DEV Community

Dwelvin Morgan
Dwelvin Morgan

Posted on

Why 90% of your LinkedIn network is already cold (and the half-life model that explains it)

You've been building your LinkedIn network for years. Connecting after every conference, every job change, every interesting conversation. But here's what nobody tells you about professional relationships: they have a half-life.
The half-life model
In physics, half-life describes the rate at which a radioactive substance decays. The concept maps surprisingly well to professional relationships.
Operationally: a relationship loses approximately 50% of its warmth every 90 days without meaningful interaction.
The formula: Score = 100 × 0.5^(days_since_interaction / 90)
Run this against your connections and the picture gets uncomfortable:

Contact you met at a conference 6 months ago, had one follow-up call, never connected again: warmth score ≈ 25%. Red zone.
Colleague you worked with closely 2 years ago, occasional LinkedIn likes since: warmth score ≈ 6%. Functionally cold.
Someone you had a genuine conversation with 3 weeks ago: warmth score ≈ 79%. Green zone.

Most people's LinkedIn network, when scored this way, looks like: 10% green, 30% yellow, 60% red.
Why this matters for outreach
Cold outreach to red-zone contacts fails at a much higher rate than warm outreach to yellow-zone contacts — not because the relationship is dead, but because you're reaching out with nothing specific to say.
The half-life model solves the "who should I contact" problem. But you still need to solve "what should I say."
The reconnection intelligence problem
When you look at a contact you haven't spoken to in 14 months and try to write a reconnection message, you're essentially writing cold outreach to someone who vaguely remembers you. The default result is either:
a) Generic ("Hey, it's been a while! Would love to catch up")
b) Nothing — you stare at the blank message box and move on
What actually works: a message that references something specific and recent about them. Their company announcement. A post they made. An industry shift that affects their role. This requires research, and research takes time, which is why most people do nothing.
Building a system around this
The basic system I've implemented:

Calculate half-life scores for all connections (based on last interaction date)
Sort by score to identify red and yellow zone contacts
For each red-zone contact, pull recent context: current company news, any public digital footprint (blog, GitHub, LinkedIn activity)
Generate a personalized reconnection message that leads with the specific context
Calendar the follow-up

This turns "I should probably reach out to some people" into a weekly workflow with specific contacts, specific context, and specific copy.
The half-life model alone won't rebuild your network. But it will tell you exactly where to start.
I built this model into SocialCraft AI's LinkedIn Network Intelligence feature — if you want to see the half-life scoring in action against your actual connections, you can upload a LinkedIn CSV export at

Top comments (0)