Topic 1 — How tech companies work

Goal: Understand what a tech company is and how it actually makes money.

Lesson 1.1 — A tech company is still a business

Teodora Vrabie — Teo to everyone — taught high-school geography for eight years before deciding to break into tech. On her first evening of this track she typed a worry into her notebook: I don't know how any of these companies actually work. Facebook, Stripe, the study-planner app her old school used — they felt like a different species of organization, run by rules she'd never been taught.

Here's the reframe that calmed her down, and where this topic starts. A tech company is a normal business. It solves a problem real people will pay to fix, finds those people and turns them into customers, earns more than it spends, and grows. Every bakery and airline lives by that same loop. The "tech" — almost always software — is the how, not a secret rulebook.

Teo's eventual target is Brightwell, a 40-person edtech startup that sells a subscription study-planner app to schools and students. It has investors, payroll, a sales team, customers who cancel, and a CEO doing the math on whether money in beats money out. Strip away the laptops and a shopkeeper from a hundred years ago would recognize it. A tech company makes its money with software instead of bricks; the business questions underneath are the ones every business has always answered.

You don't need an MBA to read one of these companies. You ask what you'd ask of any business — who's the customer, what problem, how does the money come in? — and learn the handful of patterns the answers follow. That's the rest of this topic.

Lesson 1.2 — Build once, sell many

Teo's brother builds furniture. Every table he sells, he buys more wood, more hours, more glue. Selling a hundredth table costs him almost exactly what the first one did. That extra cost to make and deliver one more unit has a name: marginal cost.

Now look at Brightwell's app. The engineers build the study planner once. The thousandth school downloads the same software the first school did — no new wood, no new factory shift. The cost of serving one more customer is close to zero. That's near-zero marginal cost, and it's the single fact that makes software behave so differently from her brother's workshop.

This is the "build once, sell many" property. Spend a lot up front to build the product, then sell it to ten, ten thousand, or ten million customers while your costs barely move.

That one property explains a lot:

• Why a tech company can lose money for years (paying to build), then turn wildly profitable once enough customers are on the same product.
• Why investors pour cash into a startup that isn't profitable yet — they're betting the "sell many" half is still coming.
• Why the industry obsesses over growth: when each new customer costs almost nothing to serve but pays full price, every customer you add drops mostly to the bottom line.

Teo wrote it plainly in her notebook: make it once, sell it a million times. That sentence, she realized, was most of what made tech feel like magic to outsiders. There's no magic — just marginal cost near zero.

Lesson 1.3 — SaaS: the rent-don't-buy model

Teo already lives inside this model and never noticed. She pays Netflix every month, Spotify every month. She doesn't own either one; she rents access, and the day she stops paying, the access stops. Brightwell's schools do the same with the study planner — a yearly per-school subscription, renewed or not.

That pattern is SaaS — Software as a Service. Instead of buying software once and owning it forever, the customer pays a recurring fee (monthly or yearly) to keep using it. SaaS is the dominant model in modern tech, and you'll hear the three letters constantly. Slack, Spotify, Netflix, and Brightwell are all SaaS.

Why does the industry love it so much? One word: predictable. If Brightwell signs 200 schools onto annual plans, it can forecast that revenue before the year starts, the way a landlord knows the rent is coming — instead of having to win a brand-new sale every month just to stand still. And because of "build once, sell many," it stacks: this year's customers mostly stay and you add new ones on top.

Bao Trinh, a senior product manager at Brightwell who became Teo's informal mentor, put it to her over coffee like this:

"The dream is boring on purpose. Customers who pay us every year without us having to re-sell them. Once you see SaaS as predictable rent, the rest of how we behave makes sense."

That predictability is exactly why a SaaS company guards its existing customers so fiercely — a theme that comes back in Lesson 1.5.

Lesson 1.4 — The other ways software makes money

SaaS isn't the only game. Reading job ads, Teo kept hitting companies that clearly weren't charging a subscription, and she wanted to name what they were doing. Four more patterns cover almost everything she'll meet:

• Transactional / marketplace. The company connects two sides and takes a cut of each deal. Uber takes a slice of every ride, Airbnb of every booking, eBay of every sale. No transaction, no revenue — so they obsess over transaction volume.
• Advertising. The product is free, and advertisers pay. Google and Meta make most of their money this way. The catch Teo learned to spot: when you don't pay, you (your attention and data) are the product sold to advertisers.
• One-time license (perpetual). Pay once, own that version forever — the older model, like boxed desktop software. Much of the industry has migrated toward subscriptions, but it hasn't vanished.
• Usage-based / metered. You pay for exactly what you consume, like electricity. AWS bills by the hour and gigabyte; the OpenAI API bills per chunk of text processed. Light users pay little, heavy users pay a lot.

Model	Who pays	One example
SaaS subscription	the user, recurring	Netflix, Slack, Brightwell
Marketplace / transactional	a cut per deal	Uber, Airbnb, eBay
Advertising	advertisers (user is free)	Google, Meta
One-time license	the user, once	older desktop software
Usage-based / metered	per unit consumed	AWS, OpenAI API

A company can run more than one of these at once, and many do. But naming the main one is the fastest way to understand what a company really cares about — which is exactly where Lesson 1.5 goes.

Lesson 1.5 — Freemium, and following the money

When Teo signed up for SkilsMVP, the platform she's reading right now, she paid nothing. A free tier let her start. Some lessons sit behind a paid upgrade. She's living inside freemium: a free version pulls lots of people in the door, and a paid upgrade converts a small slice of them — often just a single-digit percentage, commonly in the 2-5% range. Spotify Free nudging you toward Premium is the same move.

One thing to get straight, because it trips people up: freemium is a tactic, not a fifth business model. It's a way to acquire customers that sits on top of SaaS — the free tier is the storefront, the paid subscription is still where the money is.

Once she could name the model, Teo could play the game that ties this topic together. Bao calls it "follow the money." How a company earns tells you what it cares about, and what every job inside it is measured against.

• An advertising business (Google, Meta) lives on attention, so it obsesses over engagement — time on site, clicks, daily users. Roles there are judged on keeping you scrolling.
• A SaaS business like Brightwell lives on recurring subscriptions, so it obsesses over two numbers:
  • Churn — the percentage of subscribers who cancel in a given period. (As a rough benchmark, strong SaaS companies keep monthly churn under about 2% — but the healthy range depends on who the customer is: enterprise SaaS often runs under 1%, while SMB and consumer apps can sit at 3-5% and still be fine, so don't quote 2% as a universal law.) Every cancellation is a leak in the bucket.
  • Growing subscriptions — new schools signing, existing ones renewing and expanding.

Three more words every SaaS company says constantly. MRR / ARR — Monthly and Annual Recurring Revenue, the predictable subscription money added up. LTV — Customer Lifetime Value, the total a customer pays before they leave. CAC — Customer Acquisition Cost, what you spend to win one customer. The model only works if LTV is comfortably bigger than CAC — a customer must be worth more than you paid to get them, with a common rule of thumb of at least 3 times CAC. If Brightwell spends more to land a school than the school ever pays, no amount of clever software saves it.

Tell me how a company makes money and I'll tell you what it worries about at 9am.

That's the payoff of "follow the money," and it works on any company you'll ever read.

Worked example — Reading Brightwell like a pro

Renske Aldous, Brightwell's recruiter, opens Teo's informational call with a softball: "So — do you actually understand what we do here?" A month ago Teo would have said "an education app." Now she walks it like a business.

Who's the customer and what problem? Schools and students who need to plan study time; Brightwell sells them an app that organizes it. A real problem, people who'll pay to fix it.

What's the model? A recurring per-school subscription — SaaS — so the money is predictable: signed schools pay year after year. A free student tier nudges toward a paid upgrade — freemium layered on top to fill the funnel.

Build once, sell many? Yes. The engineers built one planner; the 300th school runs the same software as the first at near-zero extra cost. That's why Brightwell can add schools without ballooning costs — and why investors bet on growth.

Follow the money — what does Brightwell worry about? Two things above all: churn (schools that don't renew — a leak in predictable revenue) and growth (new and expanding schools). And it only survives if each school is worth more than it cost to win — LTV above CAC.

Then Renske asks the real question: "Why would someone with no tech background be useful here?" Teo has the answer now. Marisol Ferreira, Brightwell's founder and an ex-teacher herself, started the company because she understood the customer's problem in her bones. Davor Halász, a career-changer who pivoted from accounting into a junior data-analyst seat, got hired because he could read these same numbers — churn, MRR, LTV — and explain what they meant.

Every role at Brightwell ladders up to one chain: solve a customer problem → win and keep paying customers → grow. The product manager decides what to build to grow subscriptions. Sales and marketing win the schools. Customer success fights churn so schools renew. Designers, engineers, and QA make the product good enough that schools stay. The analyst measures all of it.

Teo can now read any tech company — already more business sense, Renske tells her, than half the candidates who walk in.