Why most ASP companies will fail
- By Steven J. Vaughan-Nichols, Sm@rt Reseller
19 May 2000 - Now, some of them will do work. The best and brightest will make it. The ones who specialize in a particular service line, like business contact management, should excel. The guys and gals who do Web hosting, e-mail serving or knowledge management and dress themselves with the sexy application-service-provider name will continue to make money. But, many of them--maybe even most--will be pushing up daisies.
The ones destined for the knacker's yard are the ones that are trying to host remote office applications. I'm sure of that because I've seen it happen before. Way, way back in the '70s and '80s, host-based computing was the only way to do office work. If you were in the mainframe world, you'd use IBM PROFS. For minicomputer users, VAX/VMS' All-in-One was your friend, and if you were using Unix, well, Sun CEO Scott McNealy is still happiest using EMACS and the like on a character-based VT-100 terminal emulation. If you were in host-based computing, ASP's ancestor, life was good.
Then in 1981, IBM rolled out a little product called the IBM PC. Now, the PC didn't completely kill host-based computing. To this day some law firms still use host-based, vertical Unix office suites from companies like the U.K.'s ICL. But, most host-based computing companies died and the few that remained were on life support.
The reasons why the PC ruled then still are valid today: They're faster.
Yes, I know all about the vast jumps in last mile broadband. That's why I think some ASPs will make it. For example, Citrix's MetaFrame and SCO's Tarantella impress me. That middleware technology makes it possible to take office applications and enable users to work with office applications remotely by using thin clients. But, and here's the important point, companies that partner with Citrix or SCO aren't in the business of trying to remotely put bread-and-butter office applications on servers. They're trying to use it to deliver programs that do well in a thin-client mode to customers.
Other companies, like Corel, Lotus and Microsoft, are pushing ASPs to deliver only slightly slimmed down fat clients over the Net. For example, Lotus' iNotes is about as fat a Java client as I've ever seen slow down my 384K SDSL line.
Do the figuring yourself. Word 2000 on my 667MHz Pentium III Dell with a quarter of a gig of memory (life's good when you're the Cyber Cynic) takes two seconds to come up. Great right? Now, cut it's size in half and it should take one second to get over the 100MBps bus to me. OK, now put that same application over a 384K SDSL connection. I'm looking at five minutes to get a radically slimmed down version of a bloody word processor on my desktop. My antique CP/M KayPro II with 4MHz of Z80 processor power and floppy disks loads WordStar 3.3, faster ... much faster.
ASP office application people like to argue that you get improved security because your applications are stored in their servers. That may be true, but on the other hand it also only takes one backhoe to dig up a fiber-optic line and there goes every application my office needs to run day to day.
Lack of speed and lack of local resources, put them together and you've got what killed off host-based computing. There's no reason to think that it won't happen again to fat-model ASP companies.
