[koolmoe]

There are many counterexamples to Godin's point, such as wikis. Wikis are useless with few contributors but extremely useful with many.

A secondary advantage (mentioned in pg's essays) is that you can more effectively take a "constant beta" mentality. Rolling a web app out to many users can certainly speed up the process of detecting bugs versus rolling it out to a few.

[nostrademons]

They aren't, though - I've used Wikis for teams of as little as 2 people. In fact, I'd argue that Wikis are more useful with few contributors: the sweet spot seems to be in the range of a couple dozen regular posters, and they tend to deteriorate (unless rigid procedures are put in place, like with Wikipedia) as the community grows beyond that.

[koolmoe]

A small base of contributors to a wiki may be OK for a very specific domain, but as the subject area broadens, you will certainly need more contributors or you risk having very superficial contributions on some of the subjects.

Anyway, if you don't like the wiki example, what about Digg or Reddit? How about a dating site? Something like epinions?

The notion that a product that isn't useful to a handful of users won't be useful to many users clearly is not one-size-fits-all.

[nostrademons]

Reddit, at least, was useful back when it was just spez and kn0thing submitting all their favorite links to their own site. There was a lot of nifty information there. Comments and the community didn't come until several months later.

I suspect several dating sites grew in a roundabout way. HotOrNot.com, for example, started as a site where you could submit photos and rate them. The initial appeal wasn't dating, and they didn't even offer it at first. It was just a lark: how much time can I waste at work, and how attractive do other people think I am.

I don't know the history on ePinions, but it looks like they could've grown in a similar fashion. Send it out to all your friends; have them share reviews on the products that they buy. It's a distributed, asynchronous form of calling your friend up and saying "Hey, you know of any good DVD players?"

A couple more examples of businesses that people think need many users, but actually were useful to small groups in their infancy:

I joined FaceBook in fall 2004 when it was just beginning to expand beyond the Ivy League. There was no messaging, no status, no platform, and no photo-sharing (they did have The Wall and Poke). Instead, people used it as a glorified directory. Meet someone at a party, and instead of having to ask their contact information, you just get their name and you can look it up on Facebook.

When Ning started, it was supposed to be a platform for creating social networks. At the time, everyone thought a social network was supposed to grow really big, get millions of users, and then get bought out and make everyone rich. Ning never really took off in this environment, because in order to get big you have to start small, and the large number of social networks hosted on it meant that each individual network had to compete for the same users.

I just checked again after reading some of Marc Andreesen's blog entries, and the current Ning incarnation is very different from when I had a friend consulting for them. It focuses much more on niche social networks, providing an online home for groups that number in the dozens, not millions. And judging from Marc's recent blog entries, it seems far more successful than the original incarnation of Ning.

[koolmoe]

When a site begins as one incarnation with a small user base and later transforms into something else when the user base has grown, it suggests that critical mass is necessary to make the transformation happen. If HotOrNot started out trying to build a dating site, they'd have been foolish not to consider scaling early on.

[nostrademons]

Right, the point is that if HotOrNot started out as a dating service, they wouldn't have gotten any users at all. People join a dating site to find other users to date - they only find other users to date if the site already has users. So you end up with a chicken-and-egg problem where everybody is waiting for everybody else to join. I'm curious how other dating sites have solved this - my guess is that they either tapped a large network of friends or they recruited users from an existing singles community.

HotOrNot actually faced a terrific scaling battle anyway: http://www.webtechniques.com/archives/2001/05/hong/. They got far more users by starting as a rating site than a dating site.

[koolmoe]

While I agree with your point, it's not the point of the blog entry:

"...if it doesn't work when you've got one, it's extremely unlikely to work when you have dozens."

Also, there is no mention of the difficulty of gaining users, merely that your idea won't be any better when you've got them.

That's clearly not true in several cases.