Define what a good prospect must help a reader do
A backlink prospect is a possible page or publisher to evaluate, not a promised link. Start with the audience, the useful destination you can offer and the specific reason it belongs on that page. “Any high-authority website” is not a reader brief. Google’s people-first content guidance asks whether content provides original value for its intended audience; use that as a question to examine, not a badge that qualifies a site automatically.
The worksheet uses an original fictional brief: help beginner urban gardeners collect a usable soil sample, using an existing English guide with sampling diagrams. The desired opportunity is a relevant resource-page reference. Changing the brief can change which prospects fit, so keep the brief with the exported decisions.
Collect candidate pages with their discovery evidence
Search for resources serving that audience, follow useful references from relevant pages, and review permitted backlink exports for possible sources. Record the exact candidate page, discovery reference and evidence date. A competitor’s referring domain can suggest a place to investigate; it does not establish that its audience, page context or submission process suits your proposal.
Backlink data for prospecting is an input with coverage limits. Google’s Links report is a sample, includes links found over time and may retain pages or links that no longer exist. Keep an export date separate from a current page observation. Preserve missing metrics as unknown rather than replacing them with zero. This worksheet uses reserved example domains and makes no network requests.
Deduplicate pages without erasing provenance
Keep the original discovery URL and a separate grouping key. The worked example removes fragments and five familiar utm tracking parameters for grouping. It retains other query parameters, path case, trailing-slash, protocol and subdomain differences. That is a conservative local rule, not a claim about Google’s canonical URL or proof that two live responses are identical.
The second record points to the first resource page through a different discovery route. It remains in the file as a duplicate of the first record, preserving both sources without creating two outreach targets. Different pages on one domain remain separate candidates until someone checks their purpose. Confirm ambiguous URL variants before merging them in a real dataset.
Use four questions and keep unknowns visible
Ask whether the audience fits, whether your resource adds useful information, whether current page evidence is available and whether an appropriate contact route is known. Record a short reason and the next check. A contact address alone does not answer the other three questions or establish permission to send a message. A dated observation should point to evidence someone else can inspect.
In this teaching rule, duplicates stay duplicates. For other rows, any No means unsuitable for this brief; otherwise any Unknown holds the record for evidence. Four Yes answers place it on a planning shortlist. The result is not a quality score or publisher approval. Editing the judgments changes the plan while the recorded example evidence stays fixed; it does not verify new facts.
Explain exclusions as carefully as the shortlist
The resource-page example has a clear audience and a specific missing explanation, so it starts on the shortlist. The industrial procurement page is excluded for audience mismatch, not declared a bad website. The already-covered page is excluded because the offered guide adds no distinct improvement in this example. A different useful resource could lead to a different judgment.
The older export stays on hold until someone checks the current page and contact route. The editorial page also stays on hold because the contact route is unknown. Neither case is proven unsuitable by missing information. Keep the rejected, duplicate and held records so another researcher understands what was considered and does not repeat the same work.
Streamline review with bounded batches and explicit checks
To speed up backlink prospecting, separate discovery from qualification. Work through a bounded batch, group clear duplicates first, review obvious audience mismatches, then spend time on contribution and current evidence. Assign a reviewer and a next check to the unresolved cases. Before a larger batch, compare a few decisions between reviewers to align what “useful contribution” means for this brief.
Track discovered rows, unique candidate pages and shortlisted pages separately; none is a forecast of links or rankings. AI can help draft questions or summarize supplied notes, but generated contact details, inferred relevance and invented page observations remain unverified until checked. This guide does not implement automated prospect discovery, an AI service or a vendor-specific CMS detection workflow. Those require their own tested integrations and evidence.
Hand off a qualified question, then prepare outreach
A usable handoff includes the candidate page, discovery trail, current evidence, proposed contribution, review outcome and next action. Download one review as TXT, all six records as CSV or printable HTML, or a header-only blank CSV for your own work. Filters change the screen, not the scope of the full exports. Local edits disappear on reload and do not create account records or messages.
For ongoing work, the existing outreach workspace accepts supplied prospect pages, stores campaign context and review notes, supports reviewed drafts, selected-record exports and follow-up records. That workspace has its own URL validation and duplicate rules; do not assume this teaching sheet imports directly into it. A shortlist precedes outreach and never guarantees acceptance. Google’s link spam policies also applies: paid advertising links should be qualified appropriately, and automated link creation to manipulate rankings is not a prospecting shortcut.