result win without proof. Check the actual results page for your main queries and confirm the FAQ display appears. Eligibility is not display. Report both: the audit result and the observed display.

What this method cannot prove

Clean headings cannot prove a ranking gain on their own. They remove a handicap; relevance and links still decide position. Pages with thin copy stay thin under a perfect outline.

The fixture also covers one template on one page. Real sites repeat the same template fault across hundreds of URLs. The method scales through the template, not through page-by-page edits. Fix the template, then re-audit a sample of ten pages before you report the win site-wide.

The bottom line

Structure is cheap and its absence is loud. One H1, stepped levels, and one valid schema block cleared every branch in this study. Audit your own template, fix it once, and re-check ten pages before you move on.