4 ms·
The problem with Mechanical Turk is that you tend to get a lot of low-quality content. If people don't have an emotional investment in the site's success, they
by nostrademons 14y ago
The problem with Mechanical Turk is that you tend to get a lot of low-quality content. If people don't have an emotional investment in the site's success, they'll usually just write drivel and collect their fee. That sets a whole tone for the site which isn't really sustainable.
Common approaches I've seen for seeding a UGC site:
1.) Create lots of sock-puppets yourself and write/find the content that you yourself would want to read. Or get your friends to help. Example: Reddit.
2.) Write for your community. Don't worry about being big, just build something you and your friends would want to use. Once they start using it, they'll tell their friends, and you can get big later. Example: LiveJournal, HotOrNot, Fark, many niche sites.
3.) Scrape the hell out of your best competitor and then either provide a better user experience or SEO your site so you rank higher than them. Note that this is almost certainly against their TOS, is probably copyright infringement, and will get you on Google's shit-list, so you're in for a fair amount of pain if you're caught. Examples: Facebook, About.com, YouTube.
4.) Contact a couple celebrities in your field and see if they're willing to let you host their works, or if you can license their works with a revshare/partnership/whatever. All it takes is one really popular author to endorse you, and you'll get the halo effect of lots of less popular authors signing up because they want to be that one. Alternatively, be that celebrity yourself, so that people want to contribute to your site to bask in your reflected glow. Examples: JoelOnSoftware, Hacker News, StackOverflow.
5.) Be an aggregator first, nail your user-experience, and then once you have the visitors coming in, start adding your own hosting as a value-add. Examples: FriendFeed, many Google products.
6.) Throw money at the problem. Examples: Google+, Bing.
- blantonl 14y ago6.) Throw money at the problem. Examples: Google+, Bing. 5,4,3,2,1 all offer good advice. 6 could (will?) be a bottomless pit
- sophacles 14y agoI thought facebook seeded by selling itself as a service to be the directory++ for colleges et al, for a long time you needed a .edu email to sign up, and many colleges just signed people up as part of enrollment. Only later did they change to everyone, all the time (ca 2007 iirc).
- nostrademons 14y agoFacebook initially started by downloading the photos off the individual house websites at Harvard. They were banned pretty quickly, but it got them attention. I think they switched to user-uploaded photos fairly soon afterwards (though I think a coworker told me that they were still scraping photos while they were Harvard-only); they'd definitely stopped by the time they got to Amherst in Oct 2004.
- objclxt 14y agoThat was Facemash, rather than Facebook. Facemash was a 'hot or not' style site that Zuckerberg put together. I don't think Facebook ever actually scraped content for seeding purposes. To be honest, the rate stuff spreads over campus list-servs I really don't think it was required at that point.
- citricsquid 14y agoSpeaking of #5: A few days ago imgur announced they now have a host of extra (proper) "social" features, comments and the like. It made me realise that #5 is exactly what imgur has done, started out as a simple hosting service fuelled by the rising success of reddit and now they're branching out into their own community and removing the reliance on reddit one bit at a time. This definitely seems like the best approach.