One of the unfortunate effects of living in another country for almost five years is that you have to almost completely rebuild your knowledge of your home city. Specifically, I find that I need to find out where the best bars and restaurants are (because, honestly, there's only so much Crackbird a man can handle).
In theory, this is where things like Yelp and Menupages are supposed to come in. The internet hive-mind is supposed to work its magic. I should be able to shout "Yelp! What is best in life?" and it will tell me "To crush your enemies, see them driven before you, and to hear the lamentation of their women.". Instead, it says "Paulie's Pizza, Kilmainham Gaol and Croke Park."
I guess it's just a fundamental problem with crowdsourcing. Rather than helping the cream rise to the top, the noise generated by these sites actively drowns out useful information, making them useless. Even large sites like Amazon suffer from the same problem. I recently tried to buy a wireless access point for work. I checked out a few tech blogs and read reviews of some products. I finally settled on a Cisco product and went to Amazon to order it. Despite the almost entirely favorable reviews I'd read, the access point had only two and a half stars on Amazon. Turns out this was based on two reviews, the first of which was a one-star review with the person saying he'd had a problem with the technical support for another Cisco product. The other review was from Cisco themselves, giving the product five stars. The text of their 'review' was "if you have an issue with a product, please email us at $blah". Both reviews were useless and, if I'd been basing my purchase on the overall score of the product, I would have walked away.
More useful than the hours I've spent trawling Yelp and Menupages has been the one post I put up on Facebook, asking my friends where they'd recommend for places to eat. This way, I've immediately got context for each one of the places that have been recommended - this friend has impeccable taste, so I'll try their recommendation first etc. It's a similar reason why I trust Brian Lam's The Wire Cutter over the countless aggregation sites, or anything that relies on the average score of a large group of people to recommend technology. A sufficiently well-curated site run by a single person can still trump the wider internet.