I learned SaaS from scratch

I didn't know SaaS or AI when I started SocialPost.ai in April 2023.

That is not false modesty. I could not have told you what a healthy activation rate looked like, how a self-serve funnel should be instrumented, or why a landing page converts at one number instead of another. I was starting from zero on the exact skills the business needed most.

What I did know was enterprise IT. Mainframes. Microsoft licensing. M&A. Almost thirty years of it. And here is the uncomfortable truth I ran into within the first month: none of that helps you run an AI SaaS for small businesses.

Not some of it. None of it.

The resume that didn't transfer

My career started in 1996 at Perot Systems. From there I went to Microsoft Consulting Services in New York, doing mainframe integration work inside Merrill Lynch, Goldman Sachs, and Lehman Brothers. Big systems, big contracts, big companies. Later I spent two decades running an enterprise services firm.

That background teaches you real things. How to close a seven-figure deal over eighteen months. How to manage a delivery team. How to read a Microsoft licensing agreement without falling asleep.

It teaches you nothing about a customer who signs up at 11pm, pokes around for six minutes, and either gets value or leaves forever. Enterprise IT is a relationship business with long cycles and forgiving buyers. Self-serve SaaS is a product business with anonymous users and zero patience. The instincts you build in one actively mislead you in the other.

I had a choice at that point. Pretend my experience covered it, or admit I was a beginner and act like one.

I chose beginner.

Thirty books before thirty features

So I brute-forced it. I read more than 30 books, and I read them like an operator, not a student. Five subjects:

• SaaS metrics
• Go-to-market
• Funnels
• Email
• SEO

Not skimming. Working through them, taking notes, arguing with the authors in the margins, then testing what they claimed against what I was seeing in my own signup flow.

I also went back to school, literally. I completed Wharton Executive Education's "Scaling Your Startup" program in 2023, in the middle of building. Some people found that funny. Why is a guy who has been building companies since the nineties sitting in a scaling course? Because the course was about a kind of scaling I had never done. Ego is expensive. Tuition is cheap.

The books did something specific for me. They gave me vocabulary and a mental map. Before them, I did not know what I did not know. After them, I could at least name every part of the machine: acquisition, activation, retention, expansion, the whole chain. I still could not operate every part well. But I knew what each part was for, and I knew what a broken one looked like.

That distinction matters. You do not read thirty books to become an expert. You read them so that experts can no longer confuse you.

Learning the tools out of necessity

Then came the tools. HubSpot. Stripe. Mixpanel. Jira. Figma. GitHub.

I learned each one the same way: not deeply, enough to operate. I can build a workflow in HubSpot but I am not a HubSpot admin. I can read a Mixpanel funnel and spot where users fall out, but I am not an analytics engineer. I can navigate Figma well enough to review our lead designer's work and give real feedback, but I do not design.

This drove some people around me crazy. The advice you always hear is "focus on what you're great at and delegate the rest." That advice is right at 50 employees. It is wrong at two.

Here is why. Early on, every tool in your stack is a window into your business. Stripe shows you the money. Mixpanel shows you the behavior. HubSpot shows you the pipeline. Jira shows you the velocity. If you cannot open those windows yourself, you are running the company on secondhand descriptions of your own business. Every question becomes a meeting. Every meeting adds days. At the earliest stage, days are the whole game.

The hiring trap I refused to walk into

I could not justify hiring experts early. But even that undersells the real reason I didn't.

If I had hired a growth expert in month two, I would not have understood what they were doing. Which means I could not have evaluated whether they were doing it well. Which means I would have been paying senior rates to be managed by my own employee.

I have watched this exact failure play out at other companies. A founder who does not understand funnels hires a "funnel person." The person produces dashboards and jargon. The founder nods. Eighteen months later there is a beautiful reporting stack and no growth, and the founder cannot even diagnose why, because they outsourced the understanding along with the work.

You can delegate work. You cannot delegate understanding. Not at the start.

The order of operations that actually works is this: learn the domain to functional competence, do the work badly yourself, learn what good looks like from your own mistakes, then hire someone better than you and know exactly why they are better. We run that way at SocialPost.ai today. The company is venture-backed and profitable, serving more than 15,000 users, and every specialist we have brought on came in after I had personally done their job badly first. I can talk to our AI PM about the product in specifics, not vibes, because I spent months in the tooling myself.

Why a veteran operator still had to start over

I want to be honest about the psychology here, because it is the part most experienced people get wrong.

When you have decades of success in one domain, your brain quietly tells you that competence transfers. It whispers that you are above the beginner phase. That the books are for kids. That the Wharton course is beneath you.

That whisper has killed more second acts than any market ever has. I have seen accomplished executives start companies and fail, not because they were not smart, but because they refused to be new at something. They kept reaching for the old playbook because holding it felt better than holding nothing.

The mainframe work at Merrill Lynch made me a strong systems thinker. It did not make me a SaaS operator. Those are different jobs, and pretending otherwise would have cost me the company.

The rule I'd give any founder entering a new domain

Here is the framework, usable this week:

List every function your business needs: product, growth, funnels, email, analytics, billing, support. Write them all down, even the ones you dread.

For each one, get to "functional understanding": you can do the basic version yourself, and you can tell competent work from theater.

Read three serious books per function before you touch the tool. The vocabulary compounds. You will move twice as fast inside the software once you know what the screens mean.

Learn every tool to operating level, not mastery. Enough to answer your own questions without a meeting.

Only hire depth once you can interview for it. If you cannot ask a candidate one question that a faker would miss, you are not ready to hire that role.

You don't need mastery at the start. You need functional understanding across the entire stack, held in one head, yours. Mastery is what you hire later, once you have earned the ability to recognize it.

Not expertise first. Literacy first.

Your turn: What tool or discipline did you have to learn the hard way, and what finally made it click?