Robert_Charlton - 8:53 am on Mar 5, 2005 (gmt 0)

On a few test searches I ran, I see that their reliance on AOL for restaurant reviews is really unfortunate. The demographic is wa-a-ay wrong. A while back, when reviews were first discussed here, I was anticipating that, for restaurants, they might try to partner up with sites like Zagat (not that great, but better than AOL).

As for result defaults... not reviews, but geography... those San Francisco hotels that come up first definitely aren't in a part of town a lot of people would like to stay in, but many folks might not know that.

I continue to think that the distance from the center of the zip code default has got to be changed. Not sure to what, but alphabetical would almost be preferable because at least users understand that alphabetical sequence is artificial. Wouldn't plug in well to a map, though.

Maybe, for something like hotels, they could ask you to choose budget locations, downtown, tourist locations, etc... build a bit of guide book type sorting into the interface.

On various city hotel searches, it's fascinating to see some link pages that I know created cross-linking problems in Google's web algo are now serving as "references" to bring various hotel groups up to the top in local. Definitely a very different kind of algo.

What I'm seeing in Google Local looks like amazing progress on the surface, and I don't mean to minimize it... but deep down it may not yet really be that useful. The results aren't to the point yet where I'd be willing to depend on them.

I'm wondering whether local is likely to go through a long phase of basically superficial sortings until a critical mass of reviews and other local sorting criteria evolves... possibly eventually entailing XML type identifiers built into a wide range of sources.