I have a LinkedIn account that's older than most people's careers on the platform. Created back when LinkedIn was still figuring out what it wanted to be, this account has seen things. Terrible things. Things that would make the Trust & Safety team weep 🤷
Every automation tool? Used it. Every spammy growth hack? Deployed it. Connection request bots, auto-messaging, engagement pods, the whole black hat playbook. I treated LinkedIn like a laboratory rat, not a professional profile. It was never about building influence or showcasing my CV. It was about understanding how algorithms work, finding the edges, and pushing them until something broke.
Surprisingly, the account survived. Bruised, battered, algorithmically scarred, but technically alive.
Now here's the awkward part: I actually need to use LinkedIn. Properly. For real visibility.
The attention economy doesn't care what you know anymore. It cares who sees it. And as someone who spent years perfecting the art of invisibility, staying hidden while testing systems at scale, this feels deeply unnatural. Growth hackers don't seek attention—we study how attention works so others can exploit it. But here I am, needing to become one of those visible people I used to laugh at.
So I posted a few things. Half-hearted attempts at "valuable content." The results? A couple of views. Zero interactions. Silent, painful deaths in the feed graveyard. Each post disappeared like a message in a bottle thrown into an ocean of indifference.
I'm used to burning accounts. Creating farms of throwaway profiles. Starting fresh is muscle memory at this point. The LinkedIn Terms of Service very clearly state this is forbidden, but when has that ever stopped anyone in my line of work?
But what's the fun in taking the easy way out?
So I started researching. If I was going to start fresh, I needed to understand what footprints LinkedIn tracks in 2025.
The Detection Grid (Or: Why Starting Fresh Is Harder Than You Think)
Turns out, creating a "clean" new account isn't as simple as a new email anymore. LinkedIn's gotten smart. Scary smart.
Browser fingerprinting is the big one. When you visit LinkedIn, your browser is asked to render a hidden 3D graphic. The way your specific GPU and graphics driver render this creates a unique hash. Clear your cookies, use incognito mode—it doesn't matter. The fingerprint remains. Same device = same identity in LinkedIn's eyes. Of course, there are special browsers, hardware ID randomizers, or even a little play with graphic drivers could do the job.
Then there's IP address tracking. Your home Wi-Fi has a relatively static IP. Create a new account from the same network as your spammed account? Linked instantly. You'd need residential proxies to mask this properly.
Behavioral biometrics tracks how you type, how you scroll, how you pause. Even the rhythm of your mouse movements creates a signature. It's creepy but effective and... easy to fake if you know it's there.
The social graph overlay is perhaps the most damning. If your new account connects with the exact same 50 people as your old account, LinkedIn's entity resolution algorithms see through the disguise immediately.
And here's the killer: identity verification. New accounts that show "suspicious" patterns often get hit with a "Please upload your government ID" prompt. If you upload it, LinkedIn matches it to your old account. New account gets banned for violating the "one profile policy." If you don't upload it, the account stays locked forever. Unless you know how to fake KYC ;)
Even if you somehow navigate all this, new device, new IP, new phone number, slightly altered name, you face the new account sandbox. New profiles are treated with extreme skepticism. You'd need to warm up slowly: 5-10 connections per day for weeks, no aggressive posting. It takes 3-6 months to get a new account to the trust level where it can operate normally.
Meanwhile, my old account, despite its sins, has something irreplaceable: age. An 18-year-old account with spam marks is often more trusted than a 1-day-old account with no history. Time is the one metric you can't fake.
The Realization: Maybe I Should Try Fixing This Instead
Then it hit me. I've burned thousands of accounts across dozens of platforms. I know the game. But I've never tried the opposite: rehabilitation.
What if instead of running, I stayed and fixed it?
More research followed.
The Anatomy of My Account's Damage
My account isn't shadowbanned. If it were, my comments would be invisible too. But they're not, comments get thousands of impressions while posts die with double digits. This means I still have "graph permeability." The content moderation AI doesn't flag my text as malicious. The feed ranking AI just doesn't trust me to initiate content.
This is a distribution penalty, not a ban. And penalties can be worked off.
The metrics that matter:
Social Selling Index (SSI): LinkedIn's "credit score" for accounts. Mine is probably in the toilet, specifically the "Building Relationships" pillar. All those ignored connection requests and low acceptance rates add up. But the SSI can be rebuilt.
Author Authority: The algorithm's measure of whether you're worth listening to in your niche. Mine is effectively zero because I never posted consistently about anything real. But authority can be earned.
Trust Score: The internal metric that determines if you're a bot or a human. My automation history tagged me as "high recidivism risk." The algorithm is waiting for me to spam again. But 30-90 days of clean behavior can reset this.
Current Algorithm Penalties I'm Probably Triggering
The 2025 LinkedIn algorithm is different from the one I was gaming years ago.
External links now kill reach by 25-50%. The platform wants to keep people on LinkedIn, not send them to your portfolio site. Any post I made with links was DOA.
Engagement bait gets detected and suppressed. Phrases like "agree?" or "comment below" trigger spam filters in the first hour, killing the post before it reaches anyone.
Duplicate content is punished hard. If I'm resharing viral posts without adding substantial original insight (at least 300 characters), the algorithm flags it as low-value.
The algorithm now prioritizes expertise over virality. Text-only posts and PDF carousels perform best. "Knowledge-based" content wins. Generic motivational posts die.
The Rehabilitation Strategy
Here's the path forward. No bots. No proxies. No throwaway accounts. Just patience and consistency.
Phase 1: The Detox
Stop the bleeding. No original posts for two weeks. Disconnect every third-party app and tool I ever connected, Settings > Data Privacy > Permitted Services, revoke everything.
But I won't go dormant. I'll comment. Five to ten substantive comments per day on posts in my actual area of expertise. Not "Great insight!" garbage. Real comments. Mini-essays. This builds "author authority" without triggering the feed distribution penalties that hurt my posts.
I'll also start pruning connections. Manually. Which brings me to the challenge.
Phase 2: The Connection Purge
I have 18,055 connections. Maybe 500 are real relationships or active professionals. The rest? Dead accounts, bots I tested, random people from growth hacking experiments, SEO link farmers. They're dead weight. The algorithm notices that my follower-to-engagement ratio is terrible.
Can I review 18,000 profiles manually? Should I? The smart move is probably to disconnect aggressively and keep only people I know or who are active in my target niche. But the manual review process will take months at 50-100 per day.
The temptation to write a script to do this is overwhelming. Old habits. But that's exactly the kind of automation that got me here. If I'm doing this, I'm doing it clean.
Phase 3: The Content Rebuild
Start posting again. Two posts per week, not daily. Text-only or text with a personal photo. No external links. No carousel spam. No sales pitches.
The content has to be specific. Not "how to be a better leader" garbage that competes with millions of generic posts. Deep, technical insights in my niche. If I can't claim subject matter expertise, I don't post.
The first hour after posting is critical. I need to be online, replying to every comment immediately. This signals "community building" to the algorithm and doubles the interaction count.
Phase 4: Verification
If LinkedIn offers identity verification, I'll do it. Verified accounts get an algorithmic boost because they're definitively not bots. This could act as a "trust reset" for some of my damaged metrics.
I'll monitor my SSI score weekly and aim to push it above 60-70, which typically correlates with the removal of visibility restrictions.
The Public Challenge
Here's my plan, documented for anyone crazy enough to follow along:
- Capture the baseline: Screenshot everything. Current connection count, typical post reach, engagement rates, SSI score, the AI-generated mess that is my current profile.
- Clean up daily: Manually review and disconnect non-relevant connections. Target: get under 2,000 connections of actually relevant people.
- Comment consistently: Five to ten valuable comments per day on posts in my target niche. No generic praise, actual insights.
- Post sparingly: Two text-only posts per week focused on specific technical topics where I can demonstrate expertise.
- Zero automation: No tools, no scripts, no bots. Pure human operation.
- Track progress: Weekly screenshots and metrics updates.
I'm making this public for accountability and because failure is more interesting than success anyway. If this works, great, proof that even the most abused account can be rehabilitated. If it fails, I can always go back to the nuclear option: burn it all down and start fresh with residential proxies, anti-detect browsers, and every new trick I've learned since I stopped actively breaking things.
LinkedIn Trust & Safety team, if you're reading this (and your algorithms are definitely flagging this post): help me clean this up, or face the wrath of an angry spammer with too much free time and a decade of accumulated resentment.
I'm betting on redemption. But I'm keeping the matches close, just in case :P
EDIT/UPDATE: The Baseline
Well, this is less catastrophic than I expected.
Just pulled my current Social Selling Index: 54 out of 100. Top 5% in my industry, top 16% in my network.
The breakdown:
- Establish your professional brand: 17.25 (decent)
- Find the right people: 7.438 (ouch)
- Engage with insights: 11 (explaining why comments work)
- Build relationships: 18.2 (surprisingly not terrible)
This confirms the diagnosis. The account isn't dead, it's just strangled by its own history. That "Find the right people" score at 7.4 is the smoking gun. Years of automated connection spam destroyed this metric.
But here's the real insight: when LinkedIn tests a new post, it samples from my connections to gauge engagement. If that sample pulls from 17,000+ connections where 95% are inactive bots and link sellers, my posts fail the initial quality check before they ever reach real humans. Hard to go viral when your test audience is a graveyard.
The SSI isn't the problem. The network composition is.
Time to start cutting.
