How Lemonade built a $2M/month engine without Google and Meta

Why Reddit and Quora worked when others failed

‍

Jason’s first assignment was to revisit two failed channels: Reddit and Quora. Previous attempts didn’t work, but he had a hunch. Unlike traditional media, these platforms rewarded specificity. So he leaned into what made each one different.

‍

On Reddit, he posted renter’s insurance facts inside subreddits like “Today I Learned.” On Quora, he hunted for questions with clear buying intent. His early creative skipped banner ads in favor of text posts that felt native. He avoided click-through CTAs and instead linked to the product at the end of each post.

These early wins helped him unlock budget and attention internally. Reddit became a top-performing channel. Quora followed close behind.

‍

The scrappy partnership model that scaled to $2M/month

‍

With Reddit and Quora driving consistent returns, Jason expanded into partnerships. He manually searched long-tail blogs, insurance review sites, and affiliate aggregators, using dozens of keyword combinations across Google. For each one, he mapped decision-makers, emailed cold, and negotiated placements.

‍

Eventually, this turned into a comparison engine that consistently ranked Lemonade in the top three results for core search terms. Jason either bought ads, built the partner relationships himself, or helped create the comparison sites from scratch.

‍

One key unlock was search arbitrage. By funding partner spend upfront in exchange for top placement, he kept Lemonade above the fold. Over time, this approach helped dominate search results without relying on owned channels.

‍

Referral programs in regulated industries

‍

Many said referrals couldn’t work in insurance. Too much compliance, too many restrictions. Jason disagreed.

‍

He started small. An email campaign targeted users who had roommates. No incentives, just a nudge: “Let your roommate know.” That one test showed a significant lift.

‍

After seeing that lift, Jason proposed adding a reward. But because of strict compliance rules, he couldn’t offer direct payments. Instead, he tested a version where Lemonade donated $20 to a charity of the user’s choice for each referral. That format fit the brand and earned approval from internal teams. Once the charity version proved successful, he got permission to test a $5 Amazon gift card.

‍

Each change was measured through incrementality. One group got the email, another didn’t. When the lift was clear, Jason lobbied to bring the referral program into the app.

‍

Once referrals became in-app, performance increased again. And by tying landing pages to user IP or name data, Lemonade stayed compliant.

‍

‍

How Jason approached attribution and testing

‍

Over four years, Jason launched more than 200 channels. He did it without models, tooling, or much internal support. His only system was a mental scoring model:

1. How fast can this launch?
2. How much time will it take to maintain?
3. Can it scale if it works?

‍

If a channel couldn’t scale, it didn’t stick. And if attribution couldn’t prove value, he used incrementality instead. Rather than comparing Meta versus Google, he looked at ROI within each channel over time.

‍

Most attribution models overvalue short-term channels and undervalue long-term ones. Jason pushed the team to zoom out. He tracked performance across all spend and optimized based on total return, not per-channel metrics.

‍

Tactics founders can steal today

‍

Pick one overlooked channel and master it. Don’t wait for tools or teams. Launch something simple, measure the lift, and only then invest more. Use incrementality where attribution falls short. And treat partnerships like sales: prospect, pitch, and close.

‍

Most importantly, move fast. Don’t aim for perfect. Just aim for signal.

‍

As Jason put it:

‍

“It wasn’t about launching perfect channels. It was about launching fast and doubling down on anything that moved the needle.”

‍

Founder’s take

‍

“Most marketers overthink. They try to model outcomes before testing anything. Jason did the opposite: launched fast, tracked lift, and only scaled what worked.”

– Jonathan Martinez, GrowthPair Co-Founder

‍