Skip to content

Dead End: PointCast and Push Technology

Abstract

In 1996 and 1997 the future of the internet was supposed to be “push.” Instead of you going to information, information would come to you: news, stock quotes, and weather delivered automatically to your screen. The company that defined the idea was PointCast, a free screensaver that filled your monitor with scrolling headlines whenever you stepped away from your desk. In March 1997 Wired put push on its cover and told readers to “kiss your browser goodbye.” News Corporation offered around $450 million to buy PointCast; the company turned it down. Within two years push had collapsed: PointCast’s constant downloads clogged corporate networks, dial-up users could not feed it, and portals like Yahoo did the same job without the bandwidth bill. PointCast sold for about $7 million in 1999. The idea it botched, though, came back with a lighter design and a different name: RSS.

What PointCast Was

PointCast was founded in 1992 by Christopher Hassett in Sunnyvale, California, and launched in beta on February 13, 1996. The product was a free Windows application that doubled as a screensaver. A user picked channels (CNN, Reuters, sports, weather, a stock ticker keyed to their own portfolio), and PointCast pulled fresh content over the internet and displayed it as animated headlines and a scrolling ticker when the machine went idle. It was advertising-supported: the downloads carried full-screen ads along with the news.

The pitch was that the web had the interaction backward. Browsing meant hunting: open a browser, type an address, wait, click, repeat, for every site you wanted to check. PointCast inverted it. You configured your interests once and the news arrived on its own. The industry name for this was push (the server pushes to the client), against the pull of ordinary web browsing. PointCast was not alone; BackWeb, Marimba’s Castanet, and others sold variations, and both Microsoft and Netscape built push features (Active Channels, Netcaster) into the browsers shipping in 1997.

The Hype

Push became the technology story of 1997. In March 1997 Wired ran a cover package titled “Push!” declaring it “The Radical Future of Media Beyond the Web” and suggesting the browser itself was on the way out. PointCast was the poster child, briefly one of the most talked-about startups in Silicon Valley.

The money followed the noise. In January 1997 Rupert Murdoch’s News Corporation offered roughly $450 million to acquire PointCast. The company rejected it. News Corporation withdrew the offer that March; James Murdoch later blamed “PointCast’s inaction.” It was the high-water mark. Hassett was replaced as CEO, David Dorman took over, and in May 1998 PointCast filed for an IPO at a $250 million valuation, a fraction of the offer it had spurned a year earlier. The offering was pulled within two months.

Why It Died

The product’s own design killed it. PointCast pulled a steady stream of content and ads over the network, and in 1997 that was expensive on both ends. On home dial-up (28.8 or 33.6 kbit/s modems), the downloads were slow and monopolized the line. On corporate LANs, hundreds of employees running PointCast in the background generated enough traffic to congest company internet links, and network administrators responded by banning it outright, cutting PointCast off from exactly the always-connected office machines that were its best audience. Users also tired of the volume of advertising the “free” service pushed at them.

Meanwhile the problem PointCast claimed to solve was being solved more cheaply elsewhere. Web portals like Yahoo and Excite let a user assemble a personalized start page (My Yahoo) that showed the same headlines, stock quotes, and weather, on demand, in an ordinary browser, with none of the background bandwidth. Push as a heavyweight client-side event turned out to be the wrong shape. By 1998 the “push” label was already dated.

David Dorman left, and in May 1999 PointCast was bought by Launchpad Technologies (an Idealab company) for about $7 million. The network was shut down in 2000. The remnants became EntryPoint, then Infogate, which AOL Time Warner acquired in March 2003.

What Survived

The underlying idea (let publishers notify subscribers of new content instead of making them check) was sound. What was wrong was the plumbing: a proprietary, ad-heavy, bandwidth-hungry client. The durable version arrived as a plain XML file that a lightweight reader could poll cheaply: RSS, whose first version Netscape shipped in 1999 for its My Netscape portal. RSS delivered on push’s actual promise (subscribe once, new items come to you) using a fraction of the resources, and it powered feed readers and, later, every podcast. See Dave Winer and RSS for how the format evolved, and Dead End: Google Reader for the client that gave it a mass audience before it, too, was shut down. The word “push” itself survived, attached to a different and narrower thing: the mobile push notification, a short server-sent alert rather than a screen full of scrolling news.

📚 Sources