Third-Party Source Placement

For recommendation and comparison questions, assistants lean on sources that read as independent: listicles, review aggregators, forum threads, industry directories. Optimising your own site harder moves none of that, because your site was never in the citation set to begin with. The work is to find which pages are in it and to be accurately represented there.

The skill file

---
name: third-party-source-placement
description: Use when assistants answer your category confidently but cite roundups, forums and review sites rather than any vendor page, including yours.
---

# Third-Party Source Placement

For recommendation and comparison questions, assistants lean on sources that read as independent: listicles, review aggregators, forum threads, industry directories. Optimising your own site harder moves none of that, because your site was never in the citation set to begin with. The work is to find which pages are in it and to be accurately represented there.

## What you need first

- The 15-25 buyer questions you already track for citations
- The cited URLs recorded from those runs, at least two rounds apart so one-off variation is visible
- A list of every place you are currently listed, reviewed or discussed
- Named owners for publisher outreach, review platforms and community accounts

## Method

1. Collect the cited URLs across every question and every assistant, then count appearances by domain. A short list of sources usually carries most of the answers in a category, and that list is the actual target.
2. Sort the recurring sources into three groups: ones you can be added to, ones where you appear but are described wrongly, and ones you will never influence. Only the first two are work.
3. Fix the wrong descriptions before chasing new placements. A stale feature list or a two-year-old price on a frequently cited source is doing active damage and usually costs one email.
4. For sources you are missing from, read the stated inclusion basis before pitching. Most roundups have one, and a pitch that ignores it gets nothing and burns the contact.
5. Where the source is a community thread, answer the question asked with the vendor badge visible. Undisclosed accounts get detected and screenshotted, and the resulting thread joins the citation set.
6. Judge the result on whether the cited URL set changed, not on traffic. The citation set is what you set out to alter, and most of these sources send little referral traffic even when they drive the answer.
7. Re-count the source list quarterly, because the pages assistants lean on shift as those pages are themselves rewritten and re-ranked.

## What this produces

A ranked table of third-party domains by citation frequency in your category, with an action per source and the observed outcome after each correction.

## Where this goes wrong

- Paying for a listing on a source that never appears in the citation set, which is the default outcome when nobody counted first
- Posting under an undisclosed account, which usually surfaces and leaves a thread about your astroturfing in the corpus
- Correcting a source once and never re-checking, so the old description returns at the publisher next refresh
- Reading a citation as proof that the source drives sales, when it drives the answer, which is a different and smaller claim

---

From the QuQi skill library - https://www.quqi.io/skills/third-party-source-placement

What you need first

• The 15-25 buyer questions you already track for citations
• The cited URLs recorded from those runs, at least two rounds apart so one-off variation is visible
• A list of every place you are currently listed, reviewed or discussed
• Named owners for publisher outreach, review platforms and community accounts

Method

1. 01 Collect the cited URLs across every question and every assistant, then count appearances by domain. A short list of sources usually carries most of the answers in a category, and that list is the actual target.
2. 02 Sort the recurring sources into three groups: ones you can be added to, ones where you appear but are described wrongly, and ones you will never influence. Only the first two are work.
3. 03 Fix the wrong descriptions before chasing new placements. A stale feature list or a two-year-old price on a frequently cited source is doing active damage and usually costs one email.
4. 04 For sources you are missing from, read the stated inclusion basis before pitching. Most roundups have one, and a pitch that ignores it gets nothing and burns the contact.
5. 05 Where the source is a community thread, answer the question asked with the vendor badge visible. Undisclosed accounts get detected and screenshotted, and the resulting thread joins the citation set.
6. 06 Judge the result on whether the cited URL set changed, not on traffic. The citation set is what you set out to alter, and most of these sources send little referral traffic even when they drive the answer.
7. 07 Re-count the source list quarterly, because the pages assistants lean on shift as those pages are themselves rewritten and re-ranked.

What this produces

A ranked table of third-party domains by citation frequency in your category, with an action per source and the observed outcome after each correction.

Where this goes wrong

• Paying for a listing on a source that never appears in the citation set, which is the default outcome when nobody counted first
• Posting under an undisclosed account, which usually surfaces and leaves a thread about your astroturfing in the corpus
• Correcting a source once and never re-checking, so the old description returns at the publisher next refresh
• Reading a citation as proof that the source drives sales, when it drives the answer, which is a different and smaller claim

Use this skill in your own AI

The download is a plain markdown file with the name and trigger in its frontmatter. Where an assistant supports skills it can load itself, that frontmatter is what it reads to decide this one applies.

Questions about this skill

More in AI search (GEO)