Professor Aven
Professional Network Analysis via LinkedIn Data


Tutorial Overview

This tutorial walks you through a structured LinkedIn network analysis using AI-assisted tools (Microsoft Copilot, ChatGPT, or whatever your organization approves - be deliberate about where your data goes) to diagnose your professional network, locate the gaps that matter for where you are going, and build a targeted plan to close them.

🎯 Learning Objectives

By completing this analysis, you will be able to:

  1. Characterize the structural properties of your own network - its size, composition, concentration, and the degree to which the people you know also know each other
  2. Detect homophily and propinquity in your own data - how far your network is a mirror of yourself and of wherever you happen to sit
  3. Locate the structural holes around you - which groups, functions, industries, and levels are disconnected from you, and which of those gaps you are positioned to bridge
  4. Interpret network patterns as opportunity and as exposure - read concentration as risk, redundancy as information you already have, and diversity as access you do not yet use
  5. Benchmark your current network against the one your goals require - and quantify the distance between the two
  6. Turn absences into specific asks - the introductions worth requesting, and who in your existing network is positioned to make them
  7. Convert the diagnosis into a sequenced plan - specific people, a defensible order of priority, and a maintenance practice you can actually sustain

🔒 Your Data Stays Yours

Nothing is collected, submitted, or shared with your employer or with Carnegie Mellon. You run the analysis on your own machine, you see your own results, and you decide what, if anything, you say out loud. Nothing is uploaded or submitted anywhere, and this page stays useful afterward, when you repeat the diagnostic a year from now.

📌 Before You Start

1. Request your LinkedIn connections file. Full instructions in Step 1 below. Do this first: LinkedIn can take up to a week to release the archive.

2. Write one sentence about where you are going. Specific, covering the next two years. The analysis is only as good as the goal you give it.

Client-facing: “Build and own a book of business in treasury management across the mid-Atlantic.”

Non-client-facing: “Expand my influence and access across lines of business so my work reaches the people who make resourcing decisions.”

3. Run the analysis. Work through Steps 1 through 5 below, or run the single portfolio diagnostic at the end in one pass if you are short on time.

🔬 Step-by-Step Tutorial

Step 1: Export Your LinkedIn Connections

On the LinkedIn desktop site:

  1. Go to www.linkedin.com/psettings/member-data, or navigate Settings & Privacy → Data Privacy → Get a copy of your data
  2. Choose “Want something in particular” and tick Connections
  3. Select “Request archive.” LinkedIn emails you a download link. This can take up to a week, which is why it is the first thing you do.
  4. Download and save the Connections.csv file
Expected Format

LinkedIn returns a CSV with the following columns. Every value in it was typed by hand, by your connections and by the employers who set up their profiles, which is why the same company shows up as “Amazon,” “Amazon.com,” and “AWS,” and why many rows carry no company or position at all. Cleaning and normalizing the file is the first real analytical step, not housekeeping.

First Name, Last Name, Email Address, Company, Position, Connected On

Step 2: Upload to AI Tool and Initial Cleaning

Upload your Connections.csv to the AI tool you prefer, then enter the prompt below. It does two jobs at once: it cleans and normalizes the file, and it returns the first cut of analysis on who is actually in your network.

Prompt 1: Data Cleaning & Basic Analysis
Help me clean and analyze my LinkedIn Contacts.

To clean:
1) Clean names of special characters
2) Normalize Company names (e.g., "Amazon.com" → "Amazon",
   "PNC Financial Services Group" → "PNC")

To analyze:
1) Count the number of contacts per Company and provide
the top 5 companies with the most contacts
2) Look up the top 10 companies and research what industries
they belong to
3) Count the unique job types by the number of contacts and
provide the top 5 job titles

What to Expect

  • The tool will clean special characters (é, ñ, symbols in names)
  • Company name normalization (consolidating variations)
  • Summary tables showing:
    • Top 5 companies by contact count
    • Top 10 companies with industry classifications
    • Top 5 job titles/positions in your network
Example Output Structure
Example from Dr. Aven’s own network: Top 5 Companies by Contact Count
Rank Company Contacts
1 Amazon 63
2 Carnegie Mellon University - Tepper School of Business 52
3 Unknown 40
4 Carnegie Mellon University 28
5 Microsoft 25
Interpretation Guide
Pattern Observed Implication
High concentration in one employer May indicate limited network diversity
One line of business dominant Influence is bounded by your own function
Client industries clustering Industry-specific network structure
Many “Unknown” entries Data quality issues, or contacts between roles

Key Questions to Ask:

  • More than 20% of contacts at one company? → Concentration risk
  • Most contacts inside your own line of business? → Limited cross-LOB reach
  • Client industries clustered in one sector? → Exposure to a single industry cycle

Step 2.2: Industry Sector Classification

Prompt 2: Deep Industry Analysis
Show industry sectors for the top 10 companies in my network.
Use official classification systems like NAICS (North American Industry
Classification System) and GICS (Global Industry Classification Standard).

What to Expect

  • Detailed industry breakdown with official codes
  • Sector classifications (Financial Services, Manufacturing, Healthcare, Technology, Professional Services, etc.)
  • A clear read on which parts of the real economy you are actually embedded in
Example Output
Example: Industry Classification with Official Codes
Company NAICS GICS Sector
PNC 522110 Banks Financial Services
Fiserv 522320 Transaction & Payment Processing Fintech & Payments
Grant Thornton 541211 Professional Services Professional Advisors
UPMC 622110 Health Care Providers & Services Healthcare
Why This Matters

Career and deal mobility: moving between adjacent sectors is far easier than jumping across them, for you and for the clients you cover

Information access: homogeneous networks return redundant information

Opportunity risk: over-concentration in a single industry or a declining one

Step 3: Sector Mapping & Network Composition

Prompt 3: Full Classification & Visual Analysis
Generate a complete list of all unique normalized companies in my network
and their respective sectors.

Then create a visual summary of sector distribution in my network
(e.g., pie chart or bar chart).

Then provide insights on:
1) Which sectors dominate my network
2) Which sectors are underrepresented
3) What this distribution reveals about my network's strategic positioning

What to Expect

  • Complete company-to-sector mapping across your whole network, not just the top 10
  • Visual representation showing % of contacts by sector
  • Narrative analysis of your network composition
  • Identification of homophily patterns (e.g., “62% of your network sits inside your own institution”)
Example Visualization: Prof Aven’s Network Composition by Sector
Example: Sector Distribution in a Commercial Banker's Network

Example: Sector Distribution in a Commercial Banker’s Network

Interpretation Framework
How to Interpret Network Composition
Sector Dominance Implication Risk
>50% in one sector Deep embeddedness in one domain Limited cross-industry mobility; echo chamber
Balanced across 4-5 sectors Diverse information access May lack depth in any single domain
Heavy inside your own institution Influence runs through internal channels Few independent sources of opportunity or market intelligence

Strategic Benchmarking

Step 4: Ideal Network Composition

Prompt 4: Personalized Benchmark
As a [YOUR ROLE AND GOAL: e.g., "treasury management officer in corporate
banking who wants to own a middle-market book across the mid-Atlantic" OR
"credit risk director who wants my work to reach the people making resourcing
decisions across other lines of business"] what would be the ideal composition
of LinkedIn contacts across industries?

Provide:
1) Recommended % distribution across sectors
2) Rationale for each sector
3) Key roles/titles I should prioritize in each sector
What to Expect
  • An “ideal” network composition tailored to your stated goal
  • Percentage targets for each sector
  • Rationale explaining why each sector matters for your trajectory
Example Output for Different Goals
Example Ideal Network Compositions by Goal (%)
Goal Target Client Industries Own Institution (Other LOBs) Peer Institutions Professional Advisors Fintech & Payments Private Capital Other
Treasury Mgmt Officer 35 20 10 15 15 0 5
Middle-Market RM 40 15 10 20 5 5 5
Capital Markets / Syndications 25 15 20 15 5 15 5
Credit & Risk Leader 25 30 15 15 5 5 5
Internal Role → Broader Influence 15 45 10 15 10 0 5
Customize This Prompt

Adapt the prompt to your own situation:

  • Client-facing, growing a book: heavier in target client industries and professional advisors (CPAs, attorneys, PE sponsors), lighter inside your own institution
  • Non-client-facing, building influence: heavier in other lines of business, especially two or more levels above you, lighter in external client industries
  • Product specialist → general management: balance your own product area against the lines of business that consume it
  • Traditional banking → payments or fintech: heavier in Fintech & Payments and Private Capital, lighter in legacy corporate

Step 5: Gap Analysis & Recommendations

Prompt 5: Visual Comparison & Action Plan
Create a visual comparison showing:
1) My CURRENT network composition by sector (%)
2) My IDEAL network composition by sector (%)
3) The GAP between current and ideal

Then provide actionable recommendations:
- Which sectors should I expand in
- Specific companies and roles to target
- LinkedIn strategies to close these gaps
- A 12-month network growth roadmap
What to Expect
  • Side-by-side comparison (bar chart or table)
  • Clear identification of over- and under-represented sectors
  • Specific, tactical recommendations
Example Gap Analysis
Example: Current vs. Ideal Network Composition

Example: Current vs. Ideal Network Composition

Example Gap Analysis Table
Gap Analysis Summary: Where to Focus Network Building
Sector Current Ideal Gap Action
Own Institution (Other LOBs) 44% 22% +22% ⚠️ Overrepresented - reach outward, and two levels up
Target Client Industries 23% 34% -11% ✅ Largest gap - add 100+ contacts in covered industries
Professional Advisors 12% 19% -7% ✅ Critical gap - add 70+ advisor contacts
Peer Institutions 11% 11% 0% ✅ On target - maintain
Fintech & Payments 6% 9% -3% ✅ Add 30 contacts in payments and treasury tech
Private Capital 4% 5% -1% ✅ Add 10-15 sponsor contacts
Tactical Recommendations Example

Underrepresented Sectors to Target:

Target Client Industries (Largest Gap: -11%)

  • Companies to target: the 20 largest employers in your footprint in the industries you already cover - healthcare systems, advanced manufacturing, logistics, regional distributors
  • Roles to connect with: CFO, Treasurer, Controller, VP Finance, Director of Procurement
  • How to connect: local CFO roundtables, industry trade associations, and the colleagues who already cover those industries - ask for a warm introduction rather than sending a cold request

Professional Advisors (Critical Gap: -7%)

  • Companies to target: regional CPA firms, middle-market law firms, PE sponsors and independent sponsors active in your footprint
  • Roles to connect with: partner, transaction advisory, outsourced CFO practice lead, deal team principal
  • How to connect: ACG and TMA chapters, referral networks you already sit in, deal announcements in your market
12-Month Network Growth Roadmap
Example: 12-Month Network Growth Roadmap
Timeframe Focus Area Specific Actions Target #
Months 1-3 Audit & Ask Reclassify ‘Unknown’ contacts; request 10 introductions into your thinnest sectors 10 introductions
Months 4-6 Expand Professional Advisors Join 2 advisor or deal-community groups; attend 1 regional conference +70 contacts
Months 7-9 Deepen Client Industries & Reach Upward Connect with finance leaders at 20 target companies; build ties two levels above you +100 contacts
Months 10-12 Visible Expertise Publish 3 LinkedIn posts on your specialty; speak on 1 panel or client webinar +30 contacts

🧭 The One-Pass Portfolio Diagnostic

If you are short on time, run this instead of Steps 1 through 5. It does everything the five steps do - cleaning, classification, sector mapping, benchmarking, and the gap analysis - and treats your network as a portfolio, in a single pass.

Attached is my LinkedIn connections export. Clean it, analyze it, and
diagnose it as a portfolio.
My goal over the next two years is: [PASTE YOUR ONE SENTENCE HERE]
Be specific and concrete throughout.

1. CLEAN THE FILE
   - Clean names of special characters.
   - Normalize company names (e.g., "Amazon.com" → "Amazon",
     "PNC Financial Services Group" → "PNC").
   - Report how many rows are missing company or position data.

2. WHO IS IN MY NETWORK?
   - Count contacts per company and give me the top 5.
   - Count unique job types and give me the top 5 titles.
   - Take the top 10 companies and classify them by industry using
     official NAICS and GICS codes.

3. SECTOR MAP AND COMPOSITION
   - Classify every unique normalized company in the file into a sector,
     not just the top 10.
   - Show the sector distribution as a bar chart with percentages.
   - Tell me which sectors dominate, which are thin, and what the
     distribution says about my strategic positioning.

4. CONCENTRATION - am I overexposed?
   - What percentage of my connections work at my current employer?
   - What percentage sit in my own line of business or function?
   - What percentage are in a single metropolitan area?
   - Flag any cluster where more than 20% share one company, school, or city.

5. DIVERSIFICATION - do I have the right mix?
   - Do I reach across seniority, including two or more levels above me?
   - Do I reach adjacent functions: who buys from, sells to, advises,
     or regulates my work?
   - Do I reach other lines of business inside my own organization?
   - Do I have any ties at competing or counterpart institutions?

6. BENCHMARK AND GAPS - what should this look like, given my goal?
   - Given the goal above, what is the ideal sector composition for my
     network, as a percentage distribution, with the rationale for each
     sector and the roles I should prioritize within it?
   - Compare my current composition to that ideal and show the gap.
   - Name the single most consequential gap in one short sentence.

7. INTRODUCTIONS - where should my next ties come from?
   - Identify the industries, firms, functions, and geographies that are
     under-represented given my goal, and that I would benefit from
     extending into.
   - Name 10 specific introductions worth asking for: the type of person
     or role, the kind of firm, and why that tie matters to my goal.
   - For each one, identify who in my existing network is best positioned
     to make that introduction, and draft the ask in two sentences.

8. REBALANCING - what do I actually do?
   - Five specific people or types of people to cultivate, with why
     and a suggested approach.
   - A 12-month roadmap, quarter by quarter, with a target number of
     new contacts per quarter.

OUTPUT FORMAT
   - Top 3 gaps, each in one line.
   - Sector distribution table: current %, ideal %, gap.
   - Introduction list: who to meet, why, and who can introduce me.
   - Priority cultivation list: name or type, why, suggested approach.
   - Three 15-minute weekly actions for the next month.

⚠️ Common Pitfalls & Solutions

Pitfall #1: “My network looks fine”

Problem: It is easy to accept the network you already have as given, without critical analysis

Solution: Force the comparison - ask “Compared to whom?” and “For what goal?”

Pitfall #2: Treating all connections equally

Problem: Not distinguishing between strong ties, weak ties, and dormant ties

Solution: LinkedIn data does not show tie strength, so answer it yourself: how many of these contacts would respond if you messaged them today?

Pitfall #3: Quantity over quality

Problem: Focusing on total number of connections rather than strategic positioning

Solution: Emphasize network architecture (diversity, brokerage potential) over network size

Pitfall #4: Static analysis

Problem: Treating network analysis as a one-time exercise

Solution: Treat this as a repeatable diagnostic you run once a year, at the same time you set your goals

Pitfall #5: Over-relying on AI without interpretation

Problem: Copy-pasting AI outputs without critical thinking

Solution: Interrogate every AI-generated insight: “Why does the tool recommend this? Does it hold up against what we know about how networks work? Does it fit MY goals?”

🎓 Key Takeaways

By the end of this tutorial, you should be able to answer:

  1. Diagnostic: How is my professional network currently structured?
  2. Comparative: How does my network compare to what I need for where I am going?
  3. Analytical: What do structural holes, weak ties, and homophily reveal about my positioning?
  4. Strategic: What specific actions will I take to architect a more effective network?
  5. Reflective: How does my network reflect my values, my habits, and the path I have taken so far?

Tutorial created by Prof. Brandy Aven, PhD Carnegie Mellon University | Tepper School of Business