Buffer went live on November 30, 2010. According to Joel Gascoigne's ten-year retrospective, the first paying customer arrived within four days. In an earlier essay on the same launch he says three. Either way, a product built in about seven weeks of evenings and weekends was taking money almost immediately, and by the time of that 2013 essay Buffer had 16 employees and roughly $2.5 million in annual revenue.
The part everyone repeats is what came before the code: a landing page that tested the idea before it existed. The part almost everyone misses is what Gascoigne was actually measuring. Get that wrong and the test tells you nothing, which is why so many founders run it and learn nothing.
Myth one: he was collecting email addresses
The common version of this story is "put up a landing page, collect emails, if you get enough, build it." Gascoigne is explicit that this was not the goal. "I wasn't optimizing for the number of signups I could get with this landing page, I was instead trying to learn," he wrote, and elsewhere in the same essay: "Do you want a billion signups, or do you want validated learning?"
The distinction is not philosophical. Email signups measure politeness. A person who likes you, or likes the idea in the abstract, will hand over an address. That number can be pushed up with better copy, a giveaway, or a bigger tweet, and none of those movements tell you whether a business exists.
Myth two: it was one page
It was two, and then three, and the third one is where the actual test lived.
The first version, described in Gascoigne's February 2011 write-up of the launch, was a two-page site. Page one explained what Buffer did and offered a button to see plans and pricing. Page two was the email capture, with a short message admitting Buffer was not quite ready.
Once clicks confirmed people wanted the thing, Gascoigne inserted a pricing page between the two. Now the sequence read: here is what it does, here is what it costs, and only then, here is where you leave your email. That inserted page changed the meaning of every signup that followed. A person who saw a price and continued anyway had answered a much harder question than a person who saw a promise.
That is the whole design. Two tests, run in sequence, each one costing a few hours: is this interesting, and is it interesting at a price?
Myth three: it went viral
The distribution was as ordinary as it gets. "I simply tweeted the link and asked people what they thought of the idea," Gascoigne wrote of the launch of the test.
His 2013 essay puts the yield in perspective: 120 signups from the landing page, of which around 50 became users when the product went live. That is not a crowd. It is roughly the number of people who might reply to a good post in a niche community, and it was enough to justify seven weeks of building, because the signal was qualitative, not statistical.
Two and a half months after launch, Buffer had more than 500 users and about 4% of them were paying. Small numbers, cleanly measured, acted on quickly. If you are hunting for a problem to test in the first place, our guide to mining public complaints covers where those signals are already written down.
What made this a real test rather than theatre
Three properties separate Gascoigne's version from the imitation version.
It could fail. If nobody clicked the plans button, or if everyone bounced at the price, there was a defined outcome that meant do not build this. Most landing page tests are constructed so that any result reads as encouragement, which makes them expensive decoration.
It was staged. Interest and willingness to pay were separated into two questions asked in order, so a failure told you which of the two had failed. A single page conflating both tells you only that something did not work.
It was honest at the point of capture. Visitors who reached the end were told the product was not ready. That mattered later: the roughly 50 people who became users at launch already knew what they had signed up for, so nobody arrived feeling tricked.
The pricing page also did something founders routinely defer for years. It forced a number to exist before there was a product to attach it to. We have written about how first prices tend to be set too low and then never revisited; putting the price in front of strangers in week one is the cheapest possible defence against that.
The decade that followed, in the founder's own numbers
The validation story usually stops at the first customer. Buffer's ten-year post is more useful because it keeps going, and the middle is not tidy.
By 2011 the company had reached what Gascoigne calls ramen profitability, around GBP 1,200 a month, and had raised $450,000. In 2012 it went fully distributed. In 2013 it introduced transparent salaries and survived a security breach. In 2014 it raised $3.5 million and turned down a nine-figure acquisition offer. In 2016 the team grew from 34 to 94 people and then 10 were laid off when cash flow tightened, a stretch Gascoigne describes as moving into a house the company could not afford. In 2017 a co-founder left and the company logged 18 consecutive profitable quarters. In 2018 it spent $3.3 million buying its venture investors out.
That last line is the one worth sitting with. Buffer's most-quoted moment is a landing page that cost an afternoon. Its most expensive lesson was that taking money is easy and unwinding it costs millions and years.
Running your own version this week
Write one page describing what your thing does for a specific person. Add a plans page with real prices you would actually charge. Put an honest "not ready yet, leave your email" behind it. Send it to the people who already have the problem, in the places they already gather, not to a general audience. Then watch two numbers: how many people click through to see the price, and how many continue after they have seen it.
Decide in advance what result means stop. That single sentence, written before you look at the data, is the difference between Gascoigne's test and a hopeful web page. And when the first handful of users arrive, treat them the way he did: as people to learn from, not as a market. Our piece on where first customers come from explains why those first fifty are almost never strangers.
Dates, figures and quotes are from Buffer's own published posts and Joel Gascoigne's essays, checked July 25, 2026. Note that Gascoigne's accounts differ on whether the first paying customer arrived three or four days after launch.
Discussion
Sign in with Google or just a name. No email link, no password to remember.