Start here
Before You Fix It: What This Check Means
Chunkability measures whether sections can be extracted as coherent units without losing meaning. In plain terms, this tells you whether AI crawlers and answer systems can understand and reuse your content correctly. Scavo simulates heading-aware chunking from extracted page sections and scores chunk quality.
Why this matters in practice: unclear machine-facing signals can reduce retrieval quality and citation consistency.
How to use this result: treat this as directional evidence, not final truth. Answer-engine retrieval behavior can shift over time even when your technical setup is stable. First, confirm the issue in live output: verify bot-facing output and policy files on the final URL Then ship one controlled change: Ensure one logical `H1` and coherent `H2/H3` tree. Finally, re-scan the same URL to confirm the result improves.
Background sources
TL;DR: Your content is one long narrative without clear sections, making it harder for AI to extract and summarise key information.
44.2% of AI citations come from the first 30% of content (Profound), so front-loading key facts matters. AI models work better with structured, chunked content — clear headers, concise paragraphs, fact boxes, and attributed claims. Walls of unstructured text force AI to guess at relevance, reducing your chances of being cited or recommended in AI-generated responses.
What Scavo checks (plain English)
Scavo simulates heading-aware chunking from extracted page sections and scores chunk quality.
Current simulation targets:
- Target chunk size: ~
380tokens - Minimum useful chunk:
140tokens - Oversized threshold:
700tokens
Scavo computes:
- Number of chunks and their sizes
- Oversized chunks
- Undersized chunks
- Orphan chunks (without heading anchors)
- Heading coverage percentage
- Largest chunk and average chunk size
How Scavo scores this check
Result behavior:
- Fail: score <
55, or 2+ oversized chunks - Warning: score <
78, or any orphan chunk, or more than 1 undersized chunk - Pass: chunk shape is stable and well-anchored
- Info: no usable HTML/section extraction
In your scan report, this appears under What failed / What needs attention / What is working for ai_chunkability, followed by Recommended next steps and Technical evidence (for developers) when needed.
- Scan key:
ai_chunkability - Category:
AI_VISIBILITY
Why fixing this matters
Better chunk structure improves extraction precision for AI answers, internal search, and summarization systems. Long unbroken sections or weak heading hierarchy increase ambiguity, especially when users ask targeted questions.
This also improves human readability and editorial quality. Chunkability and good information architecture usually move together.
Common reasons this check flags
- Huge page sections under one heading.
- Heading hierarchy skips structure (
H1then scatteredH4s). - Repeated tiny sections with no substance.
- Content lives in generic div blocks without meaningful headings.
If you are not technical
- Review page structure like a reader: can each section be summarized in one sentence?
- Ask content owners to break long blocks into clear sub-sections.
- Require meaningful headings, not decorative labels.
- Re-scan after edits and compare chunk score trend.
Technical handoff message
Copy and share this with your developer.
Scavo flagged AI Chunkability (ai_chunkability). Please improve heading-anchored section structure, reduce oversized/orphan chunks, and keep section lengths in a stable range so retrieval systems can extract precise context.If you are technical
- Ensure one logical
H1and coherentH2/H3tree. - Break oversized sections into smaller topical blocks.
- Merge ultra-thin sections that do not stand alone.
- Keep key facts near section headings for strong anchor context.
- Avoid injecting primary content late via client-side scripts only.
How to verify
- Inspect section map and heading hierarchy in rendered HTML.
- Check average and largest chunk sizes after edits.
- Confirm orphan chunk count is zero.
- Re-run Scavo and verify readiness score improves.
What this scan cannot confirm
- It does not evaluate factual correctness of content.
- It does not guarantee citation by any model/vendor.
- It is a structural simulation, not a full model inference test.
Owner checklist
- [ ] Assign one owner for heading and long-form page structure.
- [ ] Add editorial QA for heading hierarchy before publish.
- [ ] Keep template components that enforce predictable section patterns.
- [ ] Re-check chunkability after major content refreshes.
FAQ
Is this just an SEO heading check?
No. It focuses on chunk shape and retrieval-friendly structure, not only heading presence.
Can short pages fail chunkability?
They can warn as thin/orphaned if structure is unclear or fragmented.
Do we need exact token targets?
Treat thresholds as practical guidance, not law. The goal is stable, meaningful sections.
What should we fix first?
Start with oversized sections and missing heading anchors, then refine thin fragments.
Sources
- W3C WAI: Page structure and headings
- MDN: HTML heading elements
- Google Search Central: Create helpful, people-first content
- OpenAI Docs: Prompt and context optimization
Need a sectioning blueprint for your docs/blog/help templates? Send support one sample page and we can map a chunk-friendly outline.
