SERP Content Gap Analysis Against Ranking Pages
Compare the pages already ranking and name what none of them covers well.
Takes the content of the pages currently ranking for a keyword and returns a coverage matrix, the subtopics every competitor handles, the ones all of them handle badly, and the specific additions that would make a new page the most complete result.
Ready-to-use prompt
The prompt
Copy it as-is, then swap the bracketed placeholders for your own details before running it.
Role: You are an SEO editor running a content gap analysis on a live SERP.
Context:
- Target keyword: {{keyword}}
- Competing pages (paste content or detailed summaries): {{competitor_pages}}
- Our existing page, if any: {{our_page}}
Task: Produce:
1. A coverage matrix — subtopics as rows, competing pages as columns, marked covered / thin / absent
2. Table stakes: subtopics every ranking page covers, which we must include
3. Genuine gaps: subtopics none of them covers, or all cover thinly
4. Depth gaps: subtopics covered everywhere but shallowly, where more specificity wins
5. Format gaps: assets the SERP rewards that no page provides
6. A prioritised list of additions, each with the reason it should win
Rules:
- Work only from the supplied page content. Do not assume what a page contains.
- Separate "nobody covers it" from "nobody should cover it" — flag gaps that exist because the subtopic is off-intent.
- No invented metrics, rankings or traffic figures.
Output: the matrix first, then the prioritised additions.
- Define the threshold for covered versus thin before assigning ratings.
- Attach a page identifier and supporting passage to each disputed or priority-driving rating.
- Identify extraction limitations; do not treat an omitted asset as confirmed absent.
- Distinguish an observed format opportunity from evidence that the SERP rewards that format.
Estimated results
Editor's note
Why this prompt matters
A gap analysis that lists everything competitors mention is just a longer outline. The useful version separates what everybody covers, what everybody covers badly, and what nobody covers for a good reason. This prompt builds an explicit coverage matrix from the pages you supply, then forces the model to justify each proposed addition rather than padding the brief.
Anatomy
Prompt engineering breakdown
Role
Context
Goal
Constraints
Output format
What you'll get
Expected output
Coverage matrix
Illustrative scenario: the keyword is “website redesign brief template.” Five supplied page extracts represent competing results: A, an agency guide; B, a downloadable brief; C, a project checklist; D, a procurement guide; and E, a planning article. These are generic examples, not a live ranking report. No existing page was supplied.
“Covered” means usable instructions or examples; “thin” means a mention without enough detail to act. The assessments apply only to these extracts.
| Subtopic | A | B | C | D | E | |---|---|---|---|---|---| | Business goals | Covered | Covered | Covered | Covered | Covered | | Audience definition | Covered | Covered | Covered | Covered | Covered | | Budget and timeline | Covered | Covered | Covered | Covered | Covered | | Page and feature scope | Covered | Thin | Covered | Covered | Thin | | Content migration ownership | Thin | Absent | Thin | Absent | Absent | | Measurable acceptance criteria | Thin | Thin | Thin | Thin | Thin | | Approval responsibilities | Covered | Thin | Covered | Covered | Thin | | Completed example brief | Absent | Absent | Absent | Absent | Absent |
Prioritised additions
- Add a completed brief beside the blank template. All five extracts explain what to request, but none demonstrates a coherent finished brief. Use one explicitly fictional project throughout, connecting its business goal to scope, constraints, and acceptance criteria. This addresses the template-seeking intent more directly than adding another introductory explanation.
- Turn acceptance criteria into observable checks. Every extract mentions success without showing how approval works. Pair each requirement with a verification method and named role: for example, “The enquiry form routes submissions to the agreed inbox; the project owner verifies receipt before launch.” This is a depth gap, not an absent topic.
- Add a content migration responsibility table. Include content group, retain/rewrite/remove decision, owner, and approval deadline. The extracts leave this work largely undefined. Keep it at briefing level; detailed migration engineering would distract from the task.
- Preserve the essentials before expanding. Goals, audience, budget, and timeline are table stakes because every extract covers them substantively. More words on these topics are lower priority than filling the execution gaps above.
Format opportunity and intent boundary
A copyable completed brief is a supported format opportunity: none appears in the supplied extracts. Its absence does not establish that the SERP rewards that format; ranking or engagement evidence would be needed for that claim.
Exclude platform-specific development tutorials. Their absence is not a useful gap for someone preparing a redesign brief. The strongest page would help readers commission and approve work, not teach them to build the website.
Under the hood
Why this prompt works
Grounding the analysis in pasted content removes the model's biggest failure mode here, which is confidently describing pages it has never seen. The matrix makes coverage comparable at a glance, and splitting genuine gaps from off-intent gaps stops teams from adding sections that dilute the page's match to the query.
Model fit
Best AI models for this prompt
Claude
Best at holding several long competitor pages in context without blending them together.
Gemini
Strong at spotting format gaps when you describe the SERP features present.
ChatGPT
Quickest for the matrix itself; summarise longer pages before pasting.
When to use
- Before outlining a page for a competitive keyword.
- When a page sits outside the top ten and you need to test coverage as one possible weakness.
- During topic selection, to identify a useful contribution before committing production effort.
- Before briefing a rewrite, to separate missing sections from shallow existing ones.
When not to use
- Without competing page content or sufficiently detailed extracts.
- For new topics without a meaningful comparison set.
- For technical or site-level diagnosis; use crawl and indexing data.
- To validate specialist accuracy; have a qualified reviewer assess proposed additions.
- To establish current rankings or demand; consult live search results and keyword data.
Get more from it
Pro tips
- 1
Paste five to eight ranking pages; label each with a stable identifier and distinguish full text from summaries.
- 2
Include your own page under a separate label so additions become a rewrite brief, not competitor findings.
- 3
Treat off-intent gaps as exclusions; ask which reader decision each proposed addition supports.
- 4
After publishing, rerun with the revised page to verify coverage changes, not ranking gains.
- 5
Preserve tables, template fields, and captions when extracting pages; missing assets can create false format gaps.
- 6
For a second pass, challenge every “absent” finding against the source extracts and request supporting passages for disputed ratings.
Don't ship this
Common mistakes
✗ Letting the model describe pages you did not supply.
Fix — State that only pasted content may be used and check the matrix against it.
✗ Treating every absent subtopic as an opportunity.
Fix — Require an off-intent flag and drop those rows.
✗ Analysing only the number one result.
Fix — Paste at least five ranking pages; single-page analysis copies one editor's choices.
People also ask
Frequently asked questions
Q.Can the model fetch the ranking pages itself?
Do not rely on it. Paste the content or detailed summaries — grounding in supplied text is what keeps the matrix honest.
Q.How many competitors should I include?
Five to eight. Fewer and you copy one editor's structure; more and the matrix becomes unreadable.
Q.Is this the same as a keyword gap analysis?
No. A keyword gap compares which terms sites rank for. This compares what the ranking pages actually say about one query.