As WordPress Incorporated fizzled, Matt pitched a WordPress-based blogging network to his employers at CNET. Many major internet companies had blogging networks including Google, Blogger, Yahoo 360, and Microsoft Spaces. CNET also owned several domains, like online.com, that seemed perfect for a blogging network.
CNET decided against it, but the idea didn’t disappear. Matt decided to build a blogging network himself. [footnote]CNET went on to be one of the first investors in Automattic.[/footnote]
First, he needed the WordPress.com domain name. At the time, it was owned by Open Domain, an organization that registered domains and gave projects permission to use them in return for acknowledgement. It was unclear whether Open Domain was squatting on domains, or genuinely trying to help free software communities. (They’d also registered Drupal.com, then donated it to Drupal without incident). The WordPress community was understandably perturbed by the idea of someone squatting on WordPress.com. Owning the domain was key; there was little security in a blogging network without control of its own name.
After months of wrangling, Matt acquired the WordPress.com domain and work began. Donncha, as the developer of WPMU, was a perfect candidate — and fortuitously, was looking for new work opportunities. When Matt emailed the WordPress security mailing list to see if anyone was interested, Donncha got in touch.
The new WordPress.com ran WPMU trunk. Development was fast. Donncha worked on two servers and live-edited code. Without users to worry about, he moved quickly. He improved user-focused functionality and built network administration tools. The WPMU community helped, submitting patches to clean up admin screens. Features like domain mapping and tags, which didn’t sit well in WPMU, went into plugins.
Andy Skelton (skeltoac) was WordPress.com’s second employee — and first acquisition. Andy had developed Blogs of the Day, a self-hosted stats plugin that produced a list of each day’s most popular blogs. After meeting him in Seattle, Matt brought Andy on board to create a stats plugin for WordPress.com based on Blogs of the Day which, for a number of years, was featured on WordPress.com’s home page.
WordPress.com opened to signups in August 2005, by invitation only, to control user growth on untested servers. Many who were involved with the WordPress project got WordPress.com blogs, including Lorelle VanFossen and Mark Riley. Every new WordPress.com member also got one invite to share.
People could also join by downloading the Flock browser, a browser with built-in social networking tools. “I thought Flock was the future,” says Matt. “First we did invites, and then we thought well, we’ll allow you to bypass an invite if you’re using Flock because then we’ll know that you’re kind of a social, in-the-know person.” Flock integrated with WordPress.com, and users could use it to post straight to their blogs.
Both of these measures — invites and Flock — allowed WordPress.com to grow in a steady and sustainable way. Putting in sign-up barriers controlled the flow of people and helped ensure scalability. Nevertheless, demand quickly outstripped supply; one invite even landed on eBay.
The influx of new bloggers also brought demands for support. Without a clear distinction between WordPress.com and WordPress.org, many bloggers made their way to the WordPress.org forums looking for help. This caused some discontent among WordPress.org forum volunteers, who felt that they were doing free work for a service that would eventually become commercial.
The first attempt at WordPress.com-specific support was an email address. Developers responded to the messages, diverting their attention from writing code. “I found that I would spend half my day replying to users,” recalls Donncha “and then being completely wrecked in that your mind is completely taken off programming new features, or improving things, fixing bugs, just because this whole thing is so tiring.” Plan B was a mailing list, followed by WordPress.com-specific community support forums.
Eventually, support moved to a ticketing system led by Mark Riley, a veteran of the WordPress.org support forums who became WordPress.com’s sixth full-time employee. For a long time, he was solely responsible for support, closing nearly 50,000 requests on his own. The next support person joined in 2008; today, WordPress.com users have a robust community on their own forums and a huge team of “Happiness Engineers” supporting them. (Although it’s still not unusual for WordPress.com users to land on the WordPress.org forums.)
Creating a user-focused blogging platform like WordPress.com made sense in the context of the WordPress project: WordPress has always been focused on its users, on people who might not understand the mechanics of a server, but still want a website. WordPress.com took this accessibility to the next level, letting people fill out a form to get a blog with the power of WordPress. Still, WordPress.com is a balancing act, trying to satisfy those who want a simple website builder, with those who want the flexibility and power of the core WordPress software, and providing a middle ground for people who want a website without building one from scratch.
However, a hosted service like WordPress.com needs a revenue stream to pay for server costs, run the software, provide support, and pay staff. The free software project, with its haphazard donations and income from hosting companies, couldn’t maintain such a network on a wide scale. WordPress.com needed a supporting business focused on WordPress’ main audience: software users.
While the business end was coming together around the developers who had built WordPress.com, Matt was working on another product that would influence the WordPress community, an anti-spam plugin called Akismet.