LinkedIn Social Campaign — LinkedIn Data + AI Insights
Dmitri Sunshine | Solanasis LLC
Generated: 2026-03-26 Campaign purpose: Thought leadership series surfacing key findings from the linkedin-data-maximizer pipeline. Frames Solanasis as a firm that does AI-assisted data analysis — not just talks about it. Tone: Authentic, specific, no corporate fluff. Dmitri’s actual voice: affirm first, challenge second, share self, end with a hook. Post count: 7 drafts Each post: 150-300 words, standalone, includes hashtags
Post 1 — The Hook Post (Series Opener)
Suggested timing: First post in series. Sets up the rest.
I downloaded everything LinkedIn has ever recorded about me.
37 CSV files. 18 years of data. Every connection, every DM, every ad I’ve ever clicked, every company I’ve ever followed.
Then I fed it to an AI pipeline I built.
What came back was a mirror I didn’t fully expect.
LinkedIn’s inference engine — the system that decides how to categorize you for advertisers — had tagged me in seven categories. Among them: “HR professional” and “interested in technology media.”
Not cybersecurity. Not fractional executive. Not CIO or COO.
I’ve been building Solanasis as a fractional CIO/CSIO/COO firm for a year. LinkedIn’s AI had decided I was something else entirely — and was quietly optimizing my reach toward the wrong audience the entire time.
This is the first post in a series on what I found when I actually looked at my own data. The numbers are specific. The findings were uncomfortable in places. The process is replicable.
If you’ve had a LinkedIn account for more than a few years, your export is sitting there. Most people never open it.
Worth opening.
ResponsibleAI DataAnalysis LinkedIn FractionalLeadership SMB Solanasis
Post 2 — The 2,024 People I Never Reached
Suggested timing: 2-3 days after Post 1.
There’s a file in the LinkedIn data export called Member_Follows.csv.
It lists every person who has followed you — not connected with you. Followed.
People who saw your profile, decided they wanted to see more of what you post, and never sent a connection request.
I had 2,030 entries in that file.
2,024 of them had never connected with me.
That’s 2,024 people who raised their hand. Who said, effectively, “I’m interested in what you’re doing.”
And I had never reached back.
This is the largest untapped outreach pool in my entire LinkedIn history — and I didn’t know it existed until I ran the analysis. The data is public, the file is in the standard export, and most people never look at it.
In outreach terms: these aren’t cold contacts. They’re warm. They chose to follow. The barrier to engagement is lower than a cold connection request because you’re already on their radar.
If you do LinkedIn outreach and you haven’t worked your follower list, you’re starting from a harder position than you need to be.
Five minutes to export. One CSV file to open. Worth the five minutes.
What’s in yours?
LinkedInStrategy Outreach DataDriven GTM SMB FractionalLeadership Solanasis
Post 3 — What AI Found in My Own Writing
Suggested timing: 4-5 days after Post 1.
I had an LLM read 670 samples of my own writing.
Comments, posts, DMs, invitation messages — everything LinkedIn had on record. Then I asked it to find the patterns I couldn’t see from inside my own voice.
What it found:
I always affirm before I challenge. Almost every critical comment I’ve written starts with warmth. “I love that you’re doing this, and yet…” The contrast lands because the affirmation is genuine. The skepticism registers because it’s unexpected.
I write very differently in public versus private. In posts and comments, I perform warmth and expertise. In DMs, I move fast — personal context, current project, meeting request, calendar link. Four beats, consistent across 562 conversations.
Roughly 30% of my public comments include personal disclosure. A number the analysis flagged as unusually high for professional LinkedIn.
None of this was news exactly. But having it named clearly — with specific examples pulled from my actual writing — was different from vague self-awareness.
Knowing your patterns explicitly lets you use them intentionally rather than just repeat them.
That’s the capability AI brings to communication analysis. Not replacing voice. Naming it so you can work with it.
VoiceAnalysis AI CommunicationStrategy PersonalBrand LinkedInInsights Solanasis
Post 4 — The November Pattern (Data-Driven Calendar)
Suggested timing: 1 week after Post 1.
Across three completely separate data sources, one pattern showed up consistently:
November is my highest-engagement month on LinkedIn. Every year.
- November 2025: 1,805 ad clicks — more than double any other month in 21 months of data
- November 2025: 268 new connections — highest single month in 18 years
- November 2025: 88 reactions given — peak reaction month on record
I would have told you this was obvious if you’d asked me before I ran the analysis. “Of course Q4 is busy.”
But I would have been guessing.
What I actually do now: schedule the heavy outreach pushes for October/November. Don’t try to force February momentum when three separate data streams show February is historically quiet. Build content banks in September so November can run at full speed.
Data-driven calendar optimization sounds like something a consultant charges $15,000 to recommend.
The actual process was: download the export, run some Python, look at the month-by-month tables, read what my own data said about me.
You already have the data. LinkedIn has been collecting it for years. The analysis is the part most people skip.
What does your peak month look like?
DataDriven LinkedInStrategy GTM ProductivityInsights AI Solanasis FractionalLeadership
Post 5 — The Network Activation Gap
Suggested timing: 10 days after Post 1.
My LinkedIn invitation acceptance rate for outgoing requests is 49.2%.
That’s high. The platform average is closer to 20-30%. It’s a genuine signal of warm network relationships.
And yet.
When I cross-referenced my 2,478 contacts against the seven service lines Solanasis offers — security assessments, disaster recovery, data migrations, CRM, systems integration, responsible AI, fractional leadership — almost none of them had been approached about any of it.
I had built real relationship capital with people who run exactly the organizations Solanasis serves. SMBs. Nonprofits. Colorado-based founders and executive directors. People with warmth scores of 40-50 built from years of genuine interaction.
None of that capital was being converted into conversations.
This is the thing the data made undeniable in a way gut feeling never could. I knew I wasn’t doing enough outreach. Knowing it and seeing it in a table ranked by warmth score and last contact date are different experiences.
The gap between network size and network activation is one of the most common problems in professional services. You spend years building relationships and then don’t use them because asking feels awkward or the timing never feels right.
The data doesn’t care about awkward. It just shows you what’s there.
NetworkActivation ProfessionalServices Outreach CRM GTM Solanasis FractionalCIO
Post 6 — The AI Miscategorization Problem (Tactical Post)
Suggested timing: 12 days after Post 1.
There’s a file in your LinkedIn data export called Ad_Targeting.csv.
It’s tiny — maybe seven rows. It contains the categories LinkedIn’s inference engine has assigned to you. These categories determine:
- Which advertisers can target you
- How LinkedIn positions you in search results
- What kind of content the algorithm surfaces for you
Mine said: HR professional. Interested in technology media. Influences public opinion.
Not cybersecurity. Not fractional executive. Not IT leadership.
I’ve been running a fractional CIO/CSIO/COO firm for a year. The platform had decided I was something else.
This is the miscategorization problem and it matters more than most people realize. If LinkedIn’s AI thinks you’re in HR, it will surface your content to HR audiences, optimize your ad reach toward HR buyers, and deprioritize you in searches from IT buyers or technology decision-makers.
You can’t directly edit these inferences. But you can influence them through profile content, connection patterns, and engagement focus. Knowing what the algorithm thinks is step one.
Most people never check.
Your Ad_Targeting.csv is in your LinkedIn data export. Request the export, wait 24 hours, open the file.
Seven rows. Worth reading.
LinkedIn AlgorithmInsights PersonalBrand DataAnalysis ResponsibleAI FractionalLeadership Solanasis
Post 7 — The Meta Point (Series Closer)
Suggested timing: 2 weeks after Post 1. Closes the series with the “so what for you” frame.
The thing I keep coming back to from this project isn’t any specific number.
It’s how clearly the data showed the gap between my self-perception and my actual professional footprint.
I thought I was actively engaging my network. The data showed I had 2,024 warm followers I’d never reached. I thought my LinkedIn profile was positioned correctly. The algorithm had tagged me as an HR professional. I knew I had strong relationships. The table ranked by last contact date showed people I hadn’t spoken to in 16 months.
The data wasn’t wrong. It was accurate. That was the uncomfortable part.
Here’s the thing: this wasn’t a sophisticated project. It was Python scripts on 37 CSV files I already had. It was an LLM reading text I’d already written. It was connecting data that had been sitting in LinkedIn’s servers for 18 years.
The gap between “having data” and “acting on data” is exactly where most SMBs and nonprofits are stuck. They have CRMs with years of customer data that nobody analyzes. They have email histories, meeting notes, contract records — a picture of their business they’ve never assembled.
This is what responsible AI implementation actually looks like for organizations at that scale: not replacing people, not deploying LLMs into critical systems, but surfacing what’s already in your data so you can make better decisions with it.
The tools are accessible. The methodology is learnable. The data is almost always already there.
If any of this landed, I’m happy to talk through what a similar analysis would look like for your organization’s data. No pitch deck. Just a conversation about what you have and what you could see with it.
ResponsibleAI DataAnalysis SMB Nonprofit FractionalCIO BusinessIntelligence Solanasis
Campaign Notes
Sequencing
- Post 1 (Hook) → Post 2 (Followers) → Post 3 (Voice) → Post 4 (Calendar) → Post 5 (Activation) → Post 6 (Tactical) → Post 7 (Meta/Closer)
- 2-3 days between posts keeps the series visible without flooding
- Total campaign duration: ~2 weeks
Visual suggestions
- Post 1: Screenshot of the Ad_Targeting.csv with the “HR professional” inference visible
- Post 2: A simple bar chart of follower count vs. connection count
- Post 3: A word frequency graphic or quote pulled from the voice analysis
- Post 4: Month-by-month engagement bar chart (November spike visible)
- Posts 5-7: Text-only or subtle graphic; the data speaks for itself by this point
Engagement hooks
- Each post ends with a question or implicit invitation to compare notes
- Avoid direct CTAs until Post 7 (the series closer)
- Respond to every comment in the first 2 hours — LinkedIn algorithm rewards this
What NOT to do
- Don’t lead with Solanasis in every post — the brand mention should be light until Post 7
- Don’t inflate the numbers — the real numbers (2,478 contacts, 37 files, 18 years) are strong enough
- Don’t use “AI” in the first sentence — it reads as hype; earn the AI mention by being specific first