When years pass without traction, shrink back to the founding team and rewrite from scratch, then define activation by the collaborative behaviors that retain users rather than raw signups.
Notion
Notion is an all-in-one workspace that blends notes, docs, wikis, and databases into a single tool teams use to organize their work. Its story runs from a near-death rewrite that finally produced product-market fit to a growth machine built on getting one user to pull in the next, where activation was defined not by signups but by the collaborative behaviors that actually retain people.
PMF: The Kyoto rewrite and the "lost years"
The problem. Notion's first three to four years produced no product-market fit, little growth, and periods with barely enough cash to continue, a stretch co-founder Ivan Zhao calls the "lost years." The early product chased a lofty no-code vision built on Google's early Web Components tech, and it was unstable, crashed constantly, and did not scale. By 2015, with angel money running low, Zhao faced the choice of firing his small team or running out of money.
The approach. Zhao and co-founder Simon Last threw away the codebase, shrank the team back to just the two of them, sublet their San Francisco office, and moved to Kyoto to cut costs and rebuild in isolation. They also reframed the vision: instead of selling people on building their own tools, they hid that ambition inside something people already wanted, familiar notes and docs, and layered the power features in gradually. As Zhao put it, "We focused too much on what we wanted to bring to the world. We needed to pay attention to what the world wanted from us."
How it solved it. The rebuilt product shipped as version 1.0 in 2016, and the "notes and docs with a twist" framing gave users an obvious reason to start. The 2018 relaunch of version 2.0 on Product Hunt accelerated adoption, and Notion went on to reach 1 million users on only seed funding and a team of roughly 18 people, evidence the pivot had finally landed on something the market pulled for rather than something the founders pushed.
Virality: One comma standing between a user and the invite loop
The problem. Notion's growth depends on a loop: a new user invites teammates into a shared workspace, and those teammates become new users who invite still more. The final step of onboarding is the invite screen, Notion's first real referral touchpoint, so any friction there quietly throttles the entire loop. Walking the flow as a fresh signup, a growth team member found a tiny but load-bearing blocker: when adding more than one email, it was unclear whether you needed a comma between addresses before typing the next one.
The approach. Rather than treating onboarding as merely getting a user through the door, Notion fixed the friction at the exact point of activation, and it defined activation by the behaviors that predict retention: a fully onboarded user is someone who invites at least two other people or creates a meaningful amount of content. Getting a user signed up was explicitly not counted as success.
How it solved it. After the invite input was fixed, the number of people who invited someone else grew by 9x. Because that single step feeds the loop, more successful invites meant more new users who could themselves invite and create, and tying the activation definition to inviting two others or creating content kept Notion measuring genuine engagement instead of vanity signups.