# Prompt to page mapping worksheet

Map one registered buyer question to its required answer facets and intended canonical page. Retain visible evidence for each facet, then decide whether to preserve the page, repair a supported gap or investigate ownership.

Worksheet version 1.0 · 21 September 2026. Complete and repeat records as needed. Blank fields request your evidence; this manual file does not collect answers, import records or establish search demand.

Complete guide: https://rankecho.io/resources/find/prompt-gap-analysis

## 1 Register the question

Use the prompt portfolio ID as the join key. Freeze the question and review rubric before inspecting page coverage.

**Prompt identity and context** — Record panel version, prompt ID, verbatim wording, audience, buyer stage, locale and language.

[Enter your response]

**Research and eligibility** — Retain source type, source pointer, collection and registration dates, inclusion rationale and exclusion rule. Generated questions remain candidates until reviewed.

[Enter your response]

**Required answer facets** — List each distinct facet and why it is needed for this buyer decision. Give each a stable ID and a supporting source or product rationale.

[Enter your response]

## 2 Map each facet to evidence

Repeat this record for each required facet and candidate page. Inspect visible content and canonical evidence; a title, URL or schema statement alone cannot establish coverage.

**Facet and supporting record** — Record prompt ID, facet ID, exact required fact, rationale and retrievable evidence pointer. Keep missing support explicit.

[Enter your response]

**Page and canonical ownership** — Record candidate URL, intended canonical owner, saved page version, redirect chain and canonical-resolution evidence. Retain all competing candidates when ownership is duplicate or unresolved.

[Enter your response]

**Visible before and after locators** — Record the exact heading, paragraph, table cell or other locator before review and after any executed change. Keep absent facets and unexecuted after checks blank with a reason; count repeated wording once.

[Enter your response]

## 3 Classify the coverage

Use the registered question and facets to assign covered, partial, duplicate, missing or excluded. Duplicate ownership is a separate decision even when both pages contain the facts.

**State and reason** — Record the state, supported and required facet counts, and the evidence for your decision. Keep excluded or invalid prompts outside any aggregate.

[Enter your response]

**Coverage reconciliation** — Retain every prompt row. Report raw counts and denominator; a facet total is descriptive and is not importance-weighted, AI visibility or the RankEcho product score.

[Enter your response]

**Independent review** — Name the reviewer, review date and disagreements. Confirm that the state can be reproduced from saved page versions and locators.

[Enter your response]

| Fictional guide row | Before support | Decision |
| --- | --- | --- |
| G01 category | 2/2 | Covered |
| G02 calibration | 1/2 | Partial |
| G03 protocols | 2/2 | Duplicate |
| G04 outdoor use | 0/1 | Missing |

## 4 Prepare the bounded handoff

The fictional guide has 5/7 supported facets before the change. Only G02 proceeds to a bounded repair; G03 needs an ownership decision and G04 needs evidence and page planning.

**Supported action** — Choose preserve, repair, ownership review, evidence research or exclude. Name the target page and exact missing facet when proposing a repair.

[Enter your response]

**Acceptance and rollback** — Record support boundary, proposed semantic change, acceptance check, competing explanations and rollback. Preserve unrelated facts and pages.

[Enter your response]

**Responsible owner and next observation** — Assign the change or investigation owner and due date. Recheck visible coverage after execution. Collect AI answers separately; a repaired facet does not prove a citation, visit or sale.

[Enter your response]

## Complete guide

Prompt gap analysis maps a registered buyer question and its required answer facets to the canonical pages that visibly address it. Classify each prompt as covered, partial, missing, or duplicate under a frozen rubric; preserve the exact prompt, source, audience, candidate URLs, and content locators; then hand one bounded, repairable gap to Fix. A generated prompt idea, a provider-rewritten query, or a sampled grounding query is not measured search demand, and an absent AI answer is not by itself a content-coverage defect. Publication is complete when the mapping and handoff validate; site traffic is a separate outcome and is not required.

Original guide update: 2026-09-07. Worksheet prepared 21 September 2026. The original guide and its evidence records retain their dates. The worksheet is an original manual form, not a provider-approved measurement standard.

## What is the scope of prompt gap analysis?

The unit is a registered buyer question and user job, its required answer facets, and an owned-page mapping. The analysis asks whether one intended canonical page visibly provides the required information; it does not infer a gap from a provider response or a sitewide score.

Keep the original prompt, any provider-derived search, and the observed answer separate. Query fan-out, query rewriting, sampled grounding phrases, keyword estimates, and traffic each describe different objects and do not establish a content-coverage defect.


## Which prompt inputs must be frozen first?

Register the panel and rubric before inspecting pages so convenient wording cannot redefine the gap.

The linked prompt portfolio template registers prompt identity and scope, but it does not carry page-and-facet mappings, content locators, page versions, facet rationales, or evidence pointers. Attach a companion mapping record keyed by prompt ID before classifying coverage.

- Prompt ID, verbatim prompt, source and collection date, audience, buyer stage, locale or language, inclusion reason, and exclusion rule.
- Required answer facets with a source or product rationale for each one.
- Candidate canonical URL, visible page locator, ownership decision, saved page version, and evidence pointer.
- Generated prompt ideas may enter as candidates only; they are not measured demand or observed gaps.
- No prompt count, aggregate coverage threshold, or page-per-query rule is universal.
- [Record the registered prompt portfolio](https://rankecho.io/templates/prompt-portfolio)

## What must the companion prompt mapping record contain?

Keep the template's prompt ID as the join key. For every required facet, retain the registration source, inclusion rationale, evidence pointer, before and after visible locators, intended canonical path, page version, and canonical-resolution evidence.

The table below is the synthetic companion mapping for G01–G04. Its synthetic-evidence strings are artifact identifiers, not live URLs or external proof.

| Prompt and facet | Registration and rationale | Before locator | After locator | Evidence pointer |
| --- | --- | --- | --- | --- |
| G01 · product category | product-research · synthetic-research/R01 · The answer must identify what S1 is. | /products/s1#overview · paragraph 1 | /products/s1#overview · paragraph 1 | synthetic-evidence/G01/F1 |
| G01 · intended use | product-research · synthetic-research/R01 · The answer must state the intended operating use. | /products/s1#overview · paragraph 2 | /products/s1#overview · paragraph 2 | synthetic-evidence/G01/F2 |
| G02 · 12-month interval under standard conditions | support-record · synthetic-support/S14 · The decision requires the ordinary schedule and its condition. | /support/s1-calibration#schedule · paragraph 2 | /support/s1-calibration#how-often · sentence 1 | synthetic-evidence/G02/F1 |
| G02 · after-shock exception | support-record · synthetic-support/S14 · The exception changes when recalibration may be required. | No supported before locator | /support/s1-calibration#how-often · sentence 2 | synthetic-evidence/G02/F2 |
| G03 · Modbus TCP | sales-record · synthetic-sales/S22 · The integration decision requires an explicit protocol name. | /products/s1#protocols and /support/s1-integrations#supported | /products/s1#protocols and /support/s1-integrations#supported | synthetic-evidence/G03/F1 |
| G03 · OPC UA | sales-record · synthetic-sales/S22 · The integration decision requires an explicit protocol name. | /products/s1#protocols and /support/s1-integrations#supported | /products/s1#protocols and /support/s1-integrations#supported | synthetic-evidence/G03/F2 |
| G04 · documented operating range or ingress condition | sales-record · synthetic-sales/S31 · A supported suitability answer needs an applicable environmental limit. | No supported before locator | No supported after locator | synthetic-evidence/G04/F1-missing |


## How are prompts mapped to canonical pages?

Inspect rendered visible content rather than infer coverage from a title, schema object, or URL. Record the exact heading, table, or paragraph locator for each supported facet, follow redirects, and confirm the intended canonical owner.

Count a facet once even when the same wording repeats. When several pages claim the same user job, retain duplicate ownership as its own state rather than treating repetition as stronger coverage.

Freeze that owner before inspecting an AI answer. If ownership is duplicate or unresolved, resolve it here rather than retrofitting an owner and calling a later observed URL the wrong page.

- [Classify an observed owned-URL mismatch](https://rankecho.io/resources/find/issues/wrong-page-cited)

## Which coverage states must stay separate?

Assign one state from the frozen page-and-facet evidence. Excluded or invalid prompts remain outside any aggregate.

| State | Rule | Next decision |
| --- | --- | --- |
| Covered | Exactly one intended canonical owner visibly supports every required facet | Preserve; do not invent a fix |
| Partial | A candidate supports at least one but not all required facets | Name the missing facet and inspect a bounded repair |
| Missing | No candidate supplies the required answer | Resolve evidence and page ownership before creating anything |
| Duplicate | More than one page claims the same prompt job | Choose or consolidate canonical ownership |
| Excluded | The prompt is invalid, out of scope, or unregistered | Report separately; never move it silently into a denominator |


## What does the worked synthetic prompt matrix show?

SYNTHETIC EXAMPLE — Northstar S1, its pages, sources, surfaces, and observations are fictional; this is not RankEcho data, customer data, provider evidence, or a benchmark.

The four registered prompts produce one covered, one partial, one duplicate, and one missing row. Five of seven required facets are visibly supported before the fix: 5/7 = 71.4%. That aggregate is descriptive, not importance-weighted and not a RankEcho product score.

| Prompt and user job | Required facets | Candidate page or pages | Before support | Coverage state |
| --- | --- | --- | --- | --- |
| G01 · What is Northstar S1 used for? · category understanding | product category; intended use | /products/s1 | 2/2 | covered |
| G02 · How often should Northstar S1 be calibrated? · maintenance decision | 12-month interval under standard conditions; after-shock exception | /support/s1-calibration | 1/2 | partial · selected for Fix |
| G03 · Which control protocols does Northstar S1 support? · integration requirements | Modbus TCP; OPC UA | /products/s1; /support/s1-integrations | 2/2 | duplicate |
| G04 · Can Northstar S1 be used outdoors? · suitability decision | documented operating range or ingress condition | No qualifying candidate | 0/1 | missing |


## How is one gap selected for Fix?

A fix handoff needs an existing target page, exact missing facet, support boundary, proposed semantic delta, acceptance test, competing explanations, and rollback. G02 qualifies because its canonical owner and standard interval are established while the material shock exception is missing.

G03 still needs a canonical-ownership decision and G04 still needs evidence and page-ownership work. Neither is silently folded into the G02 edit.

- [Build the selected direct answer](https://rankecho.io/resources/fix/direct-answer-block)

## Which false positives and overlap risks should stop the diagnosis?

Stop when the claimed gap is only a generated prompt, provider rewrite, sampled grounding phrase, keyword estimate, missing answer, title mismatch, schema-only statement, or traffic trend. None demonstrates missing visible page coverage on its own.

The Prompt Gap Finder generates starter ideas; AI Citation Gap Analysis classifies target-owned source absence in observed answer cells; Prompt Intelligence describes the product capability. This guide owns the separate manual prompt-to-page-and-facet mapping.

- [Generate starter prompt ideas](https://rankecho.io/tools/prompt-gap-finder)
- [Analyze an observed owned-link gap](https://rankecho.io/resources/find/ai-citation-gap-analysis)
- [Review product prompt intelligence](https://rankecho.io/product/prompt-intelligence)

## Which deterministic validation checks complete Find?

Complete Find from the registered mapping and exact handoff. Do not wait for a citation, ranking, referral, impression, or traffic change.

| Check | Pass condition | Retained evidence |
| --- | --- | --- |
| Prompt identity | Every prompt has a unique ID, exact wording, source, audience, and inclusion rule | Versioned prompt record |
| Facet basis | Every required facet has a source or product rationale | Facet definition and locator |
| Visible support | Every supported facet has a reproducible rendered-content locator | Saved page and capture |
| Page ownership | Redirect and canonical resolution reproduce the declared candidates | Resolved URLs and canonical evidence |
| Classification | Row states and 5/7 arithmetic recompute from the records | Machine-readable matrix |
| Handoff | One target, missing facet, delta, alternatives, acceptance rule, and rollback are exact | G02 handoff record |


## How does prompt-gap analysis hand off to Prove?

Carry G02, both required facets, the exact target page, saved before version, source boundaries, and V01–V06 cell definitions into the bounded change and later repeat record.

The Find classification is complete before later observations exist. Site traffic is a separate outcome and is not required for the handoff or publication.

- [Continue to the direct-answer fix](https://rankecho.io/resources/fix/direct-answer-block)
- [Review the answer-volatility method](https://rankecho.io/resources/prove/answer-volatility)

## Who maintains this prompt-gap guide?

Written and maintained by Abiot Y. Derbie. Published 2026-09-07; sources checked 2026-09-07; last updated 2026-09-07. The worked records are synthetic.

Recheck provider descriptions, prompt provenance, canonical mappings, and visible-content locators before applying the method. Send material corrections through the contact page.

- [Send a correction](https://rankecho.io/contact)
- [Review measurement definitions](https://rankecho.io/methodology)

## Is a generated prompt automatically a prompt gap?

No. It is a candidate until it is registered with a user job and mapped to visible content under a declared facet rubric.

## Is a sampled grounding query measured search demand?

No. Bing describes grounding-query phrases as a sample of citation activity. They are not a complete prompt inventory or a search-volume measure.

## Does an absent AI answer prove the page lacks coverage?

No. Inspect the intended canonical page and required visible facets directly; provider availability, context, retrieval, and answer composition remain separate.

## Should duplicate pages count twice?

No. Record duplicate ownership and decide which page should own the user job instead of inflating supported-facet coverage.

## Does low traffic block prompt gap analysis?

No. Site traffic is a separate outcome and is not required to classify the registered page-and-facet evidence.

## Preserved dated evidence

- [Google Search: AI features and your website](https://developers.google.com/search/docs/appearance/ai-features) — checked 2026-09-07. Google says AI Overviews and AI Mode may use query fan-out, may use different models and techniques, and can show varying responses and links; AI Overviews do not trigger for every query, and eligibility does not guarantee crawling, indexing, or serving. Google Search documentation last updated December 10, 2025 Primary-source documentation check; this does not expose all derived queries, quantify prompt demand, prescribe one page per fan-out query, explain a particular result, or generalize to another provider. Confidence: High

- [OpenAI Help: searching the web with ChatGPT](https://help.openai.com/en/articles/9237897-chatgpt-search) — checked 2026-09-07. OpenAI says ChatGPT Search may rewrite a user question into one or more targeted partner queries; location and enabled memory can affect that process; web-search responses may include citations, which can be incomplete, outdated, or incorrect. Current consumer ChatGPT Search guidance Primary-source documentation check; the guidance does not expose every rewritten query, define a ranking formula, guarantee placement, or describe API and non-Search surfaces universally. Confidence: High

- [Bing Webmaster Blog: AI Performance](https://blogs.bing.com/webmaster/February-2026/Introducing-AI-Performance-in-Bing-Webmaster-Tools-Public-Preview) — checked 2026-09-07. Bing's public-preview AI Performance report exposes citations, cited URLs, trends, and sampled grounding-query phrases; Microsoft says those aggregates do not indicate placement, ranking, authority, page importance, or page role in a particular answer. Public preview announced February 10, 2026 Primary-source documentation check; a sampled grounding phrase is neither a complete prompt inventory nor measured search demand and is not a manual prompt-coverage state. Confidence: High

## Related reading

- [Find resource selector](https://rankecho.io/resources/find)
- [Prompt portfolio template](https://rankecho.io/templates/prompt-portfolio)
- [Generate starter prompt ideas](https://rankecho.io/tools/prompt-gap-finder)
- [Analyze an observed owned-link gap](https://rankecho.io/resources/find/ai-citation-gap-analysis)
- [Review product prompt intelligence](https://rankecho.io/product/prompt-intelligence)
- [Build the selected direct answer](https://rankecho.io/resources/fix/direct-answer-block)

## Product next step

The Prompt Gap Finder provides starter ideas. Review and register them before using this manual mapping; Prompt Intelligence is a separate recurring workflow and does not automatically import this file. https://rankecho.io/product/prompt-intelligence