Wednesday, June 4, 2003

The Slippery Slope of Statistics In a recent pitch, an entrepreneur told us that his software client was being downloaded by 10,000 new users a week. The number was an interesting one but it led to many more questions that the entrepreneur was unable to answer at that moment (e.g., how many users become paying customers?, what's the attrition rate?, etc.). To his credit, the entrepreneur had tracked all that data and gave us access to it at a later date, but at that time it left a number of unanswered questions about adoption, sales, market, etc., as well as the rigor of that entrepreneur's analysis. Statistics can be incredibly powerful and venture capitalists love them. We often invest in emerging markets or technologies. As a result, we can only predict the speed of adoption of a new technology or the ways in which a new market will develop. While we certainly can look to historical trends and past patterns of adoption in related markets, the most powerful tool we use in predicting the future of a particular product or technology is its past performance. If it took a company a year to attract its first thousand customers, six months to attract its second, and three months to attract its third, the trend is an interesting one, even if it doesn't assure that the next thousand customers will sign up in a month and a half. And that trend alone may be sufficient information to convince a venture investor that your opportunity is an interesting one. But have no doubt that statistical analysis is a slippery slope. For every statistic there are a hundred interpretations and a thousand more statistics that may be helpful to sort out which interpretation is the correct one. And each venture capitalist with whom you meet will likely focus on a slightly different aspect of that statistical analysis. So the best thing you can do is track every statistic that you believe is crucial to the understanding of your business, know those statistics cold, and be prepared to explain why it is you believe that those are the key metrics for tracking your progress. The better you understand those statistics, how they impact your business model, and how they've changed over time, the more persuasive you will be when pitching your company. [VentureBlog] I'm finishing Stats I now, and this is prescient advice. 8:10:49 PM

On the Bursty Evolution of Blogspace... On the Bursty Evolution of Blogspace (registration required) - interesting. I'm going to have to pay more attention to what the research community discovers through statistical analysis of blog usage. [Ray Ozzie's Weblog] Lordy I want to get at that information! Being able to quantitatively define the growth rate of weblogs would go a long way to validating my business model! 8:07:32 PM

Macromedia and RSS You may have noticed that over the past year we have really begun to leverage RSS through weblogs, content on the DevNet Resource Kits and the Macromedia XML News Aggregator (MXNA). This string of resources and projects utilizing and focusing on RSS has not been a coincidence, but rather the product of the realization of the power of RSS. This realization began over a year ago when we began publishing weblogs. Initially, we saw weblogs primarily as a way to... [mesh on mx] The question is, does Macromedia see RSS as the disruptive technology that it is? In other words, do they see that "browsing" has reached its peak, and that "aggregating" is the next phase? Creating server-side aggregators is all fine and good (and I nice example, since its so simple), but do they see that desktop tools for leveraging RSS are where the market will zig next? 8:05:11 PM

Why does NewsDesk integrate with NewsIsFree? Earlier this year, while NewsDesk was still in beta release, many of you told me that I simply had to make it easier for new users to get started. The two most popular requests were for (1) preloaded channels and (2) an easy way to add more. I considered the alternatives, and decided that the best solution was to partner with one of the leading RSS directories. After evaluating several of the most popular directories, I concluded that NewsIsFree offered the best service at a reasonable price. If the "Free" in NewsIsFree seems out of place, consider that it was free at one time, but demand grew so fast that NewsIsFree later added fees to cover operating costs. Their basic service is still free -- you can get headlines for most of their channels, though they limit free users to five headlines per channel and do not include headline descriptions. I contacted Mike Krus at NewsIsFree, who agreed to collaborate with me to bring preloaded channels and seamless integration to NewsDesk users. So what do you get for your $20 NewsIsFree subscription? Use the Add Channels wizard to effortlessly search the 5,700+ popular channels they provide, and add NewsIsFree channels to NewsDesk with just a few mouse clicks. You get headlines updated every hour, with headline descriptions for most channels. If you're looking for headlines from reputable sources around the world, NewsIsFree is a wonderful way to get them. NewsDesk is the only RSS headline reader to make it this easy. [NewsDesk Updates] Excellent idea! 7:59:22 PM

Sherlock vs Watson vs Macromedia Central Sherlock vs Watson: Matt Deatherage has a story in MacWorld comparing the growth of Apple's Sherlock and Karelia's Watson. Both are pluggable web-savvy shells, and overlap with Macromedia Central in some ways. It's good history, but I'm not sure I... [JD on MX] The comparisons of Macromedia Central to Sherlock/Watson come from all over, and I'm glad I'm not the only one who sees it. Konfabulator is another product that looks similar. I just don't see the market, frankly. 7:49:11 PM

Uncle Sam's Huge IT Budget There was an article in the Business Section of Sunday's San Francisco Chronicle entitled "Selling to Uncle Sam." The basic gist of the article is that while startups used to be reluctant to sell to the United States Government, as the government has increased its technology spending it has become an attractive customer for all infotech companies, big and small alike. According to the Chronicle, the Department of Defense will spend $364 Billion on technology products and services in 2003 (going to $380 Billion in 2004) and the Department of Homeland Security will spend $28 Billion in 2003 (going to $36 Billion in 2004). That's a whole lot of technology products and services. My experiences over the last year or so definitely reinforce what the article suggests. Particularly in the wake of 9/11 and the "war on terrorism," the Federal Government is working hard to integrate information gathering, sharing and analysis across and among each of the departments. That is a huge project. Add on top of that the security you need to protect that massive store of data, and you've got a whole lot of technology dollars being spent. I would say one thing, however. It remains infinitely easier to sell to the Federal Government as a subcontractor to one of their trusted suppliers than to try to sell direct. But for those startups that have managed to find the right channel into the US Government, the payoff is significant. [VentureBlog] With all the news I read about various state and local governments diving into RSS, I can see that weblog software could do very well here. 7:46:06 PM