In May 2013, a former WordPress designer named John O'Nolan put a blogging platform on Kickstarter. The software did not really exist yet. In a Founder Chats interview with Baremetrics, O'Nolan describes what backers were actually looking at as version 0.1 with mocked-up interface elements, an "absolutely pre-MVP version."
Backers funded it anyway. O'Nolan says the campaign raised GBP 196,362 from 5,236 people. TechCrunch, covering it on 7 May 2013, noted simply that it was on Kickstarter and fully funded.
Thirteen years later, Ghost's about page carries a live dashboard rather than a mission statement. At the time of checking it showed annual recurring revenue of $10,947,413, a monthly run rate of $912,284, net churn of 3.05%, and 30,493 active customers. There are no investors to report to. There cannot be: the company was set up so that no one owns it.
Disclosure before we go further: Builtplain runs on Ghost. We pay for it like any other customer and have no other relationship with the company.
Decision one: sell the idea before building the thing
Kickstarter did three jobs at once here, and only one of them was money.
It proved demand. Five thousand people paying for software that did not exist is a stronger signal than any survey, and it arrived before a year of engineering was spent.
It created a deadline and an audience simultaneously. Every backer was a person who had publicly committed to caring whether the thing shipped.
And it set the relationship. Ghost's first money came from the people who would use it, not from someone expecting a return. That single fact shaped everything downstream.
One tactical detail is worth stealing. O'Nolan says the campaign deliberately avoided physical rewards, so no one had to spend months shipping t-shirts instead of writing code. The top tier was a GBP 5,000 corporate sponsorship with eight slots. If you run a crowdfunding campaign for software, the fulfilment burden of merchandise is a real tax on the runway you just raised.
Decision two: make the company unsellable on purpose
Ghost is structured as a non-profit foundation. The reason, stated on its own about page: "We set Ghost up as non-profit foundation so that it would always be true to its users, rather than shareholders or investors."
O'Nolan gives three motivations in the Baremetrics interview. He wanted to spend his working life on open-source code. He had watched the split between the WordPress foundation and the for-profit Automattic and did not want that structure. And he admired Mozilla's model.
The practical effect is that no acquisition can happen, no investor can force a strategy change, and no future board can decide that the licence should get more restrictive. For a platform that other people build their businesses on, that is not a philosophical nicety. It is a product feature. Anyone who has watched a tool they depend on get acquired and repriced knows exactly what it is worth, which is the same reason we tell readers to weigh platform fees against portability before committing an audience.
O'Nolan is also clear about the cost, and quoting him here matters because the structure is often romanticised: "You take all the stresses and pressures of a regular business and you pile on top all of the restrictions of a non-profit." He calls it incredibly hard. There is no fundraising option when cash is tight, no equity to hand to a key hire, and no exit that ever pays the founders a lump sum.
Decision three: give the software away and sell the hosting
Ghost's revenue model is the standard open-source split done properly: the software is free and self-hostable, and the company sells Ghost(Pro), managed hosting for people who would rather not run a server.
The reason this works is that it separates two different customers who are usually forced into one product. A developer who wants to run their own instance costs the company almost nothing and produces something valuable anyway: installations, bug reports, plugins, and the 54,437 GitHub stars the about page reports. A publisher who wants to write rather than administer servers pays monthly.
The free tier is not a marketing funnel here; it is genuinely free forever, for anyone with a server. What the company sells is the removal of operational work. That is a cleaner value proposition than a feature-limited free plan, and it is why the open-source version acts as distribution rather than cannibalisation. The same dynamic explains why building an audience through what you publish beats buying attention: the free artefact is the marketing.
Decision four: publish the numbers in public
Most companies bury their metrics. Ghost puts ARR, run rate, churn, and customer count on the front of its about page, live.
This is a deliberate trade. It removes the option of quietly having a bad year. It also does something no press release achieves: it lets a prospective customer verify that the company is solvent before trusting it with their publication. Alongside 100,000,000 or more total installs and 9 billion requests a month, the numbers say the thing that matters to a publisher choosing where to build: we will still be here.
Note what the dashboard does not show. There is no profit figure, no salary data, no cash reserves. Transparency is a choice about which numbers, and Ghost has chosen the ones that answer the customer's question rather than the ones that answer a competitor's.
What a smaller builder can take from this
The structure is not the lesson. Almost nobody should set up a non-profit foundation, and O'Nolan himself describes the difficulty in detail.
The lesson is that the funding source you choose defines the company you are allowed to build afterwards. Ghost took money from users and therefore owes users a product. A venture-funded competitor takes money from investors and therefore owes investors a return, which eventually means pricing changes, acquisition, or growth at a rate the product may not support. Neither is wrong. They are different obligations, and they are chosen once, early, usually without much thought.
Three things here are directly copyable at any size: presell the concept to the people who will use it, before you build it; when you crowdfund software, refuse rewards that create fulfilment work; and publish at least one number that would embarrass you if it went the wrong way, because it is the cheapest trust you will ever buy. For solo builders weighing what is now possible without outside money, our piece on the one-person software company covers the ceiling and where it sits.
What is not copyable: 2013 Kickstarter attention for a developer tool, and a founder with an existing reputation inside the WordPress community, which is where the first backers came from. The campaign was not discovered by strangers. It was funded by a community O'Nolan had already spent years in.
Figures are from Ghost's live public dashboard, John O'Nolan's Baremetrics Founder Chats interview, and TechCrunch's May 2013 report, checked July 25, 2026. The dashboard figures change continuously.
Discussion
Sign in with Google or just a name. No email link, no password to remember.