Chapter 3: Finding Your Market

You don't need to guess what the market wants. Social media and experimentation can tell you exactly what resonates.

What You'll Learn

How to use social media as a cheap market sensor
The progression from posts to expertise
Writing for ICs vs. leadership
White-labeling knowledge as your own

If you have two or three different ideas of problems people have, test them:

Make a LinkedIn post
Make a Twitter post
Watch the engagement

It's never the case that all your posts have the same engagement. Two or three might get a thousand views while others get a dozen.

That's the market telling you something.

When everyone is writing about MCPs, clearly there's intent and desire around MCPs. When posts about hiring get engagement, people are thinking about hiring.

How to Test Ideas

Action	Purpose
Write different versions of the same concept	See which framing resonates
Notice which posts get engagement	Identify market interest
Turn high-engagement posts into blog posts	Double down on winners
Let the market guide your expertise	Build authority where demand exists

The Multi-Tweet Test

Write 3-5 different tweets about the same core idea over several days with different hooks. Track engagement carefully—note which get 700 vs. 200,000 views. Use the winning framing in your sales calls and blog posts.

From Posts to Expertise

The pattern is simple:

Post an idea → If engagement, write a blog post → If it resonates, go deeper → Build projects around it

Example Progression

Tweet: "People don't understand precision and recall in RAG"
Engagement: Gets strong response, lots of discussion
Blog post: "Why RAG Is More Than Just Embeddings"
Deeper content: Comprehensive guide to RAG evaluation
Project: Open-source evaluation tool
Consulting: Known as "the RAG evaluation person"

You can slowly become an expert this way. You're not guessing what to specialize in—you're letting the market vote.

Finding What Works

"I would honestly just take all the content you have right now, put it in the cloud, turn it into tiny blog posts and throw it into the Internet. The Internet will just vote on what is important."

This experimental approach works because: - You can test multiple framings of the same idea - Engagement metrics reveal what resonates - High-performing content becomes the foundation for your positioning

White-Labeling Knowledge

Think of content as white-labeling information. It's like Amazon realizing batteries sell well and making Amazon Basics batteries. The information might exist elsewhere, but now it's coming from you.

The Power of "I've Already Thought About This"

The magical moment is when someone asks you a question and you send them a blog post you wrote seven months ago.

When someone says they're thinking about hiring, you reply: "Before we schedule this call, here's three blog posts I've written on the topic."

You didn't get paid to write those posts. But they're getting value from having access to you. They don't have to Google and rediscover your content. That's what's valuable.

An Important Truth

People barely Google problems. You will be surprised at how many people don't actually ask for help. By proactively creating content, you're meeting a need they didn't even know how to express.

Writing for ICs vs. Leadership

Generally, don't overthink the breakdown between technical and high-level content:

If you find yourself talking to leaders, write more for leaders
If you find yourself talking to ICs (individual contributors), write for ICs

The Best Scenario

Your blog posts show up in your client's internal AI Slack channel. The team already believes in you before leadership makes the hiring decision.

Give ICs ammunition to share your content upward. If your technical content is useful enough that an IC thinks "my director should read this," you've won.

Content Strategy by Audience

Audience	Content Type	Goal
ICs (Engineers)	Deep technical content, code examples, prototypes	Build credibility, create sharers
Leadership	Business outcomes, ROI frameworks, strategic insights	Drive purchasing decisions
Both	Problem-focused posts that translate technical → business	Bridge the gap

The Dual-Reaction Sweet Spot

The most successful content often creates dual emotional reactions:

Engineers: "I knew that was the case!"
Experts: "I can't believe people don't understand this!"

This tension generates engagement and positions you as someone who can bridge knowledge gaps.

Example. The post "RAG is more than embeddings" worked because: - Engineers building RAG systems felt validated - Information retrieval researchers were surprised by common misconceptions - Result: 30,000 views on Hacker News, major catalyst for consulting business

Understanding Content TAM

TAM = Total Addressable Market. Balance reach and intent when crafting content:

High Volume, Low Intent

"3 Ways of Building an AI Agent" - Reaches many people - Lower conversion to clients - Good for awareness

Lower Volume, Higher Intent

"3 Ways of Building AI Agents with Python LangChain" - Reaches fewer people - Much higher relevance - Better for qualified leads

As you add specificity, you reach fewer people but with much higher relevance to your services.

Title Writing Tips

Instead of...	Use...	Why
"How I improved RAG"	"How you can improve RAG"	"You" has a larger TAM than "I"
Generic benefits	Specific outcomes	"15% improvement" beats "better results"
Vague timelines	Concrete timelines	"In 2 weeks" beats "quickly"

Content Strategy by Career Stage

Your approach should evolve based on your current situation:

While Employed Full-Time

Focus on purely technical content that builds your audience
Use CTAs focused on audience growth rather than service sales
Leverage your employment as proof: "Here are three things I learned this week"
Consider CTAs that align with employer interests: "We're also hiring"

Transitioning to Consulting

Begin introducing more business-outcome focused content
Shift CTAs toward your services as you prepare to leave
Create content that addresses pain points you can solve
Test different service offerings through content engagement

Full-Time Consulting

Balance technical content (for engineers) with business outcome content (for decision-makers)
Develop content targeting different levels in organizations
Structure content to pre-qualify clients and set expectations about rates

Content Defines Audience

Remember that the content you create defines the audience you attract. Technical how-to posts might get hundreds of thousands of views from learners, while content like "How to write a job description for AI engineers" might attract fewer but more valuable readers who could become clients.

High-Opportunity Markets

Beyond testing content, certain markets offer disproportionate opportunities for AI consultants. Two stand out for their combination of strong demand, healthy budgets, and clear business outcomes.

Private Equity Opportunities

Traditional industries like manufacturing and private equity represent a massive, underserved market. In private equity specifically, demand is insane. These companies know AI isn't their specialty, don't feel embarrassed about hiring external help, have clear business outcomes they want to achieve, and often have significant budgets.

PE firms are particularly attractive because they:

Have portfolio scale: One relationship can lead to dozens of engagements
Understand ROI: They think in terms of multiples and exit valuations
Move quickly: They're used to bringing in outside expertise
Provide warm referrals: Each portfolio company is a pre-qualified buyer

"A private equity company reached out and said, 'We have 25 companies, we all need AI, we're gonna sell those businesses in 7 years, so we know what the multiple is. We'll just give you 30 grand for every implementation for the next 20 companies this year.'"

That's $600,000 in potential revenue from a single relationship. And because PE firms are constantly acquiring new companies, the pipeline keeps growing.

Unlike tech startups where everyone thinks they can build AI in-house, traditional industries understand the value of specialized expertise. They're not trying to become AI companies—they just want AI to improve their operations and increase their exit multiples.

Data Engineering Opportunity

Data engineering is massively important in AI but often overlooked because it's not very sexy. There's a significant gap in the market for consultants who can help companies organize their AI data effectively.

Most AI consultants focus on the flashy parts—model selection, prompt engineering, agent architectures. But the companies actually deploying AI are drowning in data problems:

Unstructured data scattered across systems
No clear data pipelines for model training
Inconsistent labeling and quality issues
LLM conversation data with no good tooling

"It's like a ton of unstructured strings, and there's no real... I haven't seen good tools around that."

LLM conversation data is a particular challenge. Companies are generating massive volumes of user interactions but have no systematic way to analyze them, extract insights, or use them for fine-tuning.

If you can position yourself as someone who helps companies get their data house in order for AI, you're solving a problem that precedes every other AI problem. And you're doing work that most AI consultants consider beneath them—which means less competition.

Action Items

Run the multi-tweet test: Take one idea you have about your expertise. Write 3-5 different versions as tweets/posts over the next week. Track which gets the most engagement.
Audit your content TAM: Look at your recent posts. Are they high-volume/low-intent or lower-volume/high-intent? Experiment with more specific content.
Create your "white label" library: List 10 questions you get asked repeatedly. These are your first 10 blog posts.
Map your dual audiences: Identify what content would resonate with ICs and what would resonate with leadership. Plan to create both.

Key Takeaways

Social media is a free research tool—use it to test ideas before investing in long-form content
You don't need to guess your niche; let the market vote through engagement
White-label existing knowledge under your brand—make it easy to reference your expertise
Write for both ICs (to build grassroots credibility) and leadership (to drive buying decisions)
Balance content reach (TAM) with intent—specific content converts better than generic

Next: Chapter 4: The Content Flywheel →