Two weeks ago, Kirsch sent me an e-mail asking if I wanted to give AnchorDesk readers a sneak peek at Propel's new service, which will be formally introduced next Monday. Of course, I said yes.
First a little background on Propel and its product. Web acceleration is actually the company's third business plan in as many years. Started with $50 million in venture money, secured just before the dot-bomb implosion, the company worked on technologies to speed up e-commerce and databases before finally settling on trying to make Web surfing faster for dial-up users.
PROPEL DOES SO by installing client software on your system, software that directs all Web-page requests to a proxy server operated by Propel and hosted by Electronic Data Systems. This proxy server's job is to send you only the Web content you've requested that isn't already resident on your machine.
Yes, your browser already tries to do just that with its own native caching. But Propel couples some advanced Web cache-management techniques that your browser doesn't know about with other acceleration technologies to achieve its speed improvement. No one of these technologies is responsible for the increase in throughput, Kirsch told me, but together they add up to a significant speed boost.
Immediately after seeing the demo at Propel's offices, I enlisted two friends--both hard-core dial-up users--to test the service. Both are stuck with dial-up, because neither high-speed cable nor DSL services are available where they live. They both went to the Propel Web site and signed up for the 7-day free trial. Meanwhile, I played with Propel myself.
One of my two guest testers saw some immediate and fairly significant speed increases, and said he'd be willing to pay for the service. The other, however, had a more mixed experience.
For example, Kirsch claims an average 3x speed increase, with occasional pages loading five times faster, but we didn't see much of that. Our testing averaged a 2.5x speed increase. This was because the testers looked at the pages they cared about, which may not be in the Propel cache, instead of the collection of popular sites the company uses as its benchmark.
IN THEORY, while Propel should be a little faster the very first time it "sees" a page, the real speed increase should come with subsequent viewings (which don't have to happen during a single Web session). The reason: Propel needs to fill its cache on the first download and can only start using that cache the second time around.
But in our tests, the increase sometimes didn't occur until after more than two viewings: "After several visits to the NY Times site, I'm up to 3.2 times faster, and it IS finally noticeable (it wasn't after two visits)," one tester wrote. "There's a big pause at the beginning, and then the whole page suddenly appears."
Some Web pages are naturally faster than others. Pages designed with speedy downloads in mind won't benefit as much as pages covered with large graphics. In general, your Web usage pattern will impact how much of a speed increase you see. People who go to the same sites day after day will see the best increases. Those who never visit the same Web site twice will be less impressed.
IF YOU USE AOL as your Internet service provider, you'll need to use the browser included with AOL or Netscape (not the copy of Internet Explorer on your desktop) to get the greatest speed boost. One of my testers started out using IE atop AOL and was dissatisfied with her initial results before switching to the AOL browser.
Content actually hosted by AOL (which doesn't use a browser) will not be accelerated since it's not delivered as a Web page. This seems obvious enough, but some AOL users fail to recognize that the content comes from different places.
Propel makes no claim to speed up anything but Web pages. It is working on other technology to speed downloads and perhaps even streaming media, but at this time Propel won't increase their speed at all.
ONE THING Propel doesn't do that could increase download speeds: block advertising. Given that the slack online advertising market has cost many of my friends their jobs, I'm loathe to even mention this. But if blocking pop-ups (and pop-unders) would speed page-load times for the information I actually care about, I'd be all for it.
The program isn't entirely convenient. You have to launch the desktop client before beginning your Web browsing, something that's easy to forget and should be automatic. The program also seems to soak up a fair amount of memory.
The current $4.95-a-month pricing is described as "introductory;" long-term pricing may increase to as much as $6.95 a month. Besides marketing directly to users, Propel is also working on deals to have the service sold as part of a "premium" offering by dial-up Internet service providers. Over time, I believe this will be the biggest source of Propel users--people who don't necessarily realize they are using the company's product.
Because all your Web-page requests flow through Propel's servers, I'd be remiss to ignore the privacy implications. Kirsch promises that no information directly links a user to his or her browsing patterns. User information is kept in one database, managed by an outside company and separate from the database used to authenticate clients; the latter only knows which client machines are active. Users do not actually log on to Propel; the authentication takes place in the background.
BOTTOM LINE: Both testers liked Propel, one significantly more than the other; the lukewarm tester warmed to the service the more she used it, but still isn't sure she'd actually pay for it. Though Steve's 3x "average" speed increase may or may not map to your actual experience, our 2.5x increase was close enough that we won't quibble.
Is this something people will buy? I'm a speed freak and, if I found myself chained to a dial-up connection, I'd happily pay for the speed increase Propel provides.
Of course, for $25 a month more than an AOL account, you can graduate all the way to cable modem or DSL broadband Internet access. (And, no, Propel doesn't really speed these up.) But Kirsch is betting that enough people don't want to pay that extra money for DSL or cable Internet or live in places where broadband isn't available to, um, propel his company to success. After two false starts, maybe the third time really will be the charm for Propel.
What do you think? Would you pay $5 extra to get a 3x speed boost on your dial-up connection? Why not just bite the bullet and get DSL or cable? TalkBack to me below.