Google’s Matt Cutts: MetaFilter Hit By Previously Undisclosed Algorithm June 20, 2014 1:35 PM Subscribe
This article from June 12 has a lot of useful up-to-date info and context about how MetaFilter was penalized. There is also some discussion about a potential solution (although Matt Cutts is being typically cryptic) in coming "weeks or months."
What I think is really interesting is how much people in the search community like Barry Schwartz (the author here) and Danny Sullivan truly care and empathize with MetaFilter. It's also interesting to watch Matt Cutts dodge and deflect questions about what are "ranking factors" and "signals." It's a bit like Kremlinology trying to determine this stuff.
posted by KokuRyu at 1:45 PM on June 20, 2014 [14 favorites]
posted by KokuRyu at 1:45 PM on June 20, 2014 [14 favorites]
It's a bit like Kremlinology trying to determine this stuff.
I came in here to say the same thing.
posted by immlass at 1:58 PM on June 20, 2014
I came in here to say the same thing.
posted by immlass at 1:58 PM on June 20, 2014
Rhaomi's comment in the post on the Blue (The Mother of All Self-Links) links to several articles where Cutts says basically the same thing, over and over again.
posted by carsonb at 2:06 PM on June 20, 2014 [1 favorite]
posted by carsonb at 2:06 PM on June 20, 2014 [1 favorite]
Yeah, I've always called search optimization a cargo cult because Google never says anything super direct so people make wild guesses (like I had to make wild guesses about why MeFi was affected -- Matt Cutts says the backlink stuff didn't affect us). We've gotten tons of unsolicited advice about what to do with MetaFilter's SEO and so much of it was all over the map and often directly contradictory.
Matt Cutts and Google are in a really tough spot. Every non-committal word they say still ends up with hundreds of articles about how this might change everything in how people do SEO, and if they're more specific people pounce on it and do everything they can to try and make it their advantage.
And yeah, I think Danny Sullivan said it in the TLDR show that people are always saying "hey, come look at my site I got docked by Google unfairly, what gives?" and whenever he researches them he always finds something squirrelly that they were doing, but in MeFi's case he thought it really was an unfair penalty because he's seen everything under the sun and couldn't tell why MeFi would get docked for anything.
posted by mathowie (staff) at 2:06 PM on June 20, 2014 [17 favorites]
Matt Cutts and Google are in a really tough spot. Every non-committal word they say still ends up with hundreds of articles about how this might change everything in how people do SEO, and if they're more specific people pounce on it and do everything they can to try and make it their advantage.
And yeah, I think Danny Sullivan said it in the TLDR show that people are always saying "hey, come look at my site I got docked by Google unfairly, what gives?" and whenever he researches them he always finds something squirrelly that they were doing, but in MeFi's case he thought it really was an unfair penalty because he's seen everything under the sun and couldn't tell why MeFi would get docked for anything.
posted by mathowie (staff) at 2:06 PM on June 20, 2014 [17 favorites]
several articles where Cutts says basically the same thing, over and over again.
Yeah, Matt Cutts talked about it onstage at a search engine conference, and that's everyone trying to make heads and tails out of it. My twitter was blowing up that day as people were live-tweeting it. It's interesting to see how much the story changes slightly from blog to blog, it kind of shows how tough it is for the Google search team to keep their cards close but also try and say something about things that are going on. I think it's because Google is so vague that everyone tries and "reads the tea leaves" whenever Google does speak at all, to try and figure out what Google is really saying.
posted by mathowie (staff) at 2:08 PM on June 20, 2014 [3 favorites]
Yeah, Matt Cutts talked about it onstage at a search engine conference, and that's everyone trying to make heads and tails out of it. My twitter was blowing up that day as people were live-tweeting it. It's interesting to see how much the story changes slightly from blog to blog, it kind of shows how tough it is for the Google search team to keep their cards close but also try and say something about things that are going on. I think it's because Google is so vague that everyone tries and "reads the tea leaves" whenever Google does speak at all, to try and figure out what Google is really saying.
posted by mathowie (staff) at 2:08 PM on June 20, 2014 [3 favorites]
We've done a bunch of things we should have done months ago that is also hopefully going to help.
Out of curiosity, what are those things?
posted by grouse at 2:27 PM on June 20, 2014 [2 favorites]
Out of curiosity, what are those things?
posted by grouse at 2:27 PM on June 20, 2014 [2 favorites]
Wow, between this and the donations, we're going to be rich! *Daffy Duck voice* Rich, I tell you! Rich! Rich!
posted by Melismata at 2:35 PM on June 20, 2014 [2 favorites]
posted by Melismata at 2:35 PM on June 20, 2014 [2 favorites]
Out of curiosity, what are those things?
Mostly it's about removing duplicate content. People say you shouldn't have say, four versions of a post all indexed by Google. We have always tried to do that in the past, but weren't diligent about it. So we went around slapping a NOINDEX on all the tags pages, archives pages, category pages, and times where you paginate from results like the front page (hitting Older Posts a few times). Ideally, we should only have the 140k or so posts at www.metafilter.com showing up for searches on www.metafilter.com, and instead we had over 4 million results. So that makes it tough for google to match up searches with specific pages, since they might have the same text in several places.
In the end I don't even know if it helps but it certainly seems like it should, to make the index of stuff on metafilter as specific as possible. Theoretically, it seems like it would be easier to show someone every post about "competitive juggling" if there were only a handful of pages to choose from with those two words on them.
posted by mathowie (staff) at 2:41 PM on June 20, 2014 [7 favorites]
Mostly it's about removing duplicate content. People say you shouldn't have say, four versions of a post all indexed by Google. We have always tried to do that in the past, but weren't diligent about it. So we went around slapping a NOINDEX on all the tags pages, archives pages, category pages, and times where you paginate from results like the front page (hitting Older Posts a few times). Ideally, we should only have the 140k or so posts at www.metafilter.com showing up for searches on www.metafilter.com, and instead we had over 4 million results. So that makes it tough for google to match up searches with specific pages, since they might have the same text in several places.
In the end I don't even know if it helps but it certainly seems like it should, to make the index of stuff on metafilter as specific as possible. Theoretically, it seems like it would be easier to show someone every post about "competitive juggling" if there were only a handful of pages to choose from with those two words on them.
posted by mathowie (staff) at 2:41 PM on June 20, 2014 [7 favorites]
I've always found the duplication an annoyance when Googling for stuff specifically on MetaFilter, so it'll at least help out searchers, even if it doesn't help your numbers.
posted by grouse at 2:50 PM on June 20, 2014 [10 favorites]
posted by grouse at 2:50 PM on June 20, 2014 [10 favorites]
If the ad revenue does come back can we build a party palace?
posted by Justinian at 3:16 PM on June 20, 2014 [1 favorite]
posted by Justinian at 3:16 PM on June 20, 2014 [1 favorite]
Nevermind, something weird happened but it's gone now.
posted by KokuRyu at 3:22 PM on June 20, 2014
I keep seeing Google's unhelpfulness being cast virtuously as not giving MeFi "special treatment." Can someone please explain to me why Google not helping MeFi and/or assisting mathowie directly is being held as an ideal?
Google is not impartial, and it's not supposed to be. Its lack of impartiality is its defining feature! People use google precisely because it ranks search results in such a way that (mostly) reflects intuition about what constitutes a high-quality result. Its algorithms are meant to mimic what a human curator would do if he had the time, patience, and ability to trawl the entire internet. And not all of it is purely algorithmic: google's "manual actions" involve human intervention to penalize sites that had gotten false negatives from the filter algorithms.
So here is an instance where their algorithm is failing. There is no question in any domain that MeFi's rankings should be good; nobody has indicated any material problem with the quality of MeFi. Instead, everyone is speculating about what meaningless, esoteric detail google's algorithm over-reacted to. It's like some horrible combination of cargo-cult science and the sort of walking-on-eggshells that people do around mentally unstable abusers. But it's not a problem with MeFi; it's a problem with Google's Replicant not behaving like a real human.
Google can fix the MeFi false positive just as they fix the false negatives. There may be other sites hit by the 2012-11-17 update which are more questionable, and clearly google can't investigate them all as one-off instances, but what sense is there in continuing to allow an already-known, well-established bad result [MeFi's deranking] while working on solving the underlying issue? Sure, some people will complain that MeFi got a quick one-off patch while their site has to wait for the global fix, and that's not impartial. But Google isn't impartial -- the tweaks to its filters and its manual actions happen because it needs to be non-impartial! -- so their slavish adherence to unhelpfully maintaining the bad output in the name of "no special treatment" seems... off.
You reach down and flip the green over on its back. MeFi lays on its back, its belly baking in the hot sun, beating its legs trying to turn itself over. But it can't. Not with out your help. But you're not helping. Why is that, Google?
posted by Westringia F. at 3:42 PM on June 20, 2014 [19 favorites]
Google is not impartial, and it's not supposed to be. Its lack of impartiality is its defining feature! People use google precisely because it ranks search results in such a way that (mostly) reflects intuition about what constitutes a high-quality result. Its algorithms are meant to mimic what a human curator would do if he had the time, patience, and ability to trawl the entire internet. And not all of it is purely algorithmic: google's "manual actions" involve human intervention to penalize sites that had gotten false negatives from the filter algorithms.
So here is an instance where their algorithm is failing. There is no question in any domain that MeFi's rankings should be good; nobody has indicated any material problem with the quality of MeFi. Instead, everyone is speculating about what meaningless, esoteric detail google's algorithm over-reacted to. It's like some horrible combination of cargo-cult science and the sort of walking-on-eggshells that people do around mentally unstable abusers. But it's not a problem with MeFi; it's a problem with Google's Replicant not behaving like a real human.
Google can fix the MeFi false positive just as they fix the false negatives. There may be other sites hit by the 2012-11-17 update which are more questionable, and clearly google can't investigate them all as one-off instances, but what sense is there in continuing to allow an already-known, well-established bad result [MeFi's deranking] while working on solving the underlying issue? Sure, some people will complain that MeFi got a quick one-off patch while their site has to wait for the global fix, and that's not impartial. But Google isn't impartial -- the tweaks to its filters and its manual actions happen because it needs to be non-impartial! -- so their slavish adherence to unhelpfully maintaining the bad output in the name of "no special treatment" seems... off.
You reach down and flip the green over on its back. MeFi lays on its back, its belly baking in the hot sun, beating its legs trying to turn itself over. But it can't. Not with out your help. But you're not helping. Why is that, Google?
posted by Westringia F. at 3:42 PM on June 20, 2014 [19 favorites]
Bearing in mind that Google said, "Whoops, our bad," and something did go wrong back in 2012, the algorithm (which is approaching AI actually) generally *works*.
Things have changed so much since 2012, too. While the Penguin and Panda updates have penalized the spammy black-hat's that made Google Search so questionable up until 2012, the real interesting thing is Hummingbird and the shift to semantic search.
We don't even focus much (beyond technical best practices, plus some stuff schema markup) on "optimizing" pages for search, since Google is so obviously moving towards a semantic model of search that doesn't even depend on keywords.
Instead, we focus on creating great content and on-page user experience. It's very exciting because, a year from now, competition will be all about content. Period.
So I never pay very much attention to Cutts' pronouncements because the key message is "quality." That's it.
However, like in many totalitarian frameworks, "mistakes will happen" like what happened with MetaFilter.
posted by KokuRyu at 3:58 PM on June 20, 2014
Things have changed so much since 2012, too. While the Penguin and Panda updates have penalized the spammy black-hat's that made Google Search so questionable up until 2012, the real interesting thing is Hummingbird and the shift to semantic search.
We don't even focus much (beyond technical best practices, plus some stuff schema markup) on "optimizing" pages for search, since Google is so obviously moving towards a semantic model of search that doesn't even depend on keywords.
Instead, we focus on creating great content and on-page user experience. It's very exciting because, a year from now, competition will be all about content. Period.
So I never pay very much attention to Cutts' pronouncements because the key message is "quality." That's it.
However, like in many totalitarian frameworks, "mistakes will happen" like what happened with MetaFilter.
posted by KokuRyu at 3:58 PM on June 20, 2014
I agree, Westringia F. I've kind of seen this whole thing as really embarrassing for Google and the quality of their update. I mean, if you could pick the one website that is the very epitome of collaborative problem-solving and not-spam, this would be it. Except that they don't see it that way. In fact, they seem to be digging their heels in and saying that, as a policy, they won't display favoritism or do anything about it. Which is, I think, all the more embarrassing, right?
Like, if you were a knife maker — say, the world's BEST knifemaker…like you kind of invented knives level of knifemaking — and you made this new knife that had a super fancy blade and you were claiming that it could cut the toughest things in the world. Turned out it did, yay! But then kinda got stuck on butter. And the cheese and tomato people are saying, "us too!" But you, being the inventer of knives, were all, "We're not partial to any foods. Our knives cut the way they're supposed to."
It's just mind-bogglingly dumb to me.
posted by iamkimiam at 3:58 PM on June 20, 2014 [1 favorite]
Like, if you were a knife maker — say, the world's BEST knifemaker…like you kind of invented knives level of knifemaking — and you made this new knife that had a super fancy blade and you were claiming that it could cut the toughest things in the world. Turned out it did, yay! But then kinda got stuck on butter. And the cheese and tomato people are saying, "us too!" But you, being the inventer of knives, were all, "We're not partial to any foods. Our knives cut the way they're supposed to."
It's just mind-bogglingly dumb to me.
posted by iamkimiam at 3:58 PM on June 20, 2014 [1 favorite]
I'd say Westringia is because it's not impartial in a sense that it filters off specific characteristics and similar factors but not because it focuses on specific sites (at least officially if I understand correctly). Not saying it's right, but I think that's the logic.
posted by Carillon at 3:59 PM on June 20, 2014
posted by Carillon at 3:59 PM on June 20, 2014
I think it boils down to this:
-Metafilter, and probably some other sites, are false-positives on a tweak/filter to Google's search algorithms;
-if Google is seen to be giving Metafilter "special treatment" in terms of helping them tweak the site to avoid this false-positive result, they are going to have a metric shit-ton of other websites demanding the same treatment, advice, and support - only a few of whom might be legitimately affected by the same problem as Metafilter;
-that doesn't speak to the fact that such an action would be, essentially, showing a site how to avoid some of the rules in Google's algorithm - in effect, giving away part of Google's secret recipe - while not correcting the larger problem, which is that a rule is creating false-positives. Google's business interest is in the rule being correct, not in helping people finesse or dodge the rule.
-if Google reaches out to Metafilter to say "hey, we know you are a false-positive as a result of these previously unacknowledged changes we made to the algorithm, can you help us identify what about your site would've tripped these?" would essentially involve Google sharing a least a part of their secret recipe with outsiders.
So, they've done what they can - acknowledged that a change has created false-positives, that Metafilter is one of those false-positives, and that they are working to correct that. But they have to do it by themselves, and without showing any favoritism, without teaching people how to avoid the rule. They have to make it better, not bypassed.
posted by nubs at 4:03 PM on June 20, 2014 [4 favorites]
-Metafilter, and probably some other sites, are false-positives on a tweak/filter to Google's search algorithms;
-if Google is seen to be giving Metafilter "special treatment" in terms of helping them tweak the site to avoid this false-positive result, they are going to have a metric shit-ton of other websites demanding the same treatment, advice, and support - only a few of whom might be legitimately affected by the same problem as Metafilter;
-that doesn't speak to the fact that such an action would be, essentially, showing a site how to avoid some of the rules in Google's algorithm - in effect, giving away part of Google's secret recipe - while not correcting the larger problem, which is that a rule is creating false-positives. Google's business interest is in the rule being correct, not in helping people finesse or dodge the rule.
-if Google reaches out to Metafilter to say "hey, we know you are a false-positive as a result of these previously unacknowledged changes we made to the algorithm, can you help us identify what about your site would've tripped these?" would essentially involve Google sharing a least a part of their secret recipe with outsiders.
So, they've done what they can - acknowledged that a change has created false-positives, that Metafilter is one of those false-positives, and that they are working to correct that. But they have to do it by themselves, and without showing any favoritism, without teaching people how to avoid the rule. They have to make it better, not bypassed.
posted by nubs at 4:03 PM on June 20, 2014 [4 favorites]
mathowie: "So we went around slapping a NOINDEX on all the tags pages, archives pages, category pages, and times where you paginate from results like the front page"
In an abstract sense, that kinda sounds like it should be Google's problem, though. Some searchers might be better served by the tag pages, archive pages, etc. Maybe Google would be able to tell — based on weird cookie voodoo — that you're looking for detailed discussion about a particular term, or for a random assortment of interesting links for a particular day. But now they won't be able to serve that, because that information has been marked irrelevant at their command.
I understand that that's not really a juicy thing for them to encourage: it's very niche-oriented, and at the same time is very easy to exploit. I also understand that saying something is "Google's problem" is like saying global climate change is "Earth's problem" in terms of usefulness.
mathowie: "in MeFi's case he thought it really was an unfair penalty because he's seen everything under the sun and couldn't tell why MeFi would get docked for anything"
God, that is an perfect description of the problem. Every time Google changes their algorithm, the entitled singularity of the collective SEO consciousness goes crazygonuts about it. So of course Google gets lots of reports about false positives... which they then diligently investigate, only to find that they were in fact justified in saying that
What they don't see is all the legitimate sites, who spend their money on community and content and quality rather than paying people to worship and pray to Our Lord Google. Most of those sites end up fine, because for the most part the algorithms are accurately reflecting changes in the content, um, spectrum. The others, though, are silent victims: for the most part, they wouldn't be able to enact the "culture of SEO" without ruining their actual culture, so they're left rearranging deckchairs on the Lusitania.
There is a complete meta-meta-meta game going on here, and that's before "Meta" Filter even enters the picture. This is the paradox of codifiable content ranking... there's so much more content, and so much more good content, but somehow our ability to distinguish between the two remains stagnant...
... and it's because the people implementing the rankings have their own degree of epistemic closure, and the only people who can break them out of it are incentivized to stay the course.
I understand that Google increasingly thinks of itself as a private company like all other private companies, but they are losing sight of their history: they succeeded by becoming a public service. It's regrettable that that success was predicated upon ad revenue, but no other option existed at the time.
Nevertheless, Google is losing its points of differentiation: doing ten genuinely weird things because they're brilliant and no one has ever tried them; hiring brilliant people and giving them the flexibility to innovate; fighting irrational fights because they're the right things to do... this is what public services should be, but since we've abdicated the public sphere it was nice for a while that a company would just do those things, powered by spunk and idealism and ad sales.
I give them a great deal of credit; unlike Microsoft, their decline at least leaves us with open source languages and libraries, exportable data that can then be recreated, and a lingering memory of spirited tech capitalism.
However, on behalf of our former mods: fuck you, Google, whether it's still your motto or not, try harder to not be evil.
posted by Riki tiki at 4:06 PM on June 20, 2014 [6 favorites]
In an abstract sense, that kinda sounds like it should be Google's problem, though. Some searchers might be better served by the tag pages, archive pages, etc. Maybe Google would be able to tell — based on weird cookie voodoo — that you're looking for detailed discussion about a particular term, or for a random assortment of interesting links for a particular day. But now they won't be able to serve that, because that information has been marked irrelevant at their command.
I understand that that's not really a juicy thing for them to encourage: it's very niche-oriented, and at the same time is very easy to exploit. I also understand that saying something is "Google's problem" is like saying global climate change is "Earth's problem" in terms of usefulness.
mathowie: "in MeFi's case he thought it really was an unfair penalty because he's seen everything under the sun and couldn't tell why MeFi would get docked for anything"
God, that is an perfect description of the problem. Every time Google changes their algorithm, the entitled singularity of the collective SEO consciousness goes crazygonuts about it. So of course Google gets lots of reports about false positives... which they then diligently investigate, only to find that they were in fact justified in saying that
http://www.tenwierdthin.gs/ probably wasn't contributing a lot of meaning to the web.What they don't see is all the legitimate sites, who spend their money on community and content and quality rather than paying people to worship and pray to Our Lord Google. Most of those sites end up fine, because for the most part the algorithms are accurately reflecting changes in the content, um, spectrum. The others, though, are silent victims: for the most part, they wouldn't be able to enact the "culture of SEO" without ruining their actual culture, so they're left rearranging deckchairs on the Lusitania.
There is a complete meta-meta-meta game going on here, and that's before "Meta" Filter even enters the picture. This is the paradox of codifiable content ranking... there's so much more content, and so much more good content, but somehow our ability to distinguish between the two remains stagnant...
... and it's because the people implementing the rankings have their own degree of epistemic closure, and the only people who can break them out of it are incentivized to stay the course.
I understand that Google increasingly thinks of itself as a private company like all other private companies, but they are losing sight of their history: they succeeded by becoming a public service. It's regrettable that that success was predicated upon ad revenue, but no other option existed at the time.
Nevertheless, Google is losing its points of differentiation: doing ten genuinely weird things because they're brilliant and no one has ever tried them; hiring brilliant people and giving them the flexibility to innovate; fighting irrational fights because they're the right things to do... this is what public services should be, but since we've abdicated the public sphere it was nice for a while that a company would just do those things, powered by spunk and idealism and ad sales.
I give them a great deal of credit; unlike Microsoft, their decline at least leaves us with open source languages and libraries, exportable data that can then be recreated, and a lingering memory of spirited tech capitalism.
However, on behalf of our former mods: fuck you, Google, whether it's still your motto or not, try harder to not be evil.
posted by Riki tiki at 4:06 PM on June 20, 2014 [6 favorites]
And I'm not trying to defend Google, just trying to understand the situation from their viewpoint.
posted by nubs at 4:08 PM on June 20, 2014
posted by nubs at 4:08 PM on June 20, 2014
You reach down and flip the green over on its back. MeFi lays on its back, its belly baking in the hot sun, beating its legs trying to turn itself over. But it can't. Not with out your help. But you're not helping. Why is that, Google?
This makes me sad, because now I think Metafilter is a turtle, which is sad enough, but also it may have salmonella which I can catch if I handle it.
:-(
posted by potsmokinghippieoverlord at 5:35 PM on June 20, 2014 [4 favorites]
This makes me sad, because now I think Metafilter is a turtle, which is sad enough, but also it may have salmonella which I can catch if I handle it.
:-(
posted by potsmokinghippieoverlord at 5:35 PM on June 20, 2014 [4 favorites]
I worked at Google, albeit some eight years ago and not in search. But I got to know Matt and in general how the search team works and I think the ethos hasn't changed.
Matt is a decent and honest guy. He's also very careful and diplomatic. One of his great strengths is he's not some PR droid, he's honestly an engineer/manager deeply involved in search rankings. He also stepped up and volunteered to be a sort of public liaison for Google, particularly to the SEO community, and he's done a great and sincere job of it. But it's an impossible position. Sometimes I think his real job is to just give the appearance of transparency so webmasters don't freak out too much. But he's careful to try to never say things that are simply not true, and he does try to give guidance when he can.
Google's attempt at algorithmic neutrality is an important principle that is recognized by everyone to not be entirely possible. But I think they do a pretty credible job of it. To me the most tenuous moment was a few years ago when the Panda update demoted content farms like Demand Media and Mahalo. I feel certain they were careful to not target any specific company, but there's no doubt Panda wiped out a whole type of company. I feared Metafilter was collateral damage in that update, it's very interesting that now two years later we've discovered that in fact it was some other unknown Google change about the same time.
The biggest instability in this ecosystem right now is the way Google has near-monopoly power in search, keyword targeted ads, and content-targeted ads. I think Google is generally honest in not trying to crudely leverage that market power. But they have concentrated a lot of market power and in some ways their position is a natural monopoly. There's been a fair amount of anti-trust investigation in Europe (recently settled), nothing meaningful yet in the US despite Microsoft's best lobbying efforts. But I suspect it's coming at some point.
posted by Nelson at 5:51 PM on June 20, 2014 [10 favorites]
Matt is a decent and honest guy. He's also very careful and diplomatic. One of his great strengths is he's not some PR droid, he's honestly an engineer/manager deeply involved in search rankings. He also stepped up and volunteered to be a sort of public liaison for Google, particularly to the SEO community, and he's done a great and sincere job of it. But it's an impossible position. Sometimes I think his real job is to just give the appearance of transparency so webmasters don't freak out too much. But he's careful to try to never say things that are simply not true, and he does try to give guidance when he can.
Google's attempt at algorithmic neutrality is an important principle that is recognized by everyone to not be entirely possible. But I think they do a pretty credible job of it. To me the most tenuous moment was a few years ago when the Panda update demoted content farms like Demand Media and Mahalo. I feel certain they were careful to not target any specific company, but there's no doubt Panda wiped out a whole type of company. I feared Metafilter was collateral damage in that update, it's very interesting that now two years later we've discovered that in fact it was some other unknown Google change about the same time.
The biggest instability in this ecosystem right now is the way Google has near-monopoly power in search, keyword targeted ads, and content-targeted ads. I think Google is generally honest in not trying to crudely leverage that market power. But they have concentrated a lot of market power and in some ways their position is a natural monopoly. There's been a fair amount of anti-trust investigation in Europe (recently settled), nothing meaningful yet in the US despite Microsoft's best lobbying efforts. But I suspect it's coming at some point.
posted by Nelson at 5:51 PM on June 20, 2014 [10 favorites]
Hi, this is Matt Cutts and I'm happy to confirm what MattH said.
"Happy" to confirm Google isn't offering any special help to rectify a special problem they caused for a truly unique and special website. Glad he's so happy.
posted by Conrad Cornelius o'Donald o'Dell at 7:20 PM on June 20, 2014
"Happy" to confirm Google isn't offering any special help to rectify a special problem they caused for a truly unique and special website. Glad he's so happy.
posted by Conrad Cornelius o'Donald o'Dell at 7:20 PM on June 20, 2014
Rhaomi's comment in the post on the Blue (The Mother of All Self-Links) links to several articles where Cutts says basically the same thing, over and over again.
Sorry about that -- it was so redundant because there wasn't yet any video or transcript from the conference, so I threw in as many eyewitness reports as I could find to capture any tidbits of info that might have been left out of any one reaction.
Anyway, the video of the whole interview is up now; here are the relevant remarks from the ~5 minute mark:
Sorry about that -- it was so redundant because there wasn't yet any video or transcript from the conference, so I threw in as many eyewitness reports as I could find to capture any tidbits of info that might have been left out of any one reaction.
Anyway, the video of the whole interview is up now; here are the relevant remarks from the ~5 minute mark:
Danny Sullivan: What happened to MetaFilter? This is the long-time site that was sort of like the Digg before we had Digg, and the short backstory for people who don't know, it's like a longstanding site -- a lot of people were like, "Oh my God, I grew up on MetaFilter! It was like Digg before we had Digg, before we had Reddit, before we had whatever." And you know, Matt, who runs that site, he didn't know what happened to him. And I looked at it, and it sounded like he got hit by a Panda update, except he would have had to get hit by a Panda update at a time when you guys, Google, told us there was no Panda update! And then he also said that he felt like there was a lot of, maybe the ads were coming into play with it. And I know Panda had an ad component, but is it stronger than we thought? So, what happened to them, was it Panda, and are ads a bigger deal in Panda than we thought?posted by Rhaomi at 7:25 PM on June 20, 2014 [10 favorites]
Matt Cutts: Well, it wasn't Panda, so... no, MetaFilter was not affected by Panda, so we'll just take that off the table immediately. So, what happened with MetaFilter was interesting because -- he wrote a really good post, and let me just say I have a ton of respect for Matt Haughey. You know, you were in an On The Media piece about MetaFilter that just came out recently that was really interesting, and so... MetaFilter is your typical high-quality site. Now, it hasn't had a refresh of its design for quite a while, so it looks a little like Slashdot, you know, 1999-era sort of stuff. But the content itself typically has high quality. So, it was affected by an algorithmic update, it was not Panda or Penguin, so it was a different algorithm entirely. We launch a lot of different --
DS: A secret algorithm...
MC: There are a lot of different algorithms that we launch. And we take those reports as good feedback. So in particular we've got engineers thinking about, are there other signals that we could use to differentiate MetaFilter from the sites that should be affected by the algorithm. Now, we've actually been in touch with Matt, you know, since the original algorithm touched the site, and so we had not yet found any signals that would help. I think we actually have a good lead on that now, but it's probably still going to be a matter of weeks to maybe a month or two before we can see whether that pans out.
I think one of the things that muddied the water a little bit with MetaFilter is Matt Haughey wrote a blog post and he talked about all the link removal requests that he gets. And so there was one aspect to it where Matt was under the impression that Google thought that MetaFilter was spammy or a source of spammy links or we sent out example links to other people saying these links on metafilter.com are spam. And that's not what happened.
So, we looked through our records -- we have, to the best of my knowledge, never sent a message or notification to any webmaster...[Sullivan coughs, minor derail] So we had never sent a notification to any website that said a link on MetaFilter was spammy. Rather, what they were seeing were overzealous people who had been spamming and who happened to have a link on MetaFilter trying to clean it up. And the crazy thing is, he actually quotes an email that looks like it comes from the Google spam team, and we hunted down which site it was that emailed him, and it turns out we had notified that site but we had given them a different link that was spammy...
DS: And they changed it up!
MC: ...and they took the template and they started to say, "Google spam team said this--"
DS: You should ban them! Oh wait, you already did...
MC: Well... I won't go into the details of that site...
DS: Can you put them on double secret probation?
MC: Double-secret probation, yes... [laughs] So, that site, certainly we'll scrutinize, but they were doing a pretty deceptive thing. And so Matt Haughey was like, "Oh, Google says we're spammy!" And we had definitely not said anything to that effect. Now the good thing is, though MetaFilter is a high-quality site and though its rankings have been affected to some degree by the algorithmic ranking changes, the fact that it makes such big news means -- usually we get things right. Now, in this case I think we could do better, and we're gonna try to do better.
It's not just google. I spent days trying to figure out Apple's "reader" pagination so I could allow readers of my sites to get the entire article at once. Instead they only get the first half. I tried to get Apple to help and was told there was no way they were revealing how that work because if they did everyone would do what I didn't want to do.
posted by cjorgensen at 7:43 PM on June 20, 2014 [1 favorite]
posted by cjorgensen at 7:43 PM on June 20, 2014 [1 favorite]
MetaFilter is your typical high-quality site. Now, it hasn't had a refresh of its design for quite a while,
If by "quite a while" you really mean "ever", then we agree.
posted by medusa at 8:04 PM on June 20, 2014 [4 favorites]
If by "quite a while" you really mean "ever", then we agree.
posted by medusa at 8:04 PM on June 20, 2014 [4 favorites]
Now, it hasn't had a refresh of its design for quite a while,
and Cutts has repeated that mantra 3-4 times, leading one to suspect that Google's "secret algorithm" is using certain "Web 3.0 Design" standards in a way that Design starts to trump Content... which is as crappy a possible reason as I can think of.
posted by oneswellfoop at 8:41 PM on June 20, 2014 [30 favorites]
and Cutts has repeated that mantra 3-4 times, leading one to suspect that Google's "secret algorithm" is using certain "Web 3.0 Design" standards in a way that Design starts to trump Content... which is as crappy a possible reason as I can think of.
posted by oneswellfoop at 8:41 PM on June 20, 2014 [30 favorites]
Good point. Why should a site's design matter to Google? Could Google's algorithms even theoretically figure out which sites are "modern" and which are "dated" and which are "really, really old"? That would be pretty amazing if so, but I guess you can't put much past Google. Though ifthings like that are really affecting their search results, then sheez, that's ridiculous.
posted by Conrad Cornelius o'Donald o'Dell at 9:44 PM on June 20, 2014
posted by Conrad Cornelius o'Donald o'Dell at 9:44 PM on June 20, 2014
A site's design should probably matter to google exactly to the extent that it matters to the people performing the search on google. For example when I get a search result that leads to one of those list-of-40-things-on-40-separate-pages or 10 images in a slideshow format or whatever I will close that tab immediately and move on to a new result. So for me I would love it if google would deprecate those formats and drop them way down the result list.
So, yeah, I can see his point. If people who get metafilter upon searching immediately close the tab because they can't stand metafilter's design that would be something google could reasonably take into account. But I have no idea if that's what he was indicating with those comments or if it was just some kinda offhand remark or what. But site design is a perfectly valid reason to move results up or down google's search rankings.
Google's job isn't to feed people the results they should want, it is to feed people the results they do want. Within reason, of course.
posted by Justinian at 10:00 PM on June 20, 2014 [6 favorites]
So, yeah, I can see his point. If people who get metafilter upon searching immediately close the tab because they can't stand metafilter's design that would be something google could reasonably take into account. But I have no idea if that's what he was indicating with those comments or if it was just some kinda offhand remark or what. But site design is a perfectly valid reason to move results up or down google's search rankings.
Google's job isn't to feed people the results they should want, it is to feed people the results they do want. Within reason, of course.
posted by Justinian at 10:00 PM on June 20, 2014 [6 favorites]
Also if your page doesn't load almost instantly I will close the tab and move on. So I would love it if google would drop results which load too slowly. Maybe they already do.
posted by Justinian at 10:01 PM on June 20, 2014 [1 favorite]
posted by Justinian at 10:01 PM on June 20, 2014 [1 favorite]
So, yeah, I can see his point. If people who get metafilter upon searching immediately close the tab because they can't stand metafilter's design that would be something google could reasonably take into account.
Well, and usability and fashion aside, honestly a site's design not substantially changing over most of its 15-year lifespan (?!) could be construed algorithmically as a sign that the site is not a living concern with an active userbase, but that it's only being updated through spambots or whatever. I know the design here is comfort food to the hardcore users and therefore changing it is That Which Can Never Happen (and I don't say that as a dismissal, I very much value comfort food), but MeFi's design stretches way back to when webrings were a thing, and I cannot think of a single website that hasn't undergone a substantial redesign in the last decade other than this place. I can see how someone could come up with an algorithm to read something like that as a potential sign that the place is a ghost town.
But I have no idea if that's what he was indicating with those comments or if it was just some kinda offhand remark or what.
Given that he's brought it up a few times, I doubt it's just an offhand comment, but who knows.
posted by middleclasstool at 10:24 PM on June 20, 2014 [3 favorites]
Well, and usability and fashion aside, honestly a site's design not substantially changing over most of its 15-year lifespan (?!) could be construed algorithmically as a sign that the site is not a living concern with an active userbase, but that it's only being updated through spambots or whatever. I know the design here is comfort food to the hardcore users and therefore changing it is That Which Can Never Happen (and I don't say that as a dismissal, I very much value comfort food), but MeFi's design stretches way back to when webrings were a thing, and I cannot think of a single website that hasn't undergone a substantial redesign in the last decade other than this place. I can see how someone could come up with an algorithm to read something like that as a potential sign that the place is a ghost town.
But I have no idea if that's what he was indicating with those comments or if it was just some kinda offhand remark or what.
Given that he's brought it up a few times, I doubt it's just an offhand comment, but who knows.
posted by middleclasstool at 10:24 PM on June 20, 2014 [3 favorites]
I sure like that MetaFilter loads quickly and isn't junked up with extraneous triple sidebars and stuff. So many sites are so hard to read after redesign...I'm looking at you, FiveThirtyEight.
posted by leahwrenn at 11:55 PM on June 20, 2014 [4 favorites]
posted by leahwrenn at 11:55 PM on June 20, 2014 [4 favorites]
This makes me sad, because now I think Metafilter is a turtle, which is sad enough, but also it may have salmonella which I can catch if I handle it.
Hey, at least MeFi isn't an armadillo. Those things can give you leprosy.
posted by Pope Guilty at 3:26 AM on June 21, 2014
Hey, at least MeFi isn't an armadillo. Those things can give you leprosy.
posted by Pope Guilty at 3:26 AM on June 21, 2014
Hopefully they get it together.
posted by cashman
Fine with me if they get it separately. Just as long as they do get it.
posted by spitbull at 4:29 AM on June 21, 2014
posted by cashman
Fine with me if they get it separately. Just as long as they do get it.
posted by spitbull at 4:29 AM on June 21, 2014
And not all of it is purely algorithmic: google's "manual actions" involve human intervention to penalize sites that had gotten false negatives from the filter algorithms.
Yeah, I don't understand why the Google system of ethics permits manipulation of the results of their algorithms to penalize sites but not to improve their ranks.
posted by grouse at 7:56 AM on June 21, 2014
Yeah, I don't understand why the Google system of ethics permits manipulation of the results of their algorithms to penalize sites but not to improve their ranks.
posted by grouse at 7:56 AM on June 21, 2014
So google makes changes that they admit had inadvertent negative effects for mefi and there is an actual and measurable drop in income as a consequence. When might this sort of action and adverse consequences reach a level where a case for damages could be credibly brought?
posted by peacay at 9:49 AM on June 21, 2014
posted by peacay at 9:49 AM on June 21, 2014
It's not like putting AdSense text ads on your site automatically entitles you to income, right? AdSense is not a public utility. It is a platform run by a private corporation.
It has been more than 24 months since the "algo mistake" that led to a drop in traffic; the online environment changes very fast, typically from quarter to quarter.
While I'm sure the sheer volume of traffic would mean MetaFilter would get the clicks needed to generate income, on the other hand these days the only people really benefiting from AdSense are the folks at Google itself.
It's a tough, competitive market for content "publishers" (people hosting AdSense ads with their content) these days hoping to generate income from AdSense; done right, using AdSense is essentially a form of affiliate marketing, a high pressure environment if there ever was one.
There is no such thing as a sure thing, and I think the 2012 drop in traffic and the trend in declining ad revenues for the site should be considered as separate phenomena.
posted by KokuRyu at 10:24 AM on June 21, 2014 [1 favorite]
It has been more than 24 months since the "algo mistake" that led to a drop in traffic; the online environment changes very fast, typically from quarter to quarter.
While I'm sure the sheer volume of traffic would mean MetaFilter would get the clicks needed to generate income, on the other hand these days the only people really benefiting from AdSense are the folks at Google itself.
It's a tough, competitive market for content "publishers" (people hosting AdSense ads with their content) these days hoping to generate income from AdSense; done right, using AdSense is essentially a form of affiliate marketing, a high pressure environment if there ever was one.
There is no such thing as a sure thing, and I think the 2012 drop in traffic and the trend in declining ad revenues for the site should be considered as separate phenomena.
posted by KokuRyu at 10:24 AM on June 21, 2014 [1 favorite]
and Cutts has repeated that mantra 3-4 times, leading one to suspect that Google's "secret algorithm" is using certain "Web 3.0 Design" standards in a way that Design starts to trump Content... which is as crappy a possible reason as I can think of.
This was my guess during the initial revelation about traffic falloff. It coincides with a huge shift in web design UI standards. I wouldn't be surprised if this update penalized or rewarded sites for certain design in an attempt to... Maybe keep first page results look fresh and modern thus making google look like they're serving the newest and best info? I'm sure there were better reasons involved but I've been wondering if metafilter's lack of site design changes, lack of images, etc., played into all of this.
posted by milarepa at 11:14 AM on June 21, 2014 [1 favorite]
This was my guess during the initial revelation about traffic falloff. It coincides with a huge shift in web design UI standards. I wouldn't be surprised if this update penalized or rewarded sites for certain design in an attempt to... Maybe keep first page results look fresh and modern thus making google look like they're serving the newest and best info? I'm sure there were better reasons involved but I've been wondering if metafilter's lack of site design changes, lack of images, etc., played into all of this.
posted by milarepa at 11:14 AM on June 21, 2014 [1 favorite]
It seems most likely to me that a penalty for outdated (so called) web design is just down to similarity. Metafilter, as Google's algorithms see it, has web design that's similar to a lot of old, junky, spammy websites, and that's one factor that might cause the ranking algorithm to lower its score.
posted by chrchr at 11:21 AM on June 21, 2014 [1 favorite]
posted by chrchr at 11:21 AM on June 21, 2014 [1 favorite]
And I say that because a lot of machine learning algorithms are based on similarity. It's not that google has a rule that says "old design = spammy". It's that they have a set of pages that are considered good search results and a set that are considered bad search results and the algorithms look for factors that the good sites have in common and the bad sites have in common. Then you show the algorithm a new page that you wish to rank, and it decides if it seems to be more like the good sites or the bad sites. Something along these lines is very likely a component in Google's ranking algorithm.
posted by chrchr at 11:28 AM on June 21, 2014 [2 favorites]
posted by chrchr at 11:28 AM on June 21, 2014 [2 favorites]
Yeah, I actually agree on the Google devaluing "old looking" sites on the basis of design. When I'm randomly searching for stuff, I make hundreds of little snap judgments on whether to trust a site's information or not. Something that looks weird is something I'm probably going to pass over. I totally get it, because it affects my own use of Google and I'm sure they found users might have skipped over MeFi more often than other high quality sites and perhaps the design is one aspect of it.
posted by mathowie (staff) at 3:30 PM on June 21, 2014 [3 favorites]
posted by mathowie (staff) at 3:30 PM on June 21, 2014 [3 favorites]
I hope this doesn't add fuel to the fire of the current design trend of big fonts, page-width images, and low information density. It just seems really shitty for a site to be penalized for not being modern enough when the modern trends aren't necessarily better or indicative of quality. Especially when it comes to judging Google results I tend to reject anything that looks like someone threw it up days ago. Wording purposefully ambiguous.
posted by bleep at 4:16 PM on June 21, 2014 [3 favorites]
posted by bleep at 4:16 PM on June 21, 2014 [3 favorites]
I agree; nothing makes me happier then clicking on a google link and seeing a wall of text. But I'm guessing we're not most people.
posted by Justinian at 4:25 PM on June 21, 2014
posted by Justinian at 4:25 PM on June 21, 2014
I agree that he mentions design so much it must mean something with respect to the problem that happened.
posted by Mid at 4:36 PM on June 21, 2014
posted by Mid at 4:36 PM on June 21, 2014
Of course design matters. Imagine someone clicking on a Mefi/Ask/Meta link in the Google results page, seeing a wall of text, and quickly hitting the back button. Google now knows this is not a great result for that particular person. Now imagine this happening millions of times a day. Whatever the strengths of Mefi, mass appeal isn't it.
posted by dmh at 5:27 PM on June 21, 2014 [1 favorite]
posted by dmh at 5:27 PM on June 21, 2014 [1 favorite]
If that were true then it never would have gained enough popularity to support itself in the first place.
posted by bleep at 5:36 PM on June 21, 2014
posted by bleep at 5:36 PM on June 21, 2014
Yeah, I actually agree on the Google devaluing "old looking" sites on the basis of design.
Be encouraged by the words of Napoleon Bonaparte (with some slight editorial liberties):
posted by SpacemanStix at 6:27 PM on June 21, 2014 [1 favorite]
Be encouraged by the words of Napoleon Bonaparte (with some slight editorial liberties):
Soldiers of my Old Guard:We need people to stand as the Old Guard of the internet.
... For twenty years I have constantly accompanied you on the road to honor and glory. In these latter times, as in the days of our prosperity, you have invariably been models of courage and fidelity. With men such as you our cause could not be lost; but the war would have been interminable. ...
I have sacrificed all of my interests to those of the country. ... Her happiness was my only thought. It will still be the object of my wishes. Do not regret my fate; if I have consented to survive, it is to serve your glory. I intend to write the history of the great achievements we have performed together. ...Would I could press you all to my heart.
posted by SpacemanStix at 6:27 PM on June 21, 2014 [1 favorite]
I know redesign ponies are highly unlikely, but I'd love it if only one thing about the site were changed: much wider body margins. Now that so many people (by which I mean me) are using widescreen views, the long, long lines are a real hindrance to overall readability.
posted by Andrhia at 6:45 PM on June 21, 2014 [3 favorites]
posted by Andrhia at 6:45 PM on June 21, 2014 [3 favorites]
Google Algorithm Change Whacks Naked Capitalism
posted by the man of twists and turns at 7:32 PM on June 21, 2014 [3 favorites]
posted by the man of twists and turns at 7:32 PM on June 21, 2014 [3 favorites]
Maybe old-style design correlates strongly with out of date information.
posted by Obscure Reference at 10:18 AM on June 22, 2014 [2 favorites]
posted by Obscure Reference at 10:18 AM on June 22, 2014 [2 favorites]
Yeah, I actually agree on the Google devaluing "old looking" sites on the basis of design. When I'm randomly searching for stuff, I make hundreds of little snap judgments on whether to trust a site's information or not.
And Google's not 100 percent machines making decisions. One of the things that changed with Panda was increased input from the human reviewers Google's got in a basement somewhere. They're given sample search terms and a manual for rating the usefulness of the results they get, and their input informs the more automated parts of the operation.
I worked for a company that took a hard hit from Panda at the time, and I'd received a copy of that manual not a year or two earlier from a company that seemed to be contracting for the human reviewer bits. I'd built an expansive analytics framework to help us assess the sites that had been hurt the worst, and which content was causing the most damage. That human reviewer guide gave us some corners to tug at.
My favorite response to Panda at the time was a panicky manager's: "Well, if it doesn't like old content, let's get rid of all the bits of the template with the date of the article!"
Because there's nothing our audience of developers and IT professionals would have liked more than not knowing when a piece of technical information on a 15-year-old site had been written.
posted by mph at 10:24 AM on June 22, 2014 [3 favorites]
And Google's not 100 percent machines making decisions. One of the things that changed with Panda was increased input from the human reviewers Google's got in a basement somewhere. They're given sample search terms and a manual for rating the usefulness of the results they get, and their input informs the more automated parts of the operation.
I worked for a company that took a hard hit from Panda at the time, and I'd received a copy of that manual not a year or two earlier from a company that seemed to be contracting for the human reviewer bits. I'd built an expansive analytics framework to help us assess the sites that had been hurt the worst, and which content was causing the most damage. That human reviewer guide gave us some corners to tug at.
My favorite response to Panda at the time was a panicky manager's: "Well, if it doesn't like old content, let's get rid of all the bits of the template with the date of the article!"
Because there's nothing our audience of developers and IT professionals would have liked more than not knowing when a piece of technical information on a 15-year-old site had been written.
posted by mph at 10:24 AM on June 22, 2014 [3 favorites]
Because there's nothing our audience of developers and IT professionals would have liked more than not knowing when a piece of technical information on a 15-year-old site had been written.
This just happens to be one of a few things on the internet that can get under my skin pretty quickly.
posted by SpacemanStix at 10:42 AM on June 22, 2014 [1 favorite]
This just happens to be one of a few things on the internet that can get under my skin pretty quickly.
posted by SpacemanStix at 10:42 AM on June 22, 2014 [1 favorite]
I've been noticing that more and more on annoyingly "modern" sites - no dates on what are plainly date-dependent articles. Argh.
posted by LobsterMitten (staff) at 10:46 AM on June 22, 2014 [10 favorites]
posted by LobsterMitten (staff) at 10:46 AM on June 22, 2014 [10 favorites]
My understanding is that, back in the good ol' days, much of the Google traffic came via Ask MetaFilter. Now, maybe the design is unfamiliar to a lot of web surfers these days just because it's so old. But is there a way that such information could be presented better? I've occasionally run across search results on Stack Exchange or Quora (and ugh, Yahoo Answers), and I personally find them all confusing—or at least, nowhere near as simple as AskMe's. I'm curious to know if others agree or disagree (though obviously this is going to be a biased sample).
posted by Conrad Cornelius o'Donald o'Dell at 11:00 AM on June 22, 2014
posted by Conrad Cornelius o'Donald o'Dell at 11:00 AM on June 22, 2014
Even better: fake dates so the content looks "fresh". Google is definitely valuing newer pages over older ones in general; my ancient old blog has suffered that. May partly explain a drop in AskMe traffic, those 2007 answers aren't going to be as highly ranked solely based on their age. You can get some insight into how Google views the age of a page by using the "Search Tools / Any Time" configuration option to date restrict queries. It's, um, not exact.
I think folks are over-reading Matt Cutts' comments on the importance of visual design. I don't think adding some rounded corners and AJAX content to the pages is going to make a significant difference. But as someone noted above, it's all Kremlinology.
posted by Nelson at 11:01 AM on June 22, 2014
I think folks are over-reading Matt Cutts' comments on the importance of visual design. I don't think adding some rounded corners and AJAX content to the pages is going to make a significant difference. But as someone noted above, it's all Kremlinology.
posted by Nelson at 11:01 AM on June 22, 2014
But is there a way that such information could be presented better?
Yes: with the plain theme. I am not at all being snarky.
I use the plain theme and have for years. If I happen to follow a search link to AskMe on a computer that isn't logged in, I am surprised by the look of the page and I feel like the information is less trustworthy. That's a bias and ideally I wouldn't feel that way, but I'm human. I think many of my fellow humans probably react the same way to AskMe with the default theme.
You could argue that Google should rank sites based on their likelihood to present useful or reliable information for the searcher, but Google's incentive is to rank sites based on how searchers will perceive the information to be useful or reliable, which is not always the same thing. I think AskMe is likely to be an outlier in that regard, with a low perception of reliability and usefulness relative to actual reliability and usefulness.
I have no idea of Google's algorithms capture this distinction in any way, but it certainly wouldn't surprise me.
posted by ssg at 1:30 PM on June 22, 2014 [3 favorites]
Yes: with the plain theme. I am not at all being snarky.
I use the plain theme and have for years. If I happen to follow a search link to AskMe on a computer that isn't logged in, I am surprised by the look of the page and I feel like the information is less trustworthy. That's a bias and ideally I wouldn't feel that way, but I'm human. I think many of my fellow humans probably react the same way to AskMe with the default theme.
You could argue that Google should rank sites based on their likelihood to present useful or reliable information for the searcher, but Google's incentive is to rank sites based on how searchers will perceive the information to be useful or reliable, which is not always the same thing. I think AskMe is likely to be an outlier in that regard, with a low perception of reliability and usefulness relative to actual reliability and usefulness.
I have no idea of Google's algorithms capture this distinction in any way, but it certainly wouldn't surprise me.
posted by ssg at 1:30 PM on June 22, 2014 [3 favorites]
Whatever the strengths of Mefi, mass appeal isn't it.
I have to admit this was the case for me back in '06 when I first discovered the site through AskMe results popping up in my Google queries. I always assumed MetaFilter was some really technical site for, say, users of some particular Linux distro, and AskMe was an "off topic" forum for users in the presumably tight-knit community to ask each other questions unrelated to tech stuff. As I read the site more and more, I was amazed that, hey, there are all sorts of people from different walks of life asking and answering questions on this site, they aren't all Linux sysadmins. It took a few months before it hit me that Metafilter was just a general place for everybody who ponied up $5. It might have taken me a year to even read the Blue or be aware of its existence.
posted by pravit at 3:44 PM on June 22, 2014 [1 favorite]
I have to admit this was the case for me back in '06 when I first discovered the site through AskMe results popping up in my Google queries. I always assumed MetaFilter was some really technical site for, say, users of some particular Linux distro, and AskMe was an "off topic" forum for users in the presumably tight-knit community to ask each other questions unrelated to tech stuff. As I read the site more and more, I was amazed that, hey, there are all sorts of people from different walks of life asking and answering questions on this site, they aren't all Linux sysadmins. It took a few months before it hit me that Metafilter was just a general place for everybody who ponied up $5. It might have taken me a year to even read the Blue or be aware of its existence.
posted by pravit at 3:44 PM on June 22, 2014 [1 favorite]
mathowie: "So we went around slapping a NOINDEX on all the tags pages, archives pages, category pages, and times where you paginate from results like the front page"
That is interesting. If that's actually an issue, then pretty much every Wordpress (which reportedly is serving up around 20% of the web at this point) and, well, every other standard-blog-structure site (including MeFi) is being penalized in rankings by that.
I don't care overmuch about googlenaut traffic as I don't run ads on any of my sites, but I'm going to have to look into noindexing stuff selectively, I guess.
posted by stavrosthewonderchicken at 3:59 PM on June 22, 2014 [1 favorite]
That is interesting. If that's actually an issue, then pretty much every Wordpress (which reportedly is serving up around 20% of the web at this point) and, well, every other standard-blog-structure site (including MeFi) is being penalized in rankings by that.
I don't care overmuch about googlenaut traffic as I don't run ads on any of my sites, but I'm going to have to look into noindexing stuff selectively, I guess.
posted by stavrosthewonderchicken at 3:59 PM on June 22, 2014 [1 favorite]
It might have taken me a year to even read the Blue or be aware of its existence.
Same here. I actually thought for a long time that the Blue was a spin-off of the Green. AskMe results were so common in Google searches that I figured AskMe was the primary function of the site. I simply didn't bother with MeFi for ages - let alone MeTa.
posted by paleyellowwithorange at 6:36 PM on June 22, 2014 [1 favorite]
Same here. I actually thought for a long time that the Blue was a spin-off of the Green. AskMe results were so common in Google searches that I figured AskMe was the primary function of the site. I simply didn't bother with MeFi for ages - let alone MeTa.
posted by paleyellowwithorange at 6:36 PM on June 22, 2014 [1 favorite]
I think folks are over-reading Matt Cutts' comments on the importance of visual design. I don't think adding some rounded corners and AJAX content to the pages is going to make a significant difference. But as someone noted above, it's all Kremlinology.
Exactly. I'd hate for us to make this site as unreadable as the rest of the internet has become because Stalin wore a grey tie or something.
posted by winna at 6:42 PM on June 22, 2014 [5 favorites]
Exactly. I'd hate for us to make this site as unreadable as the rest of the internet has become because Stalin wore a grey tie or something.
posted by winna at 6:42 PM on June 22, 2014 [5 favorites]
I use the plain theme and have for years. If I happen to follow a search link to AskMe on a computer that isn't logged in, I am surprised by the look of the page
Me, too! Even just switching the plain theme to default and the colors as an option for members might work?
I don't find the colors untrustworthy, but, then, I do spend a lot of time googling antiquated unix/linux stuff since the server I'm forced to use at work hasn't been updated in a gazillion years. I'm used to 2001-era web sites with lots of nerd-speak. Even so, that Ask Me green can come on way strong (not bad, just jarring).
posted by bluefly at 8:37 PM on June 22, 2014
Me, too! Even just switching the plain theme to default and the colors as an option for members might work?
I don't find the colors untrustworthy, but, then, I do spend a lot of time googling antiquated unix/linux stuff since the server I'm forced to use at work hasn't been updated in a gazillion years. I'm used to 2001-era web sites with lots of nerd-speak. Even so, that Ask Me green can come on way strong (not bad, just jarring).
posted by bluefly at 8:37 PM on June 22, 2014
I always assumed MetaFilter was some really technical site for, say, users of some particular Linux distro, and AskMe was an "off topic" forum for users in the presumably tight-knit community to ask each other questions unrelated to tech stuff. As I read the site more and more, I was amazed that, hey, there are all sorts of people from different walks of life asking and answering questions on this site, they aren't all Linux sysadmins.
I saw Metafilter and thought, "Hey, this is a great site for linuxy people like me!"
posted by kaibutsu at 8:52 PM on June 22, 2014
I saw Metafilter and thought, "Hey, this is a great site for linuxy people like me!"
posted by kaibutsu at 8:52 PM on June 22, 2014
Westringia F.: "Can someone please explain to me why Google not helping MeFi and/or assisting mathowie directly is being held as an ideal? [...] you're not helping. Why is that, Google?"
Most readers of this site are still operating under the assumption that for Google to change things is as easy as altering some code somewhere. It's not like that anymore - it's all wetware as of about five years ago. Google may be buying a ton of hard drives for storage but when it comes to the nuts and bolts of the search operation it's 100% organic.
When Matt Cutts says things like, "There are a lot of different algorithms that we launch" what he really means is, "Remember those precogs in Minority Report? Basically we just plugged a few more into the big tank of Google Goo."
... and those guys in the tank? they're not really with us, you know? Their fantastically weird little minds are off doing their own thing. You can't just sit on the edge, dangling your feet in the goo, and tap each one on the shoulder and say, "Oh hey bee tee doubleyou that site MetaFilter is good, okay? So don't penalize it maybe?" It's more like winemaking where you have to blend the 'algorithms' in the tank to make sure they're all working to create the output you desire, pull out one of the ones that's too critical, install a more moderate one, et cetera. Plus, it takes time to 'grow' the replacements. It's like that old saying - Rome wasn't bioengineered in a day.
Good news, though - and keep this under your hats - is I heard mathowie got some compensation out of this. Basically Google's going to hand over one of the 'algorithms' that seemed to skew a little liberal but had a high natural language interaction factor. Now I'm not saying anything about anything, but ... I'm just sayin'. When you hear about the introduction of a new 'moderation team' some time in the next few months you keep that in mind is all.
posted by komara at 10:29 PM on June 22, 2014 [3 favorites]
Most readers of this site are still operating under the assumption that for Google to change things is as easy as altering some code somewhere. It's not like that anymore - it's all wetware as of about five years ago. Google may be buying a ton of hard drives for storage but when it comes to the nuts and bolts of the search operation it's 100% organic.
When Matt Cutts says things like, "There are a lot of different algorithms that we launch" what he really means is, "Remember those precogs in Minority Report? Basically we just plugged a few more into the big tank of Google Goo."
... and those guys in the tank? they're not really with us, you know? Their fantastically weird little minds are off doing their own thing. You can't just sit on the edge, dangling your feet in the goo, and tap each one on the shoulder and say, "Oh hey bee tee doubleyou that site MetaFilter is good, okay? So don't penalize it maybe?" It's more like winemaking where you have to blend the 'algorithms' in the tank to make sure they're all working to create the output you desire, pull out one of the ones that's too critical, install a more moderate one, et cetera. Plus, it takes time to 'grow' the replacements. It's like that old saying - Rome wasn't bioengineered in a day.
Good news, though - and keep this under your hats - is I heard mathowie got some compensation out of this. Basically Google's going to hand over one of the 'algorithms' that seemed to skew a little liberal but had a high natural language interaction factor. Now I'm not saying anything about anything, but ... I'm just sayin'. When you hear about the introduction of a new 'moderation team' some time in the next few months you keep that in mind is all.
posted by komara at 10:29 PM on June 22, 2014 [3 favorites]
Bearing in mind that Google said, "Whoops, our bad," and something did go wrong back in 2012, the algorithm (which is approaching AI actually) generally *works*.
This is absolutely not my experience. It's become orders of magnitude harder to find relevant information in places that aren't carefully tended, SEO-optimized, tastefully designed suck-pits of uselessness.
Before 2012, I would get all kinds of forum posts and product pages that actually addressed the topic I was searching on (including Metafilter, the high quality posts got me to plunk down the five bucks.)
Since then, Yahoo Answers and eHow dominates the results. On every topic. The search algorithm can't or won't deep-dive into less corporate user-generated content, like message boards or blogs, the way it once did, and it shoves legitimate web retailers (mostly small shops selling unique stuff, typically craftspeople) right into the bit-bucket. I think this is one reason why Etsy took off in a huge way, Google became useless for finding this kind of stuff.
Worse, when searching for a technical information, it will typically pull online manuals that are years out of date and clearly marked by the manufacturer as Legacy, and prominently displaying a software rev three or four years out of date. Recent docs won't show up until page three or later, after a whole crapload of spam.
So here we are in 2014. I can't trust a "search" company to come up with an algorithm that can tell which technical document to list first by comparing dates and point revs clearly listed in the document. This indicates fundamentally misguided priorities in their dev process - they're ignoring meat and potatoes issues while their fancy-pants secret anti-SEO sauce seems to be nuking actual content providers.
DuckDuckGo is moderately better, still not great, but it's much faster. I'm not at all pleased with the state of search. It may be time to go back to indexing rather than webcrawling - opt-in search using modern automation tools. Once a site is vetted, then crawlers can auto-update, and send up a signal flare if they start doing something shady.
posted by Slap*Happy at 6:47 AM on June 23, 2014 [12 favorites]
This is absolutely not my experience. It's become orders of magnitude harder to find relevant information in places that aren't carefully tended, SEO-optimized, tastefully designed suck-pits of uselessness.
Before 2012, I would get all kinds of forum posts and product pages that actually addressed the topic I was searching on (including Metafilter, the high quality posts got me to plunk down the five bucks.)
Since then, Yahoo Answers and eHow dominates the results. On every topic. The search algorithm can't or won't deep-dive into less corporate user-generated content, like message boards or blogs, the way it once did, and it shoves legitimate web retailers (mostly small shops selling unique stuff, typically craftspeople) right into the bit-bucket. I think this is one reason why Etsy took off in a huge way, Google became useless for finding this kind of stuff.
Worse, when searching for a technical information, it will typically pull online manuals that are years out of date and clearly marked by the manufacturer as Legacy, and prominently displaying a software rev three or four years out of date. Recent docs won't show up until page three or later, after a whole crapload of spam.
So here we are in 2014. I can't trust a "search" company to come up with an algorithm that can tell which technical document to list first by comparing dates and point revs clearly listed in the document. This indicates fundamentally misguided priorities in their dev process - they're ignoring meat and potatoes issues while their fancy-pants secret anti-SEO sauce seems to be nuking actual content providers.
DuckDuckGo is moderately better, still not great, but it's much faster. I'm not at all pleased with the state of search. It may be time to go back to indexing rather than webcrawling - opt-in search using modern automation tools. Once a site is vetted, then crawlers can auto-update, and send up a signal flare if they start doing something shady.
posted by Slap*Happy at 6:47 AM on June 23, 2014 [12 favorites]
Can someone please explain to me why Google not helping MeFi and/or assisting mathowie directly is being held as an ideal?
Because, in positions of influence at Google, there are a bunch of really smart transhumanist whack jobs who honestly believe that an algorithm can be made to exercise human judgment.
posted by flabdablet at 7:09 AM on June 23, 2014 [5 favorites]
Because, in positions of influence at Google, there are a bunch of really smart transhumanist whack jobs who honestly believe that an algorithm can be made to exercise human judgment.
posted by flabdablet at 7:09 AM on June 23, 2014 [5 favorites]
It's become orders of magnitude harder to find relevant information in places that aren't carefully tended, SEO-optimized, tastefully designed suck-pits of uselessness.
Yup. Mutual cargo cultism is the new Quality.
posted by flabdablet at 7:10 AM on June 23, 2014
Yup. Mutual cargo cultism is the new Quality.
posted by flabdablet at 7:10 AM on June 23, 2014
FWIW I just did a google search that returned an AskMe as the fourth hit. I had to do a double-take, not sure what I was seeing.
posted by BeerFilter at 5:57 PM on June 23, 2014
posted by BeerFilter at 5:57 PM on June 23, 2014
I just did a Google search that returned an AskMe as the first hit. I was also a little surprised (but pleasantly so).
posted by triggerfinger at 7:02 PM on June 23, 2014
posted by triggerfinger at 7:02 PM on June 23, 2014
Couldn't agree more w. Slap*Happy.
Can't count the number of times I've had remarkably bad search results from G, thought along the lines of, "How'd they come up with this garbage?"
In general, that and some other things leave me with much a sense that the co. reeks of hubris and arrogance.
posted by ambient2 at 4:47 AM on June 24, 2014
Can't count the number of times I've had remarkably bad search results from G, thought along the lines of, "How'd they come up with this garbage?"
In general, that and some other things leave me with much a sense that the co. reeks of hubris and arrogance.
posted by ambient2 at 4:47 AM on June 24, 2014
And still a simple search for "mathowie" doesn't return a single result from metafilter, and only one if I search for "mathowie metafilter". WTF?
I find this extremely frustrating; I can only imagine the aggravation it's caused the staff.
posted by paulg at 3:27 PM on June 25, 2014
I find this extremely frustrating; I can only imagine the aggravation it's caused the staff.
posted by paulg at 3:27 PM on June 25, 2014
I take it back. Clearly some changes have been made. Results for front page and Ask posts now appear on the first page of Google results. Good.
posted by paulg at 4:07 PM on June 25, 2014
posted by paulg at 4:07 PM on June 25, 2014
And still a simple search for "mathowie" doesn't return a single result from metafilter, and only one if I search for "mathowie metafilter". WTF?
Holy shit, that is surprising. I mean Matt's all over the web and has been forever, but just a simple google search for 'mathowie' not returning anything on Metafilter until the 19th result (for me, and that's the bottom of the second page of results on default settings) is ridiculous.
What's extra weird is that a search for 'stavrosthewonderchicken' gives results #2 and #3 pointing to Metafilter, and though I don't use social media and so on much, I'm kind of all over the web under that name, too.
posted by stavrosthewonderchicken at 4:09 PM on June 25, 2014 [1 favorite]
Holy shit, that is surprising. I mean Matt's all over the web and has been forever, but just a simple google search for 'mathowie' not returning anything on Metafilter until the 19th result (for me, and that's the bottom of the second page of results on default settings) is ridiculous.
What's extra weird is that a search for 'stavrosthewonderchicken' gives results #2 and #3 pointing to Metafilter, and though I don't use social media and so on much, I'm kind of all over the web under that name, too.
posted by stavrosthewonderchicken at 4:09 PM on June 25, 2014 [1 favorite]
I take it back. Clearly some changes have been made. Results for front page and Ask posts now appear on the first page of Google results. Good.
Well, that's odd. Things are going on under the hood, I guess, because I just now checked.
posted by stavrosthewonderchicken at 4:10 PM on June 25, 2014
Well, that's odd. Things are going on under the hood, I guess, because I just now checked.
posted by stavrosthewonderchicken at 4:10 PM on June 25, 2014
Yeah, I just noticed this weekend that a couple of my Google searches returned very useful AskMe results in the top 5 results, including one of those blocks with subresults and a "more results from ask.metafilter.com" link, which I don't think I've ever seen for MeFi before.
posted by Rock Steady at 3:59 AM on June 29, 2014
posted by Rock Steady at 3:59 AM on June 29, 2014
Does this mean I have to remove "Google Delenda Est" from my gmail sigs?
posted by Slap*Happy at 8:31 PM on June 29, 2014 [1 favorite]
posted by Slap*Happy at 8:31 PM on June 29, 2014 [1 favorite]
« Older Fanfare pony request: automatic tagging? | Arrested Development Fanfare Rewatch and a pony... Newer »
You are not logged in, either login or create an account to post comments
posted by mathowie (staff) at 1:36 PM on June 20, 2014 [10 favorites]