In the context of my briefing on Yelp for Business yesterday I asked Jeremy Stoppelman about the “filter by network” or friends directories (e.g., Loladex, GigPark, etc.) that are cropping up. He agreed that the idea was good in principle but that in practice it’s very hard to pull off.
He reminded me that very few people actually create content, despite the popularity of reviews, which typically means that your friends won’t have written anything about most of the businesses you’re interested in. He also said that services that ping your network when you have a request are also going to be challenging because some people aren’t going to want to receive emails all the time (see Facebook fatigue). Some of the Q&A services, however, can work quite well and I’m not sure I totally agree on this latter point.
Stoppelman added that Yelp started as an email-based system that was intended to query contacts for recommendations. He said that didn’t work for them. Hence the Yelp that exists today.
Screenwerk 
April 29, 2008 at 10:56 pm
Point of clarification…
Obviously Q&A services like Yahoo Answers can and do work well (more of a community of helpers, less of a real-life friend network).
April 29, 2008 at 11:08 pm
Sure.
April 30, 2008 at 2:05 am
Our (small) site, theSUGGESTR.com (formerly myfriendsuggests.com) has the ability to sort by the # of comments or ratings made by those in your friend network. That said I totally agree that the value of this feature is mitigated by the fact that so few people add reviews or complete their friend networks. On our site we’ve made the decision to allow ranking without having to leave a comment in order to encourage more people to at least leave there ranks.
One of the reasons we changed our name was to make our name more inline with where we see our most likely chance of success; that is in the suggestion area. We calculated personalized “predictions” of each place that a search returns based on a collaborative filtering algorithm. The “friend” notion no longer is included in our algorithm as we say it did not improve our results. We DO allow a user (who is logged in) to sort their search results by their predicted score. Another cool feature (if I do say so myself) is that our sort by is multi-dimensional, allowing a user to sort by predicted score THEN by # of comments by friends (for example). Not sure how much other people like this but I do!
April 30, 2008 at 7:02 am
I agree with Jeremy. We tried something similar on Facebook; where people could solicit recommendations from their friends. It is a tempting idea, because if it works, it would be extremely useful. So if you begin with the end in mind, it’s brilliant and valuable play. Getting there is another story.
July 2, 2008 at 9:58 pm
B. Chandra–what happened with your facebook experiment?
December 23, 2009 at 3:34 pm
[…] is because there weren’t enough reviews to make that functionality meaningful. Here’s my paraphrase of Yelp CEO Jeremy Stoppelman’s previous explanation of why they didn’t initially pursue […]