The Genesis of NewsForge As Linux and Open Source (and Slashdot's) popularity grew, Slashdot got an increasing number of reader requests to cover more Linux and Open Source news and do less coverage of "frivolous" things like movies, toys, and wing-ding science. Instead of yielding to this clamor, which came from a small percentage of Slashdot readers, and turning the site into something it was never meant to be, in early 2000 we decided to make an entire new Web site that would be the "online newspaper of record" for Linux and Open Source. The initial concept was a simple site that would do nothing but display headlines and brief summaries of stories published elsewhere about our chosen topic areas. Jamie McCarthy, a programmer who also wrote articles for us part-time on Slashdot and was writing real-time search software in his day job, joined us full-time to write a program called NewsVac that would scan hundreds of other news sites, all over the world, looking for key words like "Linux," "BSD," "Open Source," and others that seemed appropriate. Our intention was to log every single Linux-related story published online, anywhere from Linux Weekly News to the Singapore Straits Times, that was relevant to our target readers. Then came a little ethical question: If we viewed online news publishing as a huge "pot" of stories, what right did we have to dip into that pot without putting something back in? There is no legal problem doing this; we weren't republishing work published elsewhere, just linking to it, and links are the essence of the World Wide Web. But it still didn't feel right. So I cast about a bit, and hired Grant Gross, who had ten years of newspaper reporting and editing experience, plus two years of online writing behind him, as NewsForge Managing Editor. Tina Gasperson, who was already writing for us on several sites, joined Grant on NewsForge. We had several freelance columnists who had been working on a previous, non-Open Source news site, along with several part-time editors to cover our NewsVac feed and write brief link-story summaries on nights and weekends when Grant and Tina weren't around to do it. But Grant's main job was to produce real, newspaper-style coverage of as many events pertaining to Open Source, Free Software, and GNU/Linux as possible. There was a need for this kind of service. Most online Linux news coverage at the time was fanzine-like, with little hard-core reporting going on. "Gosh, isn't Linux great? We love Linux s-o-o-o-o much!" was nice to say now and then, but a constant diet of it was not useful to readers who needed a clear and accurate picture of what was going on in the Linux and Open Source communities, especially from a business perspective. I figured it would take at least a year to get our coverage up to speed; initiating coverage in a new, specialty business area takes time and work, hard-core contact-building, and a body of institutional knowledge, plus it takes time to build reader credibility for any publication that is trying to provide insider coverage of a comparatively small, "everybody knows everybody" community. And even after its rapid post-1997 expansion, the Linux community was still tiny compared to the Windows-using crowd. But Grant and Tina amazed me by starting to routinely produce credible Linux and Open Source breaking news coverage less than six months after NewsForge was launched in August, 2000, and by the time the site celebrated its first anniversary it was averaging about 30,000 pageviews on weekdays, plus another 40,000+ readers for the two daily newsletters we generated from its constantly changing content. In early 2002, we took the radical step of putting our underperforming Linux.com site under direct control of the same editorial team that had done so well with NewsForge. It proved to be an excellent decision. Three months after we relaunched Linux.com with a new look and new personnel, the combined readership of Linux.com and NewsForge was averaging at least 150,000 pagerviews per day and climbing steadily. NewsForge Financial Projections My original financial projections for NewsForge were based on an average ad banner sale price of $5 per thousand pageviews, with five banners per page and 80% of ad inventory sold, or $20 in gross income per 1000 pageviews. This was considered exceedingly conservative when I drew up the original NewsForge plan in early 2000; our (then) publishing VP was confident that he could get at least $20 per 1000 pageviews per banner and sell all five on each page, for a gross income of $100 per 1000 pageviews. Based on my projection of 25,000 pageviews per weekday and 10,000 on weekend days by the end of NewsForge's first year in existence, this would have given the site $754,000 in gross annual income vs. a budgeted annual operating cost, once it was up and running, of (roughly) $350,000. These are the kind of numbers we all like to see! My original combined budget for NewsForge's development and first year operating cost was $500,000, and I dolorously assumed no income at all during this period. Without a minimum budget commitment of $500,000 I was unwilling to build the site, despite all the optimistic sales department input I was getting. I sadly got to say, "I told you so," during much of 2001, because NewsForge ad banners like ad banners on other sites did not sell as rapidly (or for as much) as pre-recession industry wisdom (wisdumb?) had predicted. I had set my budget based on the possibility of a sharp decline in online advertising activity. I didn't predict a recession; I just allowed for one, figuring that if we really were experiencing an endless boom (which I really didn't believe was possible), the site would make a ton of money, and if the economy tanked, we'd be able to survive without backing away from my original vision of NewsForge as the world's most complete source of Linux and Open Source news. Projections are nothing but guesses, of course, and mine were neither better nor worse than anyone else's. NewsForge site readership grew more rapidly than I expected. Advertising sales were less robust than I expected, but the ad rates which we were able to get did not drop as precipitously as they did for many other online publishers, in large part because the type of news coverage which NewsForge provided proved attractive to a demographically ultra-desirable IT upper management audience segment which no other site reached as directly as we did. And now, as I write this in mid-2002, premium "top of the page" ads on NewsForge are selling quite well, and we are working to achieve that same state of grace with Linux.com. Reader Feedback as a Secret Weapon All the financial projections really did was give us a "go/no go" decision point and a budget base. They told us that, in order for the site to survive no matter what happened in the advertising marketplace, we had to keep our editorial staff small and server and bandwidth expenses to a minimum, which we did. We did not have good numerical data, at the beginning, about exactly what content we should provide. What we did have was several years of experience running Slashdot, Linux.com, freshmeat, and other Web sites popular with Open Source developers and users. This was (and is) a vocal crowd, and we have always listened to what its members say with totally open ears. I personally received (and still receive) anywhere from 500 to 1000 reader emails per week, and I read them all. This high level of feedback acceptance is the secret of NewsForge's success. Indeed, it is the secret of all OSDN sites, and why it seems like we know exactly what our readers want and who they are even though we rarely do formal readership surveys. What we learned from NewsForge readers, almost immediately, was that they liked our original reporting more than the NewsVac feed. Sure, they appreciated our links to virtually every on-topic story, announcement, or press release published anywhere on the entire World Wide Web, but they considered our in-house reporting, hardware reviews, and weekly columns the site's most attractive features. Our response, naturally, has been to gradually increase the number of original stories we publish. We have also learned that we wander beyond our original Open Source/Free Software reporting niche at our peril. Every time we have published anything even a little bit outside this self-imposed mandate, we have gotten an earful from readers. In effect, we have an editorial board with thousands of members who keep us on the straight and narrow. While we don't take every suggestion we get seriously, when we hear the same complaint or compliment over and over again, we sure as bleep pay attention. We also take the source of a comment into account; if it comes to our editors@newsforge.com email address from the CIO or CTO of a large company that uses Linux heavily, we are more likely to listen to it than if it comes from a high school student who is thinking about installing Linux for the first time. But a reader's potential purchasing power is not everything. If an email suggestion about Slashdot content comes from an accounting manager at an international consulting firm who knows little about technical matters, we are less likely to take that reader's advice than we are to take advice from a 25-year-old programmer doing freelance projects for local small businesses, because that programmer is part of Slashdot's core target audience and the accounting manager is not. Listening carefully to our readers and just as carefully choosing which readers to listen to gives each of our sites a distinct identity, direction, and readership. But listening is not something you do for a little while, then stop doing once you feel you've gotten enough feedback, because both readers' and advertisers' needs change over time. Could we decide, at some point, that it is better to combine NewsForge with Linux.com than to maintain it as a separate site? We might if enough of our readers tell us that's what they want. |
