Bookmark research

Turn LinkedIn bookmarks into original content ideas

Export the bookmarks you collected for inspiration, examine them as sources, and build angles from disagreement, missing detail, and your own evidence—not recycled hooks.

Export the bookmark source material while preserving attribution

LinkedIn labels bookmarks as “saved posts.” Install LinkedIn Bookmark Exporter, open LinkedIn’s Saved Posts page, and choose CSV or Markdown. CSV is easier for filtering bookmarks by author, date, hashtag, or engagement. Markdown is easier to read, search in Obsidian, or use as a document corpus.

Let the extension load the LinkedIn bookmark history and expand truncated text, then keep the downloaded file unchanged. Work from a copy. The exported author and original URL fields must remain attached to every post because they are the path back to attribution and context.

Build a focused LinkedIn bookmark library

Bookmarking is easy, so a raw collection mixes strong research, passing observations, promotions, and advice that contradicts itself. Start with one audience problem—such as onboarding, hiring, positioning, or outbound sales—and filter the exported file to bookmarks that actually address it.

Add a “Why bookmarked” note only to the records you keep in the working set. That note should describe the question, tension, or useful detail you noticed—not summarize the whole post. Keep the original text separate from your interpretation.

Create an evidence map before generating ideas

SourceMain claimEvidence offeredWhat is missingYour experience
Author + URLOne sentenceData, story, example, noneAssumption or gapConfirm, challenge, nuance

This stops high engagement from masquerading as strong evidence. Ten people can repeat the same appealing claim without adding independent support. Mark repeated sources, vendor incentives, and claims that ultimately point to one underlying study.

Use an angle matrix instead of copying hooks

The original contribution should survive if every bookmark link is removed. Sources can inspire and support it, but your observation, test, model, or artifact should carry the piece.

Turn the angle into a one-page brief

  • Reader: a specific role in a specific situation.
  • Problem: the decision or obstacle they face now.
  • Distinct claim: one sentence you are prepared to defend.
  • First-hand proof: numbers, screenshots, workflow details, or examples you own.
  • External support: primary sources behind changing or factual claims.
  • Counterpoint: when your recommendation should not be used.
  • Useful artifact: checklist, template, query, or decision table.

Only after this brief is coherent should you choose a hook. Otherwise a catchy opening hides the fact that the piece has no new value.

Use AI for comparison, not untraceable synthesis

An AI tool can group records, identify disagreement, or list questions the archive does not answer. Ask it to cite record IDs and source URLs. Do not ask it to “rewrite these posts into my voice,” and do not assume its summary accurately represents a source you have not checked.

Useful prompt: “Group these records by claim. For each group, cite source IDs, separate evidence from opinion, identify contradictions, and propose questions that require my firsthand experience to answer.”

Run an originality and attribution review

  1. Search for distinctive phrases that may have carried into the draft.
  2. Attribute specific frameworks, data, or examples to their source.
  3. Replace social-post claims with primary evidence where possible.
  4. Remove any idea whose only value is another person’s wording.
  5. Check that the reader receives a tool or insight unavailable from the sources alone.

To retrieve material from the exported bookmark library, use the full-text search workflow. For structured AI analysis with record-level citations, see the knowledge-base guide.

Sources