Solanasis Docs

linkedin-export-playbook-reviewed

15 min read

LinkedIn Data Export → Claude Code → Personal CRM

Reviewed playbook for exporting LinkedIn data, analyzing behavior, and building a founder-grade personal CRM

Last reviewed: March 11, 2026
Audience: Founder / operator / AI agent / Claude Code workflow
Primary use case: Export your own LinkedIn data, analyze it locally, and convert it into actionable relationship intelligence.

1) Executive summary

Yes — if you live in Boulder, Colorado, you can export your LinkedIn account data using LinkedIn’s built-in Get a copy of your data workflow.

This is the practical path for a U.S. user. LinkedIn also references a programmatic portability API, but LinkedIn’s help documentation says that API is for EU / EEA / Switzerland members, not a general U.S. portability path.

For a Boulder-based user, the important takeaway is:

  • You can export your own LinkedIn data.
  • You do not need a special Colorado-only export process just to get the data.
  • Colorado privacy law is still relevant as general background because Colorado recognizes privacy/access/portability rights, but your real-world workflow here is LinkedIn’s own export feature.

2) Official export path

LinkedIn’s current self-service workflow

From LinkedIn desktop:

  1. Click Me
  2. Open Settings & Privacy
  3. Go to Data Privacy
  4. Under How LinkedIn uses your data, click Get a copy of your data
  5. Choose either:
    • specific data categories, or
    • a larger download / archive
  6. Click Request archive

Important operational details

  • Specific data categories are typically emailed within minutes.
  • The larger download is typically emailed within 24 hours.
  • The download link is available for 72 hours.
  • LinkedIn says this feature is not available on mobile.
  • LinkedIn advises downloading only from a personal computer, not a public one.

Best practice

Request the larger archive unless you have a narrow reason not to.

Why:

  • It gives you much better raw material for AI analysis.
  • It is more useful for personal CRM building.
  • It helps preserve more context if LinkedIn later changes features, fields, or availability.

3) What data is actually in the export

LinkedIn states that available files depend on your account activity. You only receive categories that apply to your account.

That means:

  • if you never used a feature, the corresponding file may not exist
  • if you do not have certain profile sections, they may not appear in the archive

Commonly useful categories

Below are the categories that matter most for CRM, behavioral analysis, and founder workflows.

A. Connections

Usually the most immediately useful file.

Common fields may include:

  • first name
  • last name
  • email address if the other member allowed it
  • company
  • position / title
  • connected on / connection date
  • public profile URL

Reality check: missing emails are normal. LinkedIn explicitly says some emails will be missing because members control whether their email address can be downloaded by connections.

B. Messages

This can be one of the highest-value files if included.

Potential data includes:

  • sent messages
  • received messages
  • archived messages
  • dates / timestamps
  • subject lines
  • message content
  • public profile URL of the member

This is extremely useful for:

  • reconstructing relationship history
  • summarizing prior conversations
  • identifying warm contacts
  • finding stalled conversations worth reactivating

C. Invitations

Helpful for seeing network-building behavior.

Potential data includes:

  • invitations sent
  • invitations received
  • invite dates
  • names
  • URLs
  • invite messages

This is useful for:

  • seeing who you reached out to but never advanced
  • understanding your networking cadence
  • spotting patterns in acceptance / non-acceptance

D. Search queries

A surprisingly valuable behavioral file.

Potential data includes:

  • search terms
  • search phrases
  • dates of searches

This helps answer:

  • what kinds of people you keep looking for
  • whether your actual search behavior aligns with your business goals
  • where your attention is going over time

E. Comments, reactions, shares, votes

Good for content-strategy analysis.

This can help you understand:

  • what you engage with most
  • whether you mostly react vs comment vs post
  • what topics dominate your attention
  • whether your public behavior reinforces the positioning you want

F. Profile and resume-like data

May include structured profile records such as:

  • positions
  • education
  • certifications
  • skills
  • recommendations
  • endorsements
  • profile data and account history

Useful for:

  • backups
  • content drafting
  • syncing a personal knowledge base
  • resume / bio regeneration

G. Imported contacts

Potentially includes contact records you imported into LinkedIn.

May include:

  • names
  • phone numbers
  • email addresses

Treat carefully because this can contain sensitive personal data.

H. Inferences

This is one of the more interesting files.

LinkedIn’s larger export can include “information we infer about you based on your profile and activity.”

This can be valuable because it may reveal:

  • how LinkedIn categorizes you
  • possible interest/profile buckets
  • where your platform behavior is signaling something different from your intentional positioning

I. AI-powered conversations

LinkedIn says exports may include certain conversations with some of its AI-powered chat / voice features where applicable.

4) What is not in the export

This is where people get tripped up.

LinkedIn explicitly says it does not provide:

  • People You May Know
  • Who’s Viewed Your Profile

Also important:

  • you cannot export contacts that are not 1st-degree connections
  • some email addresses will be missing by design
  • the export only includes your personal data, not a full dump of everyone else’s data

So do not build your workflow assuming you’ll get:

  • full second-degree network data
  • complete email coverage
  • profile viewer history
  • full recommendation engine data

5) Colorado / Boulder angle

What matters in practice

If you live in Boulder, Colorado:

  • yes, you can export your LinkedIn data
  • the in-product LinkedIn export is the practical path
  • Colorado’s privacy regime is relevant as background context, but it does not change the basic export workflow you use day to day

Practical interpretation

Colorado’s privacy law recognizes privacy rights including data access / portability concepts, but for this use case the important operational fact is simple:

LinkedIn already offers a member-facing account export flow, so that is the path you use.

This is not a legal memo. If you ever need a legal-rights escalation or a dispute resolution path, that becomes a separate privacy-law issue.

6) Why this export is strategically useful

Most people think of LinkedIn export as just a backup.

That’s underselling it.

Used properly, it becomes a raw dataset for:

  • relationship intelligence
  • network cleanup
  • content strategy
  • founder CRM building
  • follow-up prioritization
  • behavioral self-audit

A. Reconstruct your relationship graph

Using:

  • connections
  • invitations
  • messages

You can identify:

  • who you know well
  • who you connected with but never followed up with
  • who you had real conversations with
  • who went cold
  • who is likely worth re-engaging
  • which industries and roles dominate your network

B. Audit your actual LinkedIn behavior

Using:

  • searches
  • reactions
  • comments
  • shares
  • messages

You can discover:

  • where your attention actually goes
  • whether your behavior aligns with your sales / brand goals
  • whether you are networking intentionally or just grazing
  • who you repeatedly engage with
  • whether your public footprint supports your positioning

C. Build a personal CRM from real behavior

This is the big one.

Instead of keeping contacts trapped inside LinkedIn, you can turn the export into a real relationship system with:

  • notes
  • tags
  • warmness scoring
  • last touch
  • next step
  • opportunity type
  • referral potential
  • partner potential

D. Preserve strategic continuity

If your account changes, gets restricted, loses visibility, or LinkedIn changes features, your local export becomes your safety net.

7) How people use this with Claude Code

The clean, sane workflow is not:

  • letting an AI loosely click around LinkedIn in sketchy ways
  • scraping aggressively from the live platform
  • depending on browser automation as the core system

The much better workflow is:

  1. Export your data from LinkedIn
  2. Store the archive locally
  3. Feed the files to Claude Code
  4. Have Claude parse, normalize, summarize, and score the data
  5. Push the cleaned result into a CRM / database

This gives you:

  • more control
  • better privacy hygiene
  • repeatability
  • better auditability
  • easier deduping and enrichment

Option A — Recommended

Export → local processing → structured CRM

Pipeline:

LinkedIn export ZIP
→ local extracted folder
→ Claude Code parsing scripts
→ normalized CSV / SQLite / JSON
→ Baserow / Airtable / Coda / Notion / SQL

Why this is the best default

It balances:

  • practicality
  • privacy
  • reusability
  • cost control
  • AI usefulness

What Claude Code should do

Claude Code can help with:

  • inspecting files
  • mapping schemas
  • cleaning broken encodings / weird delimiters
  • deduplicating contacts
  • normalizing titles and companies
  • extracting message summaries
  • assigning tags
  • ranking re-engagement opportunities
  • generating follow-up lists
  • producing CRM import files

9) Suggested local folder structure

linkedin-data/
  raw/
    2026-03-11-linkedin-export.zip
  extracted/
    2026-03-11/
      Connections.csv
      Invitations.csv
      Messages.csv
      SearchQueries.csv
      Reactions.csv
      Comments.csv
      Profile.csv
      Inferences.csv
      ...
  working/
    schema-notes.md
    field-maps.json
    clean_contacts.csv
    clean_interactions.csv
    message_summaries.csv
    relationship_scores.csv
  output/
    linkedin-network-summary.md
    crm-import.csv
    reengage-priority-list.csv

Best practices

  • Keep the original ZIP untouched in /raw/
  • Extract to a date-stamped folder
  • Never overwrite raw files
  • Keep a schema-notes.md file because LinkedIn fields may change over time
  • Version your cleaned outputs

10) Suggested CRM schema

Below is a pragmatic starter schema for converting LinkedIn data into a personal CRM.

Table: contacts

Recommended fields:

  • full_name
  • first_name
  • last_name
  • linkedin_url
  • email
  • current_company
  • current_title
  • connected_on
  • source
  • city_or_region
  • relationship_type
  • network_segment
  • warmth_score
  • strategic_value_score
  • last_interaction_date
  • last_interaction_type
  • notes
  • next_action
  • next_action_due
  • owner

Table: interactions

Recommended fields:

  • contact_key
  • interaction_date
  • interaction_type
  • direction
  • subject
  • content_summary
  • raw_reference
  • sentiment
  • followup_needed

Table: relationship_scores

Recommended fields:

  • contact_key
  • message_count
  • comment_count
  • reaction_count
  • invite_status
  • days_since_last_touch
  • warmth_score
  • reengage_priority
  • partner_fit
  • client_fit
  • referral_fit

A simple first-pass scoring system is often enough.

Warmth score example

Score based on:

  • recent message exchanges
  • accepted invitation + real follow-up
  • repeated comment or DM interaction
  • mutual relevance to your business goals

Example rough logic:

  • 90–100 = active relationship
  • 70–89 = warm
  • 40–69 = known but underdeveloped
  • 10–39 = weak tie
  • 0–9 = dormant / no real evidence of relationship

Strategic value score example

Score based on:

  • role seniority
  • business relevance
  • influence / connector potential
  • target market fit
  • partnership likelihood
  • investor / advisor / buyer alignment

Do not overcomplicate the first version. Build version 1 fast, then refine.

12) High-value questions Claude Code can answer

Once the export is normalized, Claude Code can answer much better questions than LinkedIn’s UI can.

Relationship questions

  • Which 1st-degree connections have I never messaged?
  • Which people messaged me in the past but I never replied to properly?
  • Which old contacts are now in more senior roles?
  • Which people look like the best partner candidates?
  • Which connections are likely buyers for Solanasis-style offers?

Behavior questions

  • What kinds of people do I search for most often?
  • Do my engagement habits align with the market I say I want?
  • Which topics dominate my comments and reactions?
  • Am I spending attention on peers instead of buyers?

Content questions

  • Which industries do I visibly interact with most?
  • Which people’s content do I repeatedly engage with?
  • What content themes should I post more about based on actual network overlap?

CRM questions

  • Who should I re-engage in the next 30 days?
  • Who belongs in partner vs prospect vs peer buckets?
  • Which contacts likely belong in a local founder / Boulder / ops / nonprofit segment?

13) Strong use cases for a founder / consultant

For personal brand

Use the export to see whether your visible behavior aligns with how you want to be known.

For sales / partnerships

Use it to identify:

  • warm intros
  • neglected leads
  • connector relationships
  • community nodes
  • partnership candidates

For memory augmentation

Use message history and connection records to build a system that remembers:

  • where you met
  • what was discussed
  • what they care about
  • what you should follow up on

For local ecosystem mapping

With some tagging and enrichment, you can create buckets like:

  • Boulder / Denver founders
  • nonprofit leaders
  • SMB operators
  • cybersecurity buyers
  • AI / ops consultants
  • community builders

14) Security, privacy, and ethics

This is where people get sloppy.

Sensitive-data reality

Your export can contain:

  • private messages
  • personal emails
  • imported contacts
  • sensitive relationship context

Treat it like sensitive business data.

Rules of thumb

  • Store it locally in a controlled folder
  • Limit access
  • Do not dump raw message archives into random web tools
  • Prefer local processing when possible
  • Redact before sharing with assistants or contractors
  • Keep the raw archive as restricted as possible

Safer operating model

  • raw export stays local
  • AI works on local files
  • only cleaned / summarized outputs get pushed to CRM
  • message bodies are optionally summarized and then excluded from operational CRM views

This keeps the CRM useful without making it a privacy mess.

15) Practical limitations and caveats

Limitation 1: The export is incomplete

It is very useful, but it is not a full graph of LinkedIn.

You will not get:

  • full second-degree network data
  • profile viewer data
  • people-you-may-know suggestions
  • perfect email coverage

Limitation 2: Schemas can change

Field names, file names, formats, and availability can change.

That is why your Claude Code workflow should:

  • inspect files dynamically
  • map fields explicitly
  • log assumptions in schema-notes.md

Limitation 3: Message quality varies

The messages export can be powerful, but it may require cleanup depending on format and time ranges.

Limitation 4: Not everything should go into the CRM

A personal CRM should not be a raw dump of every private conversation.

Instead:

  • summarize
  • classify
  • keep only what is operationally useful

Minimum cadence

Export your larger archive:

  • quarterly, or
  • before major account changes, or
  • before closing / merging accounts, or
  • before a major outreach push

Better cadence for a founder

  • monthly light review of CRM outputs
  • quarterly full LinkedIn export refresh
  • annual deep analysis of network behavior and positioning

17) Suggested Claude Code task list

Below is a concrete handoff another AI can follow.

Task 1 — Inspect archive contents

  • list files
  • identify file names and delimiters
  • detect encoding issues
  • create schema-notes.md

Task 2 — Normalize contacts

  • merge connections into a clean contacts table
  • standardize names, URLs, titles, companies
  • retain source provenance

Task 3 — Normalize interactions

  • parse invitations, messages, comments, reactions
  • create one unified interactions table
  • classify interaction types

Task 4 — Summarize message history

  • summarize by person
  • extract last known topic
  • identify pending follow-up opportunities

Task 5 — Score relationships

  • assign warmth score
  • assign strategic value score
  • assign re-engagement priority

Task 6 — Create CRM import files

  • export contacts.csv
  • export interactions.csv
  • export reengage-priority.csv

Task 7 — Create founder insights memo

Generate a Markdown report answering:

  • what my LinkedIn behavior says about my priorities
  • which network segments dominate my activity
  • who I should re-engage first
  • what content themes best align with my real network

18) Good prompt ideas for another AI

Prompt: parse the export

Review the extracted LinkedIn export folder. Identify all files, infer their schemas, and create a schema-notes.md file summarizing file names, columns, delimiter/encoding issues, and which files are most useful for CRM building.

Prompt: build contact master

Using the LinkedIn export files, create a normalized contacts master table with one row per person. Include full name, LinkedIn URL, email if available, current company, current title, connection date, and inferred relationship type. Deduplicate carefully and preserve source provenance.

Prompt: build interactions log

Create a unified interactions table from messages, invitations, comments, reactions, and other engagement data. Normalize dates, classify interaction types, and create concise content summaries where useful.

Prompt: score relationships

Score each relationship for warmth, strategic relevance, and re-engagement priority. Use recent interaction activity, message presence, role relevance, and connection age. Explain the scoring logic in a separate Markdown note.

Prompt: produce founder memo

Analyze my LinkedIn export and write a candid founder memo covering: who dominates my network, whether my behavior aligns with my business goals, which relationships I should re-engage, and what content themes are most aligned with my actual network and search behavior.

19) What to push into a CRM vs what to keep out

Push into CRM

Good candidates:

  • contact identity data
  • LinkedIn URL
  • connection date
  • current company / title
  • relationship tags
  • interaction summary
  • last touch date
  • next step
  • follow-up priority

Usually keep out of CRM or heavily summarize

  • raw full message bodies
  • highly personal conversation details
  • imported contact fields you do not operationally need
  • sensitive AI-chat data

Rule:

the CRM should help you act, not become a surveillance archive

20) Recommendation for this specific use case

For a founder/operator building a serious relationship system:

Best path

  1. Export the larger LinkedIn archive from desktop
  2. Store it locally in a versioned folder
  3. Let Claude Code inspect and normalize the files
  4. Create:
    • contacts.csv
    • interactions.csv
    • relationship_scores.csv
    • network-insights.md
  5. Import the cleaned tables into your working CRM
  6. Repeat quarterly

Why this wins

Because it turns LinkedIn from:

  • a noisy feed
  • a fragile platform dependency
  • a half-usable contact graph

into:

  • a portable relationship asset
  • a decision-support system
  • a founder memory layer
  • a real personal CRM seed dataset

21) Bottom line

Yes — you can export your LinkedIn data while living in Boulder, Colorado.

The meaningful workflow is not “just download a CSV.”

The meaningful workflow is:

  • export the larger archive
  • process it locally
  • use Claude Code to analyze your behavior and relationships
  • convert it into structured founder intelligence and a living CRM

That is the real leverage.

22) Official source notes

The following points were verified against official sources before this playbook was revised:

  • LinkedIn’s current export path is Settings & Privacy → Data Privacy → Get a copy of your data.
  • LinkedIn says you can request specific categories or a larger download.
  • Specific categories are generally available within minutes; the larger archive is usually available within 24 hours.
  • The download link remains available for 72 hours.
  • The feature is not available on mobile.
  • LinkedIn says it provides only your own personal data and does not provide People You May Know or Who’s Viewed Your Profile data.
  • LinkedIn notes that some connection emails are missing because members control whether their email can be downloaded.
  • LinkedIn’s help page for exporting connections says the larger archive includes connections, verifications, contacts, account history, and inferred information based on profile and activity.
  • LinkedIn’s help page on email visibility confirms users can control whether their primary email address can be downloaded by connections in data exports.
  • Colorado’s privacy framework is real and relevant as general context, but the operational export path here is still LinkedIn’s built-in member export.

23) Source references for verification

  • LinkedIn Help — Download your account data
  • LinkedIn Help — Export connections from LinkedIn
  • LinkedIn Help — Visibility of Your Email Address
  • Colorado Attorney General — Colorado Privacy Act (CPA)

Graph View