SMX Advanced: You & A with Matt Cutts

Filed in Conferences/Educ. by Matt McGee on June 4, 2007

SMX logo(This was posted as-it-happened during the session. I will try to do some clean-up later. Lots of juicy stuff in this, I would recommend reading carefully….)

Q – supplemental pages don’t come up when you search for unique phrases?
MC – pages in supplemental can come into main index if they get more links; supplemental index is not a penalty; we parse and index pages differently in supplemental — we may not index every word and every phrase if the page is in supplemental; we have to compress pages to store them in supplemental; if I had my way, we’d take the label off

Q – how confident are you when identifying person’s location?
MC – not my area of expertise, but we can have a fairly good idea (tells story about his hotel room

Q – why are webmaster guidelines so vague?
MC – don’t want to get in to having to mention every new scheme, saying “no, don’t do this”; easier to just say “avoid link schemes” rather than list each new scheme that people try; would like to get more detailed going forward

Q – buying links / what are you doing about it?
MC – wouldn’t be surprises if we add something to webmaster console allowing you to report paid links that way, like they can report spam; we do have existing algos and classifiers to handle this, but we’re not averse to using some manual intervention; some people think of Google as pure algorithm, no people involved; NYTimes article over the weekend showed we have a lot of people thinking about this; webmasters can do whatever they want, and we as a search engine can do whatever we want to return the best search results

Q – is linking out to authority sites a good idea?
MC – it’s good for your users, so it’s good for your web site; but I’m not gonna say we have a signal that measures that

Q – what about SERP pages appearing in the index?
MC – we said “avoid search result pages” in our technical guidelines, but not in our webmaster guidelines, because it’s a best practice; we do reserve right to remove pages that don’t add value; users don’t like search results appearing in search results; as a webmaster, look at the value of your search results pages — category pages are great; we do reserve the right to remove search pages, but we don’t consider it to be spamming necessarily

Q – impact of click-thru on rankings?
MC – it can affect personal search. and we won’t confirm or deny if it affects general search; if we were to use it, it would be very noisy; historically, I’ve been leery of using it because of the attempts to spam/game it;

Q – why does Google love Wikipedia?
MC – almost no one in this room is a regular user; my Mom would love Wikipedia; regular users do like Wikipedia a lot; (gives examples of when Wikipedia is right result); it can be a good fallback site on many searches

Q – can other domains you own affect your main site? why is it Google’s business what other sites I own when grading one site? (references PubCon site review story)
MC – I was trying to figure out what kind of webmaster the person was; if you have a gambling site and sweater site, it doesn’t mean the first will hurt the second;

Q (follow-up) – so, algorithmically, it has no impact?
MC – if you have spammy sites, it’s “fair game” to consider that when reviewing all your other sites; if you have 200 spammy sites, shouldn’t that be a signal when we look at your 201st site?

Q – thoughts on Mahalo?
MC – too early to say how it will do; this whole idea that people have nothing to do with search at Google is a mistake; all kinds of opportunity for human involvement — in webmaster guidelines, we’re changing wording where we say “100% algorithmic” process; it’s okay to say ‘we’ve used humans in many ways’

Q – how to categorize tens of thousands of products? and, how does one compete with one’s resellers?
MC -if a product is in many categories, try to figure out which one is best;

Q – more on algorithmic vs human intervention; what about the Stephen Colbert Googlebomb
MC – Googlebomb algorithm is completely algorithmic; we ran the algorithm again and it picked up the new phrases that were targeted; the Googlebomb algorithm runs every few months; we didn’t change the algorithm for this case, we just pressed the button to run it again;

Q – Danny follows about recent Bush/failure SERP
MC – there’s a component that runs every few months, and a component that is part of the live algorithm; the live part picked up that the word “failure” appeared back on Bush’s page

Q – image results in main SERPs; do a search for “retarded” and it shows people’s pictures
MC – we try to figure out what the user wants; we’ll take a look at the query you mentioned; we’re getting better at image analysis, but we’re not perfect — this is why Jimmy Carter’s photo might show up for “George Bush” search

Q – theming, LSI, similar words – how has that technology progressed?
MC – try it and see what works for you; we don’t confirm or deny the theories out there; in general, I think it’s a good practice to naturally work in synonyms and related words in your copy; we try a lot of different approaches at Google — we have different philosophies and work behind the scenes to figure out the best way; you can use a tilde to see related words to a query

END: Matt takes requests on what we’d like to see in Webmaster Central, and we’re done.

Comments are closed.