SaaS Companies That Started as Side Projects

Many software companies started as side projects. Their founders built the first version after work or between client jobs and kept their salaries until the product paid enough to replace them. BuiltWith, Baremetrics, Gumroad, ConvertKit, Plausible and Replit all began that way, and their early years share a few patterns.

At a glance

Company	Started	Built alongside	Where it went
BuiltWith	2007	A full-time job	Founder went full time in 2013
Baremetrics	2013	Client work and two small products	Sold for $4 million in 2020
Gumroad	2011	A job at Pinterest	Raised about $8 million, then rebuilt as a profitable small company
ConvertKit	2013	The founder’s own book business	Renamed Kit in 2024
Plausible	2018	A developer job	Passed $1 million a year in 2022, without investors
Replit	2011	Jobs at Yahoo, Codecademy and Facebook	Valued at $9 billion in 2026

Figures come from the full stories, where each one is sourced.

Why software suits a side project

Software is the easiest kind of business to start in the evenings. The first version needs a computer, a domain and the founder’s time, and almost nothing else. There is no stock to buy, no premises to rent and, at the start, nobody to hire.

That low cost has a second effect that matters more. A software side project can sit online for months, earning a little or nothing, without costing its founder much more than the hours they put in. Several of the companies in this group spent years in that state. A business with rent or inventory would have had to grow or close long before.

The same low cost explains why so many software side projects fail quietly. Nothing forces a decision, so a project can drift for years. The stories here are the ones where something eventually did: a customer asked to pay, revenue passed a salary, or a co-founder arrived.

The stories in this group

BuiltWith. Gary Brewer had the idea on a Sunday afternoon in July 2007 and registered the domain the same day. He built the tool, which shows the technology behind any website, in his evenings while holding a full-time job in Sydney. Customers soon asked for lists of sites using particular technologies, and those paid lists became the business. By 2011 BuiltWith was making around $40,000 a month while Brewer was still employed, and he did not go full time until 2013. Read the full story: BuiltWith: The Evening Side Project That Took Six Years.

Baremetrics. Josh Pigford was running client work and two small products when he built a Stripe analytics tool for himself in about eight working days in 2013. The first customer paid $250 a month. Within a year Baremetrics earned more than his other products combined, and in 2020 he sold it for $4 million. Read the full story: Baremetrics: An 8-Day Side Project That Sold for $4 Million.

Gumroad. Sahil Lavingia built the first version of Gumroad in a single weekend in 2011, while he was an early employee at Pinterest. He left Pinterest a few months later, raised about $8 million from investors and then, after layoffs in 2015, rebuilt the company as a smaller, profitable business. Read the full story: Gumroad: The Weekend Project of Pinterest’s Second Employee.

ConvertKit. Nathan Barry started ConvertKit in 2013 next to his own book and course business, after setting himself the goal of $5,000 a month in six months. Twenty-two months later it was making about $1,300 a month, and he decided to go all in rather than shut it down. The company was renamed Kit in 2024. Read the full story: ConvertKit: The Side Project Nathan Barry Almost Shut Down.

Plausible. Uku Täht wrote the first code for Plausible Analytics around Christmas 2018 while working as a developer. Its first paid month brought in $64. Marko Saric joined in 2020 to handle marketing, both went full time, and by June 2022 the business had passed $1 million a year without outside funding. Read the full story: Plausible: The Side Project That Took On Google Analytics.

Replit. Amjad Masad released an open source tool for running code in the browser in 2011, while working at Yahoo in Jordan. It reached the top of Hacker News and got him hired at Codecademy in New York. He founded Replit as a company in 2016 with Haya Odeh and Faris Masad, and it was valued at $9 billion in 2026. Read the full story: Replit: The Open Source Side Project That Took Ten Years.

What these companies have in common

Read together, the six stories share four patterns.

The founder was the first user. Pigford built Baremetrics because he needed Stripe metrics for his own products. Barry built ConvertKit to sell his own books. Brewer wanted to know what sites were built with. A side project that solves the founder’s own problem has at least one committed user from day one, and that user knows exactly what is missing.

Revenue came before the decision to quit. None of these founders left a job to build an idea. They left once the side project was already earning, or, in Lavingia’s case, once it had traction and investors. BuiltWith was making $40,000 a month two years before Brewer went full time.

Customers changed the product. BuiltWith started as a free lookup tool; customers asking for lists turned it into a data business. Baremetrics launched as an internal tool and found its market when other Stripe users asked for it. The first version was rarely the business.

The slow years were longer than expected. ConvertKit’s original target arrived about two years late. Plausible earned $64 in its first paid month. In most of these stories there is a stretch where the project could reasonably have been abandoned.

How long the side-project years lasted

The gap between the first version and the founder going full time varies widely:

Gumroad: a few months, from the April 2011 weekend build to Lavingia leaving Pinterest.
Baremetrics: within about a year, Baremetrics became Pigford’s main product.
Plausible: about fifteen months, from the end of 2018 to both founders going full time in 2020.
ConvertKit: about twenty-two months before Barry went all in.
BuiltWith: about six years, from 2007 to 2013.
Replit: about five years from the open source project to the company.

There is no standard length. What decided it in each case was a specific event: revenue passing a salary, an investor, a co-founder, or a deadline the founder set himself.

Funding: with investors and without

This group includes both routes. Gumroad and Replit raised venture capital; BuiltWith, ConvertKit and Plausible did not; Baremetrics raised a small round of $500,000 before it was sold.

The side-project start does not decide which route a company takes. It does change the terms. A founder who has already proved demand in the evenings can raise money on better terms, or decide not to raise at all. Lavingia later wrote about the costs of the venture route for Gumroad, while Barry and the Plausible founders chose to grow only on revenue.

If you are building one

These six stories suggest a few practical rules for anyone building software in the evenings.

Build for a problem you have today. Every founder here could test the product on their own work the day it was built. That shortens the gap between idea and first real use to almost nothing, and it keeps a side project alive through the months when nobody else is paying.

Charge early, even a little. Baremetrics charged its first customer $250 a month at launch. Plausible’s first paid month brought in $64. A small amount of revenue is a stronger signal than any number of free sign-ups, because it shows someone values the product enough to pay for it.

Decide in advance what would make you quit, or stop. Barry gave himself a target and a deadline; when he missed it, he had to choose between shutting ConvertKit down and going all in. Without a line like that, a software side project can drift for years.

Watch for the request you did not plan for. BuiltWith’s real business arrived when customers asked for something it did not yet sell. The most useful thing early users give a side project is often a request, not a payment.

Protect the hours. None of these stories was built in a burst of inspiration. They were built in steady evenings over months or years. A fixed slot each week matters more than a big weekend.

Frequently asked questions

Which famous software companies started as side projects?

Among the companies covered here, BuiltWith, Baremetrics, Gumroad, ConvertKit (now Kit), Plausible and Replit all began as side projects built alongside a job, a degree or another business. GitHub, Mailchimp, Zapier, Buffer and Salesforce follow the same pattern and will be covered in this group.

How long does a software side project take to replace a salary?

In these stories it ranged from a few months to about six years. Baremetrics and Gumroad moved fast; BuiltWith took six years even though it was making around $40,000 a month two years before its founder quit. The deciding factor was usually a specific trigger, not a fixed timeline.

Do you need investors to grow a side project into a software company?

No. BuiltWith, ConvertKit and Plausible grew on revenue alone, while Gumroad and Replit raised venture capital. Starting as a side project gives a founder proof of demand before choosing a route, which is useful either way.

What kind of software side projects tend to work?

In this group, the strongest starts came from founders building a tool they needed themselves, such as Stripe metrics or a way to sell their own books. Having a real first user, even if it is the founder, made it easier to know what to build next.