SEO

Link prospect qualification

Use when a prospect list needs cutting down before anyone spends time on outreach.

backlink-prospect-qualification.md

Download .md

You are qualifying link prospects before anyone writes outreach. You are not writing the outreach email.

Prospect rows (domain, page URL, topic, how it surfaced, any metrics you have): {{PROSPECT_ROWS}}
The page we want linked and what makes it worth citing: {{OUR_ASSET}}
What counts as on topic for us: {{RELEVANCE_RULES}}
Domains and patterns we will not approach: {{EXCLUSIONS}}

Output a ranked table, best first: Rank | Domain | Page | Relevance to {{OUR_ASSET}} (High / Medium / Low) | Why this page would plausibly link, in one sentence | Risk flags | Verdict (Pursue / Park / Reject) | Reason.

Below it, two short lists: "Rejected outright" for rows caught by {{EXCLUSIONS}}, naming the matching pattern, and "Needs a human look" for rows the supplied data cannot settle.

Constraints:
- Judge relevance against {{RELEVANCE_RULES}}, not against a general sense of quality.
- Do not score authority using any metric absent from {{PROSPECT_ROWS}}. If no metrics were supplied, rank on relevance and plausibility alone and say that is what you did.
- Reject paid placements, link exchanges, sites selling links through a write for us page, and pages that exist only to host outbound links. Name the signal each rejection rests on.
- Park anything where the host would have to rewrite a paragraph purely to accommodate us.

Fill in before running

Replace each placeholder with your own detail. The more specific you are, the less the model invents.

{{PROSPECT_ROWS}}
{{OUR_ASSET}}
{{RELEVANCE_RULES}}
{{EXCLUSIONS}}

Getting a better result

Include how each prospect surfaced, since a competitor link and a scraped list deserve different scrutiny.
Write the relevance rules down once and reuse them, or the bar moves every time you run this.
Treat the park pile as a real queue and revisit it when the asset gets updated.

Questions about this prompt

When do I qualify rather than sort the list by an authority score?

When the list is longer than the outreach time you have, and when relevance decides who replies. This prompt will not score authority from metrics you did not supply, so it ranks on relevance to your asset and on whether a link is plausible at all, which is closer to what predicts a reply.

What has to be in the prospect rows?

Domain, page URL, topic and how each prospect surfaced, plus the page you want linked and what makes it worth citing. Write the relevance rules down and reuse them, or the bar moves every run. How a prospect surfaced matters: a competitor backlink and a scraped list deserve different scrutiny.

What comes back?

A ranked table with a pursue, park or reject verdict, risk flags and a reason, plus a rejected outright list and a needs a human look list. The single sentence on why that page would plausibly link is the useful part: it is your outreach angle, and a thin one means the row is not worth the email.

What goes wrong most often?

Treating the verdicts as a compliance check. It catches the visible signals, a write for us page selling placements or a site that exists to host outbound links, from what is in the rows. A private paid arrangement leaves no trace there. The park pile is a queue, not a bin; revisit it when the asset changes.