← all stories
1999–2003acquired by Google · 2003

Blogger

Pyra Labs

A tool for publishing web pages instantly, grown out of a side internal tool at a company that was selling something else entirely: Pyra, a web-based task and project manager. Between 1999 and 2003 it went from a free form with automatic FTP publishing to a service with a million users, surviving the departure of the whole team and almost a year in which one of its two founders carried it alone, and it was bought by Google in 2003.

Founders Evan Williams · Meg Hourihan
Domains pyra.com · blogger.com · blogspot.com · pro.blogger.com
bloggingpublishingcmssolo-founderviral-badgeinternal-tool-pivotbootstrapped

In brief

Blogger began not as a blogging service but as a side internal tool inside a completely different startup: Pyra, a web application for managing tasks and projects. When Pyra ran out of money and the entire team, including the other co-founder, went their separate ways, one of the two founders, Evan Williams, kept the service alive single-handedly for more than a year on nothing but enthusiasm and other people's goodwill. What saved the company was not an acquisition but a licensing deal, agreed in outline over a sushi dinner that cost about $50. Two years after that rescue Blogger was bought by Google, and it became one of the first genuinely mass tools for self-publishing on the internet, the one the word "blog" itself grew up around.

How it started (the founders)

Evan Williams and Meg Hourihan met in 1998; in Hourihan's own words, "Ev was the first person who I met that was like, '[The Internet] is incredible and I want to build things for it'" estimate. In January 1999 the two of them founded Pyra Labs in San Francisco, and Williams himself wrote about it in 2001 without naming an exact day: "Meg and I started Pyra in January 1999." Later sources give the precise date of January 1, 1999, but that precision is not confirmed by a primary source estimate.

The company's first product was not a blogging platform but the Pyra web application, a "power tool designed to help you manage tasks and related project information in an easily accessible, shared space": a manager for tasks, contacts, and team discussions. The official 2001 press release described the company's mission more broadly still, as "power tools for regular people." While Pyra was still in beta, part of its code was reworked, mostly by the engineers Paul Bausch and Matt Haughey, into an internal tool for quickly publishing notes on how development was going. That side tool turned out to be so convenient in its own right that they decided to release it as a separate public product; several independent modern sources converge on the date August 23, 1999, while Williams and Bricklin themselves spoke of it a little later in vaguer terms, as "released in late August 1999."

At the start Blogger was not aiming at a mass user at all. Hourihan herself recalled the original idea this way: "When we first started Blogger, it was more of the idea to just make it a little bit easier for people who already knew how to do it," meaning the tool saved time for those who already knew how to assemble HTML pages and had their own hosting, rather than opening publishing up to anyone at all. The product described itself as plainly as possible: "Blogger is an automated weblog publishing tool... No muss. No fuss. Total control," and it was free, its one technical trick being the automatic FTP upload of the finished page to the user's own server. Tellingly, at this stage Blogger was publicly listed as a module of the parent company, with a "Powered by Pyra" badge at the bottom of the page. At about the same time, among early web designers and tech bloggers, the word "blog" itself was born: in April or May 1999 (not even he names the exact day) Peter Merholz jokingly proposed in the sidebar of peterme.com pronouncing "weblog" as "wee'-blog," shortening it to "blog," as he himself described it three years later: "Sometime in April or May of 1999 (I can't say for sure when I exactly did it)... I've decided to pronounce the word 'weblog' as wee'-blog. Or 'blog' for short."

Year-by-year timeline

Lesser-known but significant facts

  1. The deal that saved Blogger was discussed and agreed in outline over dinner at a sushi bar. Williams and Bricklin met in San Francisco on February 13, 2001, and went to eat at Niko Niko Sushi; the receipt survives, $41.10 plus $8.50 in tip. The next day Williams got Trellix CEO Don Bulens on the phone, and a day after that, on February 15, received a draft of the deal terms from him by mail; the formal signing, with the money transferred, took place 17 days after the dinner, on March 2, 2001.
  2. At the height of the crisis one server was holding 25,000 sites and had never once been rebooted. A few days before the whole team broke up, Williams wrote on his personal blog: "Gutenburg hosts 25,000 sites and has never been rebooted," a note on the scale of the load that was by then being physically kept up by very nearly one person.
  3. A tool that was free and open to everyone was only half free in its first year. To publish anything at all through Blogger, a user still needed their own web hosting and a minimum knowledge of FTP, the very technical barrier the tool was supposed to remove. The service became genuinely open to everyone only a year later, with the launch of BlogSpot.
  4. Williams seriously considered closing off free access a year and a half before the paid plan appeared. In January and February 2001, trying to survive, he thought about charging for Blogger immediately; what stopped him was not a change of heart about the idea but the fact that the Trellix deal arrived faster and gave him breathing room.
  5. On the day the Google deal became public, what the company feared most was not competitors but a rush of the curious. A post on the Blogger.com home page dated February 17, 2003, said that the only immediate change was an urgent purchase of servers to handle "the extra load this news has caused."
  6. When there were not enough servers and investors had said no, Pyra simply asked its own users for money and had it within a few hours. On January 2, 2001, blogger.com openly launched the "Blogger Server Fund," a drive for $5–10 donations toward a new server. Within a few hours 191 people had chipped in $3,200 in total through PayPal, including $25 from a direct competitor, Dave Winer, the author of his own tool, Edit This Page.

The first growth lever

Blogger's growth was built on a pair of mechanics separated in time by about a year. The first was radical simplification of the act of publishing itself: the user filled in a form on the site, the tool automatically uploaded the finished HTML page to their server over FTP, and all of it was free, "No muss. No fuss. Total control." A built-in side effect of that mechanic was a free viral trail: next to the "Powered by Pyra" link on Blogger itself, and on tens of thousands of personal sites including Williams's own evhead.com, sat a "this web site is powered by Blogger" badge that clicked back through to blogger.com, a direct ancestor of today's "made with" and "powered by" badges that site builders and AI tools now embed in their free plans.

But this first version had a hidden barrier: it was designed for people who already had their own hosting and a basic knowledge of FTP, or, as Hourihan herself admitted, simply to "make it a little bit easier for people who already knew how to do it." The real explosive reach came from the second part of the pair, the launch of BlogSpot about a year after Blogger started, which took hosting on itself for those who had none; after that, more than half of all new blogs were created there straight away. This was the moment the tool for geeks became a tool for everyone.

The user numbers grew as the barriers came down: about 2,300 users by January 2000, an estimate from secondary sources (ultimately tracing back to Scott Rosenberg's book "Say Everything") and not independently confirmed by a primary document estimate; by early November 2000, about 65,000 registered users; and by January 2, 2001, when Pyra openly launched the donation drive for a new server (the "Blogger Server Fund"), about 86,000 (+21,000 in two months, by Williams's own account on the back of coverage in Wired and the New Yorker), both figures coming from a Wired News article written in the immediate aftermath of the events; by the time of the Trellix deal in April 2001, 150,000 registered users, a figure confirmed by two independent primary documents at once, the official press release and Williams's own essay; by July 2002 the rate of new blog creation reached 1.5 a minute, or about 65,000 new blogs in a single month; and by the time of the Google purchase in February 2003, more than a million registered users and about 200,000 active blogs, a figure that converges across three independent modern sources.

What changed during verification: the figure of about 100,000 active accounts by the end of the first year (that is, around the fall of 2000), which stood here in the first pass on the strength of a single secondary source, was not confirmed and has been removed: a primary source of the period records only about 65,000 users at the start of November 2000. Growth, it appears, was smoother and more drawn out over time than the earlier source suggested.

The price of that growth for the one person left was high and not abstract: for about 13 months, from the layoff of the whole team in December 2000 to the Trellix money arriving in March 2001, Williams alone was responsible for the code, the servers, the overdue bills, and communication with users, while seriously considering an emergency switch to a paid model as a last lever. What probably let him survive alone in that role was the extreme technical simplicity of the product: one form and one FTP call, with no infrastructure of its own for storing users' content.

Parallels today (projects from the catalog)

What a builder can take from this in 2026

  1. The product the money was raised for and the product that took off are often not the same thing. Pyra attracted investors on the strength of a task manager; what survived and eventually became the company was a side internal tool built just in case for the team's own needs.
  2. Virality can be wired into the very fact of using the product, and it costs nothing. The "Powered by Blogger" badge, spread across tens of thousands of other people's sites, was a built-in free distribution channel rather than a separate marketing campaign.
  3. Free and simple is half a product if a hidden barrier to entry remains. The first version of Blogger was free but required your own hosting; the service became genuinely mass-market only a year later, when that second, less obvious barrier was removed.
  4. Having no way to make money is not freedom but fragility. A year and a half without a single paid option brought the company to the brink of closing and forced it to rely on luck, on an outsider moved by a blog post, rather than on a decision it could have controlled in advance.
  5. A simple architecture is built-in insurance for the day the team disappears. Blogger could be carried alone for more than a year precisely because the whole system came down to one form and one FTP call; a service with complex infrastructure in the same position would most likely not have survived the whole team leaving at once.

Discrepancies and what we could not verify

Sources (primary first)

Primary (period documents and the founders' own words):

Secondary (analysis, retellings, and encyclopedia entries):

The same thing, about today

The same breakdowns, but of projects launching right now: what the product is, where the first users came from, how they charge. The card is free, the full dossier is $5 (the dossier itself is written in Russian).

Browse the dossier catalog →