Growth Through Rapid Experiments.

MasterClass started almost by accident. David Geiger had a hunch: maybe people would pay to learn from top experts. But he had no resources—just curiosity and a willingness to experiment.

He built a fake website listing a hundred courses he hadn’t yet created, bought ads, and waited. Surprisingly, people signed up—lots of them. Feeling guilty, he sent out gift cards and set up a 1-800 number that rang directly to his cellphone.

When Geiger showed this early traction to his professor—venture capitalist Michael Dearing—Dearing immediately recognized its potential and wrote a check for $50,000 on the spot.

That’s how MasterClass began—not from intricate business plans, but from curiosity, a simple experiment, and the boldness to follow wherever it led.

This approach is called a “fake door test,” where you measure interest in an unbuilt feature or product. David Geiger wasn’t alone; countless successful (and unsuccessful) startups have done this. You never really know what people want until you try to sell it to them.

Your riskiest assumption often holds your biggest opportunity—test it first. Every human creation begins as an experiment. Some think you need knowledge to test things; I say knowledge isn’t free—it’s bought with curiosity.

Initially, doing something that a big company can’t sets your startup apart. This unique ability to experiment is exactly what drives growth. Each experiment offers rewards—perhaps insight, success, or even a valuable lesson in what not to do. Whatever you learn fuels your company for years to come.

I’ve seen experienced founders treat growth experiments as unrewarded risks—risks they’re unwilling to take. Often, their fear of failure threatens their credibility. Eventually, they lose momentum compared to startups willing to embrace these risks.

Experiments equal momentum. To get a startup off the ground, you must keep experimenting—building momentum—until you finally break through the clouds. Every new tactic you test—whether onboarding users, keeping existing ones happy, or uncovering new revenue streams—contributes to your innovation. Experiments reshape your idea into something better, open new revenue channels, or even reveal improved ways of running your business.

• Airbnb began by renting air mattresses in the founders’ apartments. The experiment wasn’t glamorous, but it proved people would pay to sleep in strangers’ homes, unlocking an entirely new market.

• Slack initially tried building an online game that flopped. However, the internal messaging tool they developed became their real product, transforming a failed experiment into billion-dollar success.

• Instagram first launched as “Burbn,” a complicated check-in app nobody wanted. Simplifying it to just photo-sharing transformed it into the platform we know today.

• Netflix tested mailing DVDs when everyone else relied on physical stores. Later, another experiment—streaming video—changed their trajectory again.

Each small experiment built momentum until these startups broke through. Momentum isn’t luck; it’s the cumulative effect of constant experimentation.

Some of the most successful startups are known for small experiments they conducted in their early years—things that seemed impossible to scale but dramatically improved user experience. Surprisingly, they found ways to scale these seemingly unscalable ideas, differentiating them in the market:

• Spotify’s “Discover Weekly” started as a tiny internal hackathon project to test personalized playlists. Initially manual and seemingly unscalable, it became the core feature users loved most. Today, Spotify is famous for personalized recommendations.

• Dropbox’s viral referral program began as just an experiment—giving extra free storage in exchange for inviting friends. It was uncertain if it could sustain growth, yet Dropbox scaled this approach, becoming a textbook example of viral growth.

These startups stood out precisely because they dared to start with the unscalable. The effort put into those early experiments paid off in lasting differentiation. These experimental hacks eventually became their identity.

When I worked with an edtech platform, we encountered a common problem: our webinar funnel depended entirely on paid ads, and customer acquisition cost (CAC) kept rising. We tried typical fixes—optimizing ads, tweaking the funnel—but nothing moved the needle.

What finally worked wasn’t a better ad strategy; it was something that didn’t seem scalable.

One day, a student reached out, asking if he could have a quick call with a mentor. It wasn’t part of our process, but I set it up anyway. The call lasted just five minutes. That was enough—he became excited to join. This made me realize something important: people don’t just want a course; they want reassurance that the course is right for them.

So, we tried an experiment. Instead of just allowing students to attend the webinar, we asked them to fill out a form to receive personalized career guidance. Writing hundreds of these by hand wasn’t feasible, so we grouped students based on skill level, experience, and common struggles. Working with a mentor, we created several versions of a personalized letter—specific enough to feel personal, yet broad enough to cover recurring patterns. We used AI to tweak these messages, ensuring no two emails felt identical.

The results surprised us. Students who received these emails were more engaged, more confident in their decisions, and more likely to sign up. It wasn’t because we tricked them into buying—it was because they received something valuable before paying.

I’ve noticed this pattern frequently. Things that work best usually don’t appear scalable initially. But they create the trust and momentum that simplify everything else.

Experiments provide insights that no external expert can match. Even failed experiments put you ten steps ahead, offering new perspectives on your solution. Every experiment unlocks possibilities, and good founders know which opportunities to pursue and which to avoid—considering the opportunity cost. Not every possibility you uncover is worth chasing.

The best way to measure an experiment’s success is by working backward from your customers. Many founders waste months building features no one asked for and spend heavily on launch day. Instead of falling in love with your ideas, figure out exactly what customers need first, then build precisely that.

Every experiment should be data-driven and closely tracked. Early-stage assumptions and guessing what users want are fatal—because nothing is costlier than spending time on the wrong things.

I’ve seen startups waste months chasing random growth experiments, thinking they’re building momentum. Usually, these experiments go nowhere—no insights, no traction, just confusion about what they were testing.

Good experiments aren’t random; they’re measured precisely at every step. Each data point should teach something concrete. Real insights come from data—not assumptions or guesswork.