Mapping Existing Content Against Real Prompts

Most content audits ask one question: does a page exist covering this topic. That question produces a comforting but misleading answer, because a page can be topically correct and still fail the prompt it was supposed to serve. This sub-chapter introduces a stricter audit, run at two resolutions rather than one: topic-level, whether a page covers the right subject, and fragment-level, whether a specific passage inside that page actually answers the prompt directly. The gap between those two resolutions is where most real coverage problems hide, and closing it is the working deliverable this chapter has been building toward.

The Two Resolutions of Alignment

Content-to-prompt mapping runs at two distinct resolutions, and conflating them is the single most common reason an otherwise careful audit produces a falsely reassuring result.

Topic-level alignment asks whether a page’s overall subject matches an intent cluster. This is the coarse resolution, and it is where most content audits stop. Fragment-level alignment asks something narrower and more demanding: does a specific, identifiable passage within that page directly and completely address the cluster’s underlying need, in a form a reader, or a generative system extracting a fragment, could use without reading the rest of the page. A page can pass the topic-level check completely and fail the fragment-level check entirely, because being about the right subject is not the same as containing an answer to the specific request. This distinction connects directly to the fragment selection mechanics covered in Chapter 3.4, where exactly this gap determines whether a system can actually use a piece of content.

Building the Mapping Document

The mapping document is a structured record connecting each intent cluster from Chapter 7.3 to the strongest existing content candidate for resolving it, an honest alignment strength rating, and a specific description of the gap where one exists.

Each row of the document should carry the cluster, the candidate page or asset, the topic-level assessment, the fragment-level assessment, and a plain-language note describing exactly what is missing when alignment is partial or absent. The fragment-level note is where the real diagnostic value lives: “covers the topic but never states a direct recommendation” is a specific, actionable gap description. “Needs improvement” is not. The document’s usefulness scales directly with how precisely each gap is described, since the next step, prioritization, depends on knowing exactly what has to change, not just that something does.

Why Topic-Level Alignment Alone Is a False Positive

Topic-level alignment alone produces a false sense of coverage, and this false positive is specifically dangerous because it is the version of the audit most teams already know how to run and therefore the one most likely to be trusted without question.

A page titled “Understanding CRM Systems for Small Business” that discusses CRM concepts thoroughly can still fail an evaluative cluster asking “which CRM should a small team actually use,” because thorough conceptual discussion is not the same as a direct, extractable recommendation. Topic-level alignment says the content is in the right neighborhood. Fragment-level alignment says the content actually answers the door when the specific question knocks. A content inventory that only checks the first produces a coverage picture that looks considerably healthier than the coverage a generative system, evaluating fragments rather than page topics, will actually find.

Auditing Existing Content Honestly

Honest auditing includes every page relevant to a cluster, including flagship, high-traffic, or long-standing pages that a team has emotional or historical investment in. These pages are not exempt from failing fragment-level alignment, and in practice they fail it more often than newer content, because they were frequently built under an older content standard, before fragment-level thinking was part of the process.

The audit should resist the pull to assume a page is fine because it performs well in traditional search, or because it represents significant past investment. Traditional search performance measures something different, established directly in Chapter 5.1, and past investment is a sunk cost that has no bearing on whether the page’s current fragments actually resolve a specific prompt cluster today. The honest version of this audit sometimes produces uncomfortable findings about a site’s best-known content. Those findings are exactly the ones worth having.

Turning the Map Into a Prioritized Action List

The finished mapping document is not the end product. It becomes an action list once each gap is paired with a specific next step: build new content, restructure existing content to surface a buried fragment, add a missing element like a direct recommendation or a comparison table, or, in cases of strong existing alignment, leave the asset alone and move attention elsewhere.

Prioritization here follows the same logic established in Chapter 7.4: weigh cluster frequency, business relevance, and competitive gap together, rather than defaulting to whichever fix looks fastest. A restructuring fix on an already-strong page is often faster and higher-leverage than a from-scratch build, which is part of why the fragment-level resolution in this mapping matters: it frequently reveals that the fix needed is smaller and more surgical than “write a new page” would suggest.

Handing the Map to Architecture and Operations

The completed mapping document is the direct input to two later parts of this framework. Chapter 8 uses it to inform architecture decisions: which clusters need a dedicated spoke page, which can be resolved through restructuring an existing asset, and where silo boundaries should sit based on where real coverage strength and gaps actually fall. Chapter 13.2, the mapping phase of the broader operational sequence, treats this document as a starting artifact rather than something built again from scratch.

This handoff is the reason the mapping document needs to be built with genuine rigor rather than treated as a one-time internal exercise. Decisions made downstream, what gets built, in what order, with what priority, inherit whatever accuracy or inaccuracy this map carries forward.

Building the Map That Actually Gets Used

Michael Rubinstein treats the fragment-level resolution as the part of this process teams most often skip under time pressure, precisely because topic-level auditing is faster and feels sufficiently thorough. It isn’t, and the gap between the two resolutions is exactly where a site’s real coverage picture usually turns out to be worse than its content inventory suggests.

ScribePress runs new content plans against this two-resolution mapping standard by default, checking fragment-level alignment specifically before treating a cluster as resolved, since topic-level coverage alone is precisely the false positive this sub-chapter warns against.

Learn more about the work behind this framework at michael-rubinstein.com.

Put the framework to work