Phil Deakins

Grabbing one-word names to "protect" a brand wasn't an unreasonable idea because it's the name that shows up in the results and not the display name. That was until wednesday when LL dealt with it. Those one-word names are no longer listed at the top. I doubt that any searcher would have clicked on an avatar's listing for say, "sex", when plenty of place listings were immedately below it but, if they were of any benefit in the results, the benefit was short-lived. If Jack, as the manager of the searchteam, did anything to create a need for his existance, he's run out of luck. He's been given his notice and will be gone soon.

Those #1 rankings won't do them any good, so let them play with their new toy. Ask yourself this:- if you were searching for shoes, would you click on an avatar link or on a store link?

I wouldn't worry about those names ranking #1. I can't believe that LL will leave it like that. Also, they aren't going to help those people's sales. Nobody will follow a trail via an avatar to find a place that sells what they are looking for when there are plenty of listings of such places immediately below the name's listing. Grabbing names for that purpose is and was a waste of time.

Jack: about the spam filter part:- You say that "Edits are typically reflected on the World page in a few minutes ..." but you are wrong. Edits used to be reflected on the page in a few minutes but not now. It now takes hours for an edit to be viewable in the page, and sometimes not until the next day. That alone makes your new system virtually unworkable for place owners. To start with, all the biases will be included in the one "boost" figure, so we won't know if a negative boost is caused by page content or by something else. We could work with that if we could do as you suggested - experiment with the words on the page and see the updated page and boost in minutes. But, since it takes at least hours for any edit, and its effect on the boost, to be seen, the new system will be virtually impossible for users to work with. It could take goodness knows how many days of testng to find out which part of the page is causing the negative boost, when it may not even be the content at all. That's *really* bad. Do you really want people to empty their pages altogether and spend goodness knows how many days of adding bit by bit to find out which bit causes a negative boost? Judging by your blog, you didn't know that pages now take such a long time before any edits are viewable. Get that back to minutes and it will be workable but, as long as an edit takes hours to be reflected in the page, the new boost system will be virtually unworkable for users. You say that "the nature of the objects on your parcel" can be a negative factor, but you don't say what the "natures" are and we will be blind. I can only come up with three "natures" - normal, not-for-sale and html in the line. Normal objects can't be faulted and html is stripped out of the data that's fed to the GSA, so there's no need to penalise that. That only leaves not-for-sale items and they can't be penalised because there are lots of freebie items around. Perhaps you'd like to tell us what you mean by "object natures" and what types of objects will cause a negative value. Even if there are types of objects that cause a negative value, why not simply leave them out of the page data that's fed to the GSA instead of giving the page a negative boost value because of them?