Episode 20: Matt Cutts keynote at Pubcon 2013

[Music] welcome to om report by Andre Alpa your interview Focus podcast on topics from online marketing to internet startups all right how's everybody doing this morning yeah all right cool uh I've got some slides but I think Brett said we could talk until like 10:00 a.m. so we'll have time for Q&A too so not too worried about it uh if we can bring up the uh the presentation on the deck so thank you everybody for coming I know like day one is not as

hard to get up at 9:00 a.m. Day 2 is a little harder to get up at 9:00 a.m. when you're in Las Vegas so I appreciate everybody coming here I was uh making my slides and I realized the same thing I realized the very first pubcon that I went to was uh London and Boston in 2003 and it was a little bit smaller then uh great set of folks but uh you know I found a few pictures

uh Brett didn't always wear a suit and Pub didn't always stand for Publishers by the way and uh yeah it's a little weird that's what I looked like 10 years ago uh that was before all the spammers gave me gray hair so all right but you guys want to hear about what's going on with Google what's the future you know what direction is Google moving in and one of the best ways to do that is to look at where

we've gone so I want to talk about the state of the index for a few minutes and talk a little bit about the moves that Google's done so that you can predict where is Google going to go in the future so we've actually doubled down on a lot of what I call moonshot changes so these are changes that you know Larry Page Larry Page says think big like think of the impossible so the knowledge graph that's not just taking Wikipedia

and making it into something pretty it's actually trying to understand the entities the people the stuff that's really in the world so my officemate amet singal calls this thing not strings so instead of just matching key wordss you really know what's behind a query so York New York New York Times Time Square knowing what the difference is between all those and not just matching up the individual key words we've also gotten a lot better at voice

search I thought about trying to do a demo of that today and I was like okay live mic Reverb inside a faraday cage inside basically a bunker uh probably the odd of not good that that'll work well but if you haven't tried it voice search is getting better and better conversational search sounds like voice search but it's actually the ability to do queries over a session so think pronouns I'll show you an example of that in just a second Google Now

sometimes you don't even think to do a query sometimes it would be really nice if Google would tell you hey you've got a meeting at 10:00 a.m. it's over at the bagio the traffic's pretty bad you'd better leave now if you want to make it to your meeting so it's it's looking for that space where people aren't even thinking to necessarily do those queries yet and one of the biggest changes we've been starting to do is something called

Deep learning it's basically like these neural networks that people did in the 1980s except it's multiple layers of neural networks and it's using thousands of computers to try to improve what you can learn and the stuff that it can learn is pretty crazy amazing so let me show you just one example if you take a word and you boil it down into a thousand dimensional space so you make it a vector where it's a series of a thousand numbers and you

make every word into this thousand dimensional Vector space the relationships between those words actually tell you something about the semantic meaning of those words so for example given China and Beijing you can say okay this word China is here in the space this word here Beijing is here in the space and the difference between those actually encodes something meaningful it kind of encodes what is the capital of and so if you take another uh country like Russia or turkey

and you apply that same Delta that same Vector you get Moscow or you get anara it doesn't just work for like capitals of countries it works for all kinds of things this is a paper that Jeff Dean who's one of our uh senior fellows in fact he used to be a Google fellow which was level 10 on the career ladder and then they made him a senior fellow so his level actually goes up to 11 like these guys have worked on deep

learning where they try to figure out what are the relationships between words so the first row of this is just like France is to Paris as Italy is to Rome or Japan is to Tokyo but it starts to get smarter you take a word like big and you compare where it goes to bigger and then you apply that same Vector to a word like small and you get a word like larger and so we're actually trying to

teach Google to learn and to read at you know some sort of Elementary School level or grade school Google's 15 years old it ought to be able to do some of these standardized tests and it gets to some pretty crazy stuff so like you can see this road that has copper the atomic symbol for copper is cu and then it can interpolate or extrapolate okay gold is au and zinc is ZN and it can kind of

learn some crazy things like Einstein is to scientist as Mozart is to a violinist so over time we're getting better and better at doing that sort of deep learning we're also working on this voice search so you can say who is the Prime Minister of Turkey thanks to the knowledge graph we actually know Turkey is a proper noun It's actually an entity in the world it should be capitalized and so you can say t er erdogan is the

Prime Minister of Turkey but then you can keep going you can say how old is he and we know that you were talking about the Prime Minister of Turkey from the previous query and we'll say okay he's 59 years old so Google is trying to figure out answers we're trying to organize the world's information make it universally accessible and useful and actually figure out what is the person exactly asking for here's a query I was doing last night I asked will it rain

tomorrow and you can tap on the voice search you can say that and Google will actually reply back to you don't try it right now but when you get home just go to Chrome and you'll notice if you go to google.com there's actually a microphone there and you can click and you can talk directly to your computer and it will talk back so it knows will it rain tomorrow it knows my location is in Las Vegas or Paradise Nevada and it gives me

an answer but then you can tie that together with conversational search and you can do a query like what about Mountain View all you have to do is say what about Mountain View for the next query and it knows you're still looking for the weather and it says oh no it's not going to rain in Mountain View either tomorrow and then you can keep going and you can say how about this weekend I encourage you to try these

queries out and it we'll say oh you're talking about how about well it will rain in Mountain View this weekend and it says no on Saturday in Mountain View it's not going to rain so we're starting to figure out the structure of what people are actually saying we're starting to get a little bit better about that and understanding what people are asking for so those are the sort of moonshot changes that we've G got going on I want to start big and then we'll

drill down to Quality then we'll drill down to web spam then we'll talk about the future go the same way so one of the big changes that we rolled out in the last few months is a change called hummingbird the idea behind hummingbird is if you're doing a query it might be a natural language query and you might include some word that you don't necessarily need like uh what's the capital of Texas my dear well my dear

doesn't really add anything to that query it would be totally fine if you said just what is the capital of Texas or what is the capital of ever Loven Texas or what is the capital of crazy Rebel beautiful Texas some of those words don't matter as much and previously Google used to match just the words in the query now we're starting to say which ones are actually more helpful and which ones are more important and so

hummingbird is a step in that direction where if you are um you know saying or typing a longer query then we're going to figure out which words matter more and give that more intelligent scoring now there's a lot of Articles written about hummingbird like even when just the code name was known people were like okay how will hummingbird affect SEO and even though people don't know exactly what hummingbird is they're still going to write 500 words about how hummingbird

affects SEO and the fact is it doesn't affect it that much it affected 90% of queries but only to a small degree and we rolled it out for over a month without people even noticing so it's a subel change it's not something that you need to worry about it's not going to rock your world like Panda or penguin it's just going to make the results a little bit better and especially on those longtail queries or really

specific queries make them much better so unless you're you know a spammer and you're targeting you know how many seos does it take to change a light bulb and you've got like all the keywords and you've got 15 variants of it you got a page for each one you know if you're doing those really long tail things then it might affect you but in general people don't need to worry that much about humming bird uh another change that we did is we

were looking at softening Panda so there's always people who we think should be affected by panda and there's sites that we think are really high quality and shouldn't be affected by Panda and then you have a gray Zone and with that Gray Zone you basically basically have to guess whether a site is high quality or not and so we found some new signals that basically help us disambiguate that a little bit and move some of the sites that were in the gray

Zone toward the higher quality area so softening the effect of panda in some instances we've also been looking at detecting the boosting Authority so take uh medical for example if you're an authority in the medical site in the medical space we want to be able to know that and start to push you up a little bit higher whenever a medical query comes along now this is not something that's done by hand it's not like we pick the individual topic areas it

actually applies to thousands of different topic areas so nothing that you have to do but if you are a topical Authority keep writing about it keep developing keep deepening the amount of content that you have you really want to be a resource you do want to be an authority and if you turn out to be an authority then you're more likely to be boosted by that particular change and we've also been working on smartphone ranking so if you have a a

phone that doesn't do flash then we're less likely to show you a page that contains Flash for example uh over time we might start to think about whether a uh a site is slower on smartphones uh We've also done ranking changes that say if every single uh page on a smartphone redirects to the root page and there's not you know you don't get the individual Pages then we might start to rank that lower as well okay so now let's drill down a

little bit deeper let's talk about web spam changes in the last few months penguin 2.0 and 2.1 launched um it was kind of funny because we were working on the next generation of penguin which we call Penguin 2.0 and uh we were trying to get a really soft Landing we wanted something that was you know wouldn't set everybody's hair on fire running around screaming and they were crazy like oh no this is horrible ah and so penguin 2.0

launched and the spammers were actually like you call that a Spam change that didn't affect me at all and so a lot of people were a lot of the black cat spam forums were like oh that didn't have any impact and so we're like okay well we can turn that knob a little bit higher and so that's what we did with penguin 2.1 and uh and we're going to keep iterating on those those kinds of uh

methods of detecting spam so that people don't have the incentive just to just create nasty ugly stuff that doesn't help anybody the last one say four to 5 months we've also been working on very spammy areas so think like payday loans in the UK where you can actually like see the Fingerprints of the Russian mafia being involved this sort of thing so so uh we've come up with a couple different algorithms uh they don't just apply forer like payday loans they'll

they can apply for like mesothelioma and a whole bunch of like spammy areas car insurance you know stuff where there's a lot of tricks that go on pornographic queries those sorts of things uh we're going to keep iterating on that we're going to keep trying to recoup that we've also been taking action on advertorials or native advertising now just to be clear there's nothing wrong with advertorials or native advertising as long as you mark it clearly as long

as people know that it's an advertisement but we've taken action on several dozen newspapers in the US and in the UK when they had paid content that wasn't labeled as paid in any way and was flowing page rank it's pretty common sense we've said it since 2006 like and probably before that you shouldn't be paying for links that past PID rank that's a high-risk area and that applies to advertorials and Native advertising just like it applies to

everything else uh and then we continue to to take action on spam n works so uh at one point we were like okay maybe we should uh you know take a poll about which spam link Network should we take down next but instead we've got a pretty good list and we're just working our way down it um I can actually see that we're having a pretty good impact because if you go to random black hat forums you

get comments like this who wants to punch mat Cuts in the face so anytime I'm being threatened with bodily violence on a black hat Forum I know my team is doing their job uh note that this is a this user is Jason a this is not Jason if you you guys want to hear about Jason you have to you know take some time out of the Q&A and ask about it specifically I'm not going to get into that unless people

really want to hear about it so if the spammers are unhappy on the black hat forums then some you know not always but often that's a good signal we've also been working on communication I think a lot of people know these things so I'll cover them relatively quickly uh Miley oay made a ton of like hours of video in case you've been hacked or to help you with malware there's a lot of sites that have problems with that and so we're trying

to figure out how we can help people preventively not get hacked or not get malware and then give them a warning and more heads up whenever they do get hacked uh we we increased the amount of examples that we included in our quality guidelines because a lot of people were like okay I understand you know maybe you don't want me to pay for links but are there other examples of spammy links that you could give me and then the team

there's like you know a dozen people a lot of people like want to hear about okay what does Matt cut say but there's actually like a dozen people at Google who do various types of web Master communication right John Mueller Pierre far Miley o zanev Gary there's there's whiz there's Maria there's a ton of people who go to all kinds of places you know they go to Russian Church conferences that's Vladmir or Maria they go to all sorts of different places uh

John Mueller does Web Master office hours a bunch of different people talk about different topics so if you want to talk to someone at Google we try to make that easy and uh and I think if you want to you know dial into a Web Master office hours that's fantastic we also launched a website this summer called or earlier this year called House search works and I I want to ask I want you to raise your hand if you've been to the

house Search Works website okay so that's that's a fair number of people uh maybe 30% maybe 25% but most people in this room have not been to that website if you're willing to pay hundreds of dollars to come to a search conference I highly encourage you to come check this out because we actually tell you the categories of spam that we take manual action on so you can actually see these are the things the categories that we

consider spam and then when we break down what are the categories that we take action on the most so black hat pure spam gibberish autogenerated Just Junk the sort of stuff that you look at it and you immediately know it's spam that's the number one category but then you can look down and you can see oh I see orange that orange is the next one what's that oh that's hacked it turns out hacked content is the next most

common category so a lot of the stuff that that you're wondering where's the webpam team spending its time you can find those kinds of answers on this website I highly encourage you to check it out you can even see live screenshots of spam that we're taking out as we take it out it's like you're watching over our shoulder as you as we're fighting spam and then we'll even show data like the number of requests reconsideration requests that we get each week and so

you can find out a lot more information about the volume of data that we process and all that sort of stuff okay so all of that serves to give you an idea of where Google has gone in the last year or so and from that you can sort of try to extrapolate where Google's going to go but why don't I just spell it out for you the sort of trends that I think are really going to

matter so let's start high level again let's start really big what are the mega Trends the big future Trends one of them is machine learning Google is going to continue to try to be smarter and smarter so our mission statement is to organize the world's information and make it universally accessible and useful that doesn't include the word search engine in there if we could find ways to solve equations and tell you whether it's going to rain tomorrow and

and extrapolate all kinds of useful information we're going to try to do we're going to keep trying to figure out how to add more value for users and for Searchers mobile is huge what do you think these numbers represent 6% 25% 40% anybody want to guess in 2011 YouTube had 6% of its traffic coming from mobile phones in 2012 YouTube had 25% of its traffic come from mobile phones in 2013 40% of YouTube's traffic comes from

mobile phones mobile is coming faster than anyone expects it there's a ton of Savvy people in this room and yet no matter how Savvy you are I think you might be surprised at how quickly mobile is growing in some countries mobile traffic has already surpassed desktop traffic in a ton of other countries it will surpass desktop traffic in the next one or two years so if you haven't thought about mobile if you haven't figured out what your strategy is if

your website looks really sucky on mobile you want to start thinking about that now another big trend is social identity authorship I think Facebook did a fantastic job of recognizing the value of Social and the value of knowing you know who people are on the web and if web spam knew who people were on the web it would be much easier to keep the spammers out of search results so things like authorship things like knowing who

you are and identity can make a big difference now a lot of people are like okay if I go get in a lot of retweets on Twitter or a lot of likes on Facebook or a lot of plus ones on go+ does that mean my ranking will go higher tell me tell me now is that the signal is that what I should be chasing and my answer is not in the short term right it's not the

case that we're able to crawl all of Facebook they block a lot of different pages it's not the case that we're necessarily able to access every page on Twitter and it's not the case that plus ones give you a boost in Google's ranking right now however in the long term having good social signals is a reflection of being an authority it's a reflection of being the sort of person that people listen to and the to the

degree that social reflects the fact that it mirrors the fact that you are someone worth listening to then search engines want to listen to you as well so don't get it backwards don't say I have to boost up my social so I'll rank higher think I want to be an authority I want to be someone that people listen to I want to be an expert so that all those signals that accompany that tell the search engines that hey this is someone

who should be ranking and ranking well okay let's drill down a little bit deeper web spam Trends it's going to look for the next six months like web spam's not doing much I'm just going to tell you that right now we're going to be working on things that most people won't see so hacking remains one of the big areas that we haven't yet tackled we've well we've tackled but we're working on the next generation of hacking detection uh

so if you do a query like by Viagra it's still spamming I'll be happy to admit that it's still spammy because there are people willing to do illegal things and hack websites we're working on that we're trying to figure out the Next Generation so that we can catch that and make sure that those guys the people who are willing to do those illegal things not just black hat but like would go to prison black cat are not going to

succeed it will take a little bit of time but we're going to keep working on it we've also been working on um some some really hot topics internationally things like child sexual abuse imagery because I started out on webpam because I was working on safe search I still get pulled in every so often whenever there's something that we need to work on improving and we want to make sure that if you type really nasty queries you do not find what you're looking for

on Google so uh we don't want people to be able to find you know child forign even if they're searching for it on Google um and then just one tidbit a lot of people ask when are we going to get the next page rank update I'll just give you the skinny right now we have our own internal version of page rank it's always updated it's continuous and continual every single day we have new page ranks then there's also an export

that says okay given our internal page ranks export that to the Google tool bar and normally it runs once every 3 months or so maybe every 3 or 4 months a while ago like earlier this year that pipeline broke and we were kind of like you know people get a little too obsessed about page rank anyway maybe it's okay to just leave that for a little while and so uh we we don't have anybody staffed on

trying to revive that Pipeline and we don't want people to get too obsessed about page Rank and so we're probably not going to update page rank throughout the rest of the year and then we'll see whether anything happens in 2014 okay so let me give you some very concrete advice getting down to the the end of the presentation and looking forward to Q&A so what are some things that I would look forward to in the next

year and some things I would do on my website um I mentioned mobile right you need to get ready if you're not thinking about whether you're going to go with M do whatever or responsive web design or you know however you're going to handle it start thinking about it if you're still using a site that has flash you are not able to be seen on a large variety of mobile phones so you need to start thinking about how you're going to

solve that sort of issue this is tactical advice but there's a really cool standard coming out it's I think maybe already in the beta version of Chrome called request autocomplete how many people like filling out forms and typing in your credit card and all that stuff when you buy stuff online no one no one I would expect at least least one weird person okay one one like yeah I really like typing in forms they're awesome I type in forms

all the day people hate it if you have a form on your website people do not like that part of the experience if you've got a funnel where you want to start with awareness and drive it down to engagement and convert that to someone who's a pain user or someone who subscribes to your newsletter or whatever your conversion goal is having that form makes things more difficult so request autocomplete is a standard whereby you can basically mark up your

forms so the the easiest thing to do is just make sure you annotate your forms with some standardized markup you can do it literally in an afternoon I'm talking like 3 or 4 hours it's just the HTML you just add an extra attribute to your form that says this is an address or this is the city it's not hard to do that's the bare minimum but request autocomplete will land probably in the next few months in Chrome in the main Channel and

basically the sites that support that can say you know what do you want to one one click fill this form out and people are like yes I would like to fill out this form in one click so if you look at all the friction all the people who abandon the shopping cart all the people who leave where you don't get the traction autocomplete is a really good way to try to reduce that and minimize friction okay we're working on a new

version of uh you know things like add heavy pages so we rolled out a version I actually sort of talked about it first at pubcon uh a while ago and we're turning the crank to do another version um it will actually have more of an impact in some other languages like Russian or Arabic than it will have in English but if you look at the top part of your page and the very first thing you see front and center top above the

fold is ADS right there then you might want to ask yourself do I have the best user experience because we are working on an algorithm and the next iteration of that algorithm to try to take some action on that um authorship we want to make sure that the people who we show as authors are high quality authors and so we're looking at the process of possibly tightening that up it turns out if we reduce the amount of authorship that we

show by just like about 10 or 15% we're radically able to improve the quality of the authors that we show which is another nice signal for the Searchers and the users who are typing into Google to say oh I see this picture I see this person is a is a is an author this is something that I can trust this is content that I really want to see so it's not just going to be about markup

it's going to be about the quality of the authorship and then in the same way we're thinking about whether Rich Snippets it might make sense to take into account the quality of the site Rich Snippets when it started out we actually had to have the rich Snippets team say yes you are approved to have Rich Snippets and then it mov to everybody's approved and then we'll take out a little bit of spam peace meal if we find it and we're starting to find a

better middle ground and that middle ground is if you're a reputable site then we'll trust Rich snippets on your site if you look like you might be a Spam or a lowquality site then you're less likely to have Rich Snippets so that's some advice to stuff to get ready for and things to be thinking about in the next uh few months one last thing is we have been working on JavaScript not just uh you know Ajax sort of stuff but actually the

ability to fetch execute render and index things created by JavaScript so you want to be a little bit cautious about this don't just take a flying leap and make your entire site one big hairball of JavaScript but Google is starting to get smarter and if you want to use common JavaScript Frameworks very common libraries that a lot of sites use it's more likely that you'll be able to do that and still be indexed well so we

are doing some limited testing with a small number of sites things look pretty good and we're going to keep iterating on that to try to make sure that we do better JavaScript indexing uh a few things worth doing now sign up for Web Master tools if you haven't there's a Twitter account Google webm Central uh and so if you ever watch my videos on uh on YouTube Google WMC usually tweets it out before I do so if

you want to get in early and find out what what's the newest video what are people going to talk about uh today that's a great way to do it and then our Web Master blog inide search web Master Video channel the main thing I want to say is it's been a fantastic experience coming to pubcon for the last 10 plus years met a lot of people that I've really enjoyed talking to uh getting to be friends with some of them I really

appreciate the feedback that we get you know whenever there's bugs whenever there's stuff that we can be doing better um that makes a huge difference because I can come back and I can say look the outside world is complaining about this here's something that we need to do better and so I really appreciate all the people who give us that kind of feedback and with that I think we'll open it up whoever wants to ask [Applause]

questions testing one two testing one two Mr Angy thanks Brett that was good and Loud um I have one quick question and and then one topic area the topic area I'd like to address have you address I perhaps you could do that second which is negative seal M uh but before that whenever you talk about Google Plus signals you always specifically say plus ones and I want to see if you can differentiate between a plus one that

does not come with a sharing action a comment versus a share where you're actually sharing something or even potentially sharing link cuz I think it makes sense that those are a little bit different scenario because when you do a share there's a little bit more of the person doing the share invested in the event yep absolutely so when I I talk about plus ones a lot but mainly that's just because it's emblematic of you know the sort of signals that you get from

go+ so it it's not the case that a plus one or or a share or a comment or something like that makes a makes a difference in ranking today now that's in generalized ranking I will say for personalized search if you're friends with someone and they post something or or you know they've claimed authorship of it that sort of thing then that can you know affect whether a picture shows up and maybe you'll click on it more

often because you know that person but I'm talking about the generalized web ranking um so the high order bit is yeah I talk about plus ones specifically because it's easier to just talk about plus ones but the the broader picture is the same which is it's still short-term for social uh we're still investigating those signals long term I'm very bullish on the idea of Social and identity okay so let's talk about negative SEO great question um we have thought about

negative SEO for years and years and years and years we have to design our algorithm such that you know one person can't torpedo another person and we think hard about that now um with penguin the sorts of things that you're doing the sort of you know if there's black hat spam going on it doesn't just remove you know the Boost that you might get it can actually have a slightly negative effect and because of that we've tried to be

very very careful and we haven't gone as strong in some aspects of penguin in order to try to make sure that we minimize the potential for negative SEO I think negative SEO is something that a lot of people worry about but is actually relatively rare so I I got a complaint from a a site that sells wigs and they said you know what my other wig competitor is doing negative SEO on me and it's not fair my rankings have

dropped and you guys should do better at designing algorithms and I sent him back a link that had been live since 2010 I think and it was a paid link it was abundantly clear it was a paid link it was directly to his site keyword Rich anchor text and it had been there for 2 or 3 years I was like you really think that your competitor started doing negative SEO in 2010 before Panda even launched right

and so there's a lot of people who are worried about it and who are thinking a I've got this this compet compor who's doing something to me when we're designing the algorithms to try to make it as difficult as possible for that to happen and a lot of people don't know what in-house or outsourced SEO might have you know made a lot of a mess now one of the things that we did is last year we introduced the disavow tool in

fact I think I announced it a pubcon last year and so if you have been cleaning up you know the mess that your in-house or outsourced SEO did and you've gotten as many links as you can down and we've seen reconsideration requests that include postcards where people have sent a stack of physical letters to people said please take these down once you have exhausted those kinds of options that's the perfect thing for disavowing so the disavow tool is

intended for that sort of cleanup and to make cleanup easier but if you are worried about negative SEO if you do think that someone has really got it in for you then you can use the disavow tool you can go to web Master tools you can sort links by the most recent so you can see the links that we're seeing you can see how we're garnering them and then you can actually you know go through and if you see a new link that

you really think some competitor is doing you can disavow it and you can disavow at the Domain level but in general um it's not as big a phenomenon as most people are are actually worried about so that's that's just a little bit about it the the other thing I'll mention is on Web Master tools we've started to give better backlinks more diverse backlinks we used to give you basically sorted by the alphabet and so if you were say Amazon you get like

100,000 backlinks and it would stop at like AA aaa.com or something like that now we actually give a much better random sample of those backlinks they're still sorted alphabetically but after we've sampled from the entire set that we have in our base index and so if you are worried about negative SEO or if you think that you've had a SEO who did bad work for you you can get a much better picture of what's actually been going on

so that launched just the last few months so if you haven't gone to download your links 100,000 links from Web Master tools I would recommend that because you'll get a much better picture of what's actually going on Matt one one one two quick announcements while you're looking for somebody to ask a question from there Jo I have one oh you have one well that's where I was going too uh two quick things number one those going to pubcon

Labs that'll be in the middle of exhibit hall that was not announced earlier it should have been uh in the middle of the exhibit hall number two before I forget I want to personally thank Mr Jim boyin and internet marketing Ninjas for sponsoring the conference this year they underwrote everything this year thank you Jim let's give them a round of applause and and last thing for those watching at home uh hi Jason I'm I'm sure you're

watching so so did you catch yesterday's keynote by chance Matt uh so I was flying in right while Jason's keynote was happening and uh I I gather it was kind of interesting yeah somebody called somebody else a liar something like that I don't know evil I don't know not a bad not a good partner uh I I thought about I actually made some slides go for it well no let me let me put it to the

crowd no no so I asked Jason on Twitter whether he wanted the polite response or the thorough response you're not going to get the thorough response at most I will give you the polite response I I did make some slides but I'm happy to do Q&A you know everybody wants the bloodthirst everybody wants the SEO cage match and smack down I'm not going to I'm not going to smack down Jason I'm just going to give a careful view of of my

considered opinion of how things went down but that's like 5 10 you know 7 minutes away from regular questions we'll wait we're not going anywhere you guys actually want to hear like our let let the record show people ask for it okay so the appendix all right it's all a set up this is not a wrestling match you guys shouldn't want us to fight like here's the thing I have huge I'm I'm going to look straight into

the camera in case he's watching I have huge respect for Jason kakas he's a smart guy he's a passionate guy he calls it the way he sees it and he was trying to do the best he could with his business and his business is like his baby right he's taking money he's got employees he's working as best he can to try to make sure that Mahalo succeeded right and I can understand why he would be frustrated so I I don't begrudge Jason

anything at all but I'll I'll try to give a little bit of context here so on Twitter yesterday somebody asked him Jason why would mat Cuts or Google apologize when Panda made the results significantly better and I thought the response from Jason was pretty interesting he said a lot of babies got thrown out with the bath water which I would actually agree the first launch of panda was pretty uh substantial right it it was a jolt um he said they could have

rolled out slowly I actually disagree with that they could have shown changes I think he mean shown changes to Partners that sort of thing so let me just take this idea that we could have rolled out slowly and let me go backwards in time a little bit I'm not sure if everybody remembers what it was like in 2011 we actually started worrying about lowquality content and content farms in 2010 so we were working on this problem for eight or nine months

before it really broke big but then on January 1st there was an article by Vivic wadwa and TechCrunch that said why we desperately need a new and better Google uh Jeff Atwood from Coden said Google The Once essential tool is somehow losing its Edge the spamers scrapers and SEO to the hillt content Farms are winning Christine haverson said I believe if Google does not get on top of the content Farm situation it's going to be their downfall and I mean that

sincerely Marco amen a lot of people have heard of said massive amounts of technically not spam sites are generated by Penny hungry affiliate marketers and sleazy web content startups to Target longtail Google queries on maass so this is what it was like in 2011 we've been working on the problem for a long time and suddenly everybody decided to be really angry at once it got so bad here's read white web somebody said Can quality survive they

said Google needs to wake up and smell the coffee I have a half dozen articles like this they're really painful to read when you already know there's a problem when you're already working as fast as you can on that problem as a stop Gap measure someone on my team made a Chrome extension that would let you block sites I'm not sure if people can see the red box 197,000 people have installed that Chrome extension so there were a lot of

people who were not happy with the quality of Google in early 2011 who did not like the content Farms it was our internal metrics our internal feelings our internal data was saying that the outside world was saying that and then 200,000 people installing this Chrome extension all shows that people were not happy about condent Farms they did not want content from content farms in January February of 2011 now where does mahalo come into that picture well here's a here's an

article by Business Insider back then you know people have made fun of Business Insider saying oh sometimes they take content from other guys and they summarize it and they're you know they don't always have the deepest most introspective reporting and Business Insider calling out Mahalo right they called it hilariously useless uh the paragraph at the bottom the bad news for Jason kakas is that the Mahalo page was laughable and then they said seriously like not joking they did not like the

content the quality of Mahalo so it it got so bad there were actually people making jokes about content Farms there were parody sites how to stub your toe how to stub your toes so there's a site called the content farm. tumblr.com by the way the instructions on how to stub your toe choose a place to walk pick one of four cardinal directions walk continue in this fashion until you stub your toe right no joke there were people making

fun of content farm so much there was actually an article about how to pour milk okay so let me just you know here we are in the time machine remembering what was going on grasp the handle of the container in which the milk is with the hand lift it in an upwards direction towards the ceiling by moving your arm while while still holding the container in the handle with your hand position the opening or orifice of the container

over or on top of some type of receptacle such as a glass cup mug Bowl teacup small pitchure measuring cup or saucer right people have seen that kind of keyword stuffing going on so Google did have to move fast as soon as we had the signals ready we wanted to launch them we knew that we could iterate we knew that we could improve over time and in fact we launched uh this personal block list pretty much to just tie people over for

a month or two because we knew we had the signals coming and this was something where we could sort of have a valve where people could get a little bit of Rage out of their system now the other uh the other issue that I think Jason had is I think and I didn't see the keynote and I haven't watched on Ustream so you know correct me if I get it wrong but he seemed to be implying

that Google was not a good partner um that Google wasn't doing as good as it could as far as being a partner now that's kind of interesting because that's not how we think about it in search quality Mahalo was a partner with YouTube but if you say look if you're a special partner with YouTube should you get to you know have a one-on-one meeting with Matt at the Google Plex I think most people would say no that doesn't seem fair it doesn't

seem right if you're a partner you know you know with YouTube or some other property that you automatically get to talk to the the search quality team so that's kind of tough I mean demand media was a partner on YouTube did that mean that demand media should get special treatment that they should get a special meeting I don't think that that's the case now Jason can actually be pretty insistent so he wanted to have a meeting

with me and talk about things um has anybody else had a one-on-one meeting with me at the Google Plex to talk about panda don't raise your hands cuz I know the answer is zero in this room okay but I actually agreed to meet with Jason he can be pretty assistant he really wanted to make his case that Mahalo was high quality the problem is I was the messenger Jason wanted to convince me that there was high quality content on

Mahalo and I believe that there was some high quality content but in aggregate the signals that we had the data in the algorithm said that it was not in aggregate high quality site the other issue is it's really tough to go into a meeting because you don't want to you don't want to be the guy to say hey I know you love your business I know you're passionate about your business I know you're working hard and you're trying to adjust and you're

trying to improve the quality of your business but your site is actually one of the most blocked sites in this chrome exension like that's really hard news to break to somebody and when you're when the other person really wants to convince you that it's high quality content I think ultimately we came to an impass at that meeting and you know for good reasons he wanted to convince me it was high quality content I was trying to

break the news that even if he convinced me it wouldn't make any difference it's sort of like getting mad at a traffic cop whenever the government is shut down it wasn't P Panda wasn't in the web spam team it was actually in the larger quality team the initial launch didn't come from web spam at all so I was happy to try to explain all of this to Jason But ultimately he was not happy with the outcome and I could understand why he's

frustrated so content Farms were a serious issue we couldn't roll out changes slowly and we really do not want to give Partners special treatment in fact after that meeting with Jason I pretty much you know decided I don't ever want to meet and I didn't talk to Jason because he was a partner I talked to him because he had a high impact site you know in some ways he'd said he was a search engine and so I wanted to bend

over backwards to try to give him information and give him context but we've actually got a pretty good policy now I have an internal website you know has anybody watched a YouTube video that I've made a few people and everybody's like I don't want to raise my hand Leave Me Alone we've also made a couple internal videos so I've made a 5 minute internal video when basically a salesperson comes to me and they say hey this person

spends x millionar a year while you talk to them or hey this is a really valued partner or this is a great client I send them to this little mini website that we have inside Google that basically basically explains why we won't talk to them because we don't want to give special treatment to Partners advertisers or clients and that's basically just the way it works so 6 months after Panda launched and in roughly 6 months in June there was a

quote in the New York Times you can't mess with Google forever uh just a short time ago the web seemed ungovernable bad content was driving out good but Google asserted itself and credit is due Panda represents good cyber governance it has allowed Google to send untrustworthy repetitive and unsatisfying content to the back of the class so that's the kind of verdict that we have on Panda we did have to launch it quickly there was some collateral

damage we had to figure out how to iterate and improve But ultimately there was a fundamental disagreement between Jason and I about the the of a particular site and more importantly there was a disagreement between Jason and the algorithm and we we have no exceptions to Panda we have no exceptions whatsoever so even if he had convinced me that that Mahalo was the best site in the world I at best I could have gone to the engineers and said can

you try to find some signals that show that this is a good site we did that analysis we didn't find the data to support that so I understand if Jason's angry I understand if he's frustrated he had a lot of people depending on him he was under a lot of pressure I guess get a lot of complaints I get a lot of anger I get a lot of frustration and that's okay right it's all right if a black hat

spammer wants to say I want to punch Matt Cuts in the face that's all right what I have to do is look at what the other person's saying and look and see is there a kernel of Truth in there is there something that we need to do better and I think there were things that we needed to do better we've tried to improve our communication we tried to make things more scalable we're sending out example URLs and messages so that

people know what's exactly wrong if we've taken manual action we've launched the manual action viewer so you can actually see whether Google has put a manual webpam action on your site or if it's only algorithmic ranking so there were a lot of things that we needed to improve and I I appreciated Jason's feedback if he's still angry at me then that's his call but I I think in every one of these encounters whether it's you know Michael Gray calling us out Aaron

wall calling us out Jason kakas calling us out in every one of those instances that's an opportunity for us to say what's the kernel of truth behind what they're saying and what can we do better and I think that we're doing better now than we did in 2011 and we'll do better in the future future but we have to keep working on that so if you guys wanted a cage match I'm sorry but you know there's a couple well-meaning people who

both are trying to do what they think is right that's okay all right so back to the pubcon type questions uh I spoke to a bunch of speakers they kind of gave me a few questions they want to run by we'll start with the first one we'll move to the audience in July Google updated the links sches and there was a lot of changes a lot of tactics that were used by many people in this room are no

longer valid do we have to worry about old content that may violate some of that yeah so I think this question especially seems to apply to press releases you know I was getting a lot of links from press releases and then Google specifically called out press releases so do I need to worry about that do I need to go back in time it's kind of weird cuz people act like are guidelines change a lot but the

fundamental ideas behind our guidelines haven't changed that much so people for the most part should know if you're paying for something then you shouldn't get page rank out of it that's that's basically you know money under the table it's it's poola we don't want that sort of stuff to count so if you were doing a press release and there's no keyword text hey great you're just trying to convince a bunch of editors to write those articles if you're doing a press

release and you're like embedding and I've you know you've certainly seen plenty of those kinds of press releases where they're throwing three or four keyword Rich articles or anchor text in there then that is something to pay attention to It's not that it's new we've always had the principle if you're paying a 100 bucks or whatever for a press release then that shouldn't count that shouldn't flow page rank that shouldn't flow anchor text we're just clarifying our existing principles in

that regard now with especially with regard to press releases what we've done is we've identified a lot of the top press release sites and all we have to do is we say okay you know what maybe we don't trust links or anchors coming out of those sites we don't count them as spam we just say just ignore those links and that's okay because the idea behind a press release is to convince someone to write about it if you can get an

editor or reporter to take that seriously that's fantastic that's great that newspaper article hopefully it'll include a link but at the least it increases awareness but if you're worrying about press releases for the most part we've already taken care of those if you thought you were paying $100 and you were getting keyword anchor text what you were really getting is an article you know a press release and then you weren't necessar you weren't you probably weren't getting that page

rank you weren't getting that anchor text so it's not that would necessarily C cause you a penalty I wouldn't panic but going forward I would say if someone approaches you and they're like hey you know do you want to be in my directory well it's 100 bucks but I'll guarantee you whatever listing you want yes you should be cautious about that you should not do that hopefully that's relatively baked in as far as common sense by now

but hi um a lot of Y uh more and more sites are using infinite scrolling can you address that absolutely so infinite scrolling is getting more common and that's the sort of thing like like JavaScript um it is it is tricky because it's almost like a calendar where you can say okay give me all the dates in 2012 and then 2013 and then 2014 and in theory Google B could keep crawling all the way up to the year 10,000 you know and we

never know when to stop so you might want to think about building in some stuff you might not want to have infinite scroll you might want to have the scroll finish at some point but Google is getting better at being able to access those links and say okay do those Scrolls scroll events that sort of thing I wouldn't necessarily count on Google being able to like if you're going to move the mouse or something like that and you have to wait until the

scroll bar gets to the bottom of the page that's the sort of thing where Google bot might not be literally physically scrolling so you probably do want to think about having some paginated kind of interface where people can still scroll through if you can we'll keep working on that because we want the web to be beautiful and modern and and you know seamless experience um but you still probably want to build in those safety guards cuz even if Google

googlebot does well you know smaller search engines you know duck ducko blacko whatever they might not spend quite as much time on the page processing so it's still probably a good idea to have the fall back and have those static links somewhere on your site at least for the you know sort of foreseeable future Matt hey there I have a qu two questions um the first one is for people who are trying to establish Authority how do they compete with

people who have been around for a long time and I'm looking for some new ideas um for example um with like usability could they improve their product descriptions to help um improve the you know establish themselves I'm thinking of local businesses and how they can compete with the larger ones and perhaps they know their products a little bit better can they you know enhance their product descriptions or captions um naturally and you know will Google notice that um

also are ebooks like um PDFs do they contribute to Authority okay cool okay so good questions um how can a small site build Authority and it is possible but it's definitely a lot of elbow work right there's there's not a lot of shortcuts um but there are some really good best practices it never hurts to go back to basics like some of the things that work really well are having a blog or having a forum you know having that unique

content having something that's you know if if your business model is I'm going to grab an affiliate feed I'm going to just slap that affiliate feed up on the web or I'm going to spin it with a thesaurus and I'm just going to slap that up on the web that is not the sort of thing that stains the test of time but if you're writing unique descriptions if you make them a little bit more catchy a little bit more

compelling something like that that that could be a great way to work anything that's creative right if you think about what were the last five things you shared on Facebook or Twitter they're probably something creative if it's a video it's fun it's coming at it from a different angle so I'd say a combination of back to the basics things like a Blog things like a forum you know reaching out to the local newspaper all that sort

of you know just straight normal vanilla got to put in the time kind of stuff and then at the same time think about what is the creative thing somebody was like Plumbing how are you going to make Plumbing interesting you know what there are probably a ton of Plumbing stories right I had to get my uh uh the pipes in our house snaked out recently and I spent the whole afternoon with the plumber and he kept me laughing the

whole time right he was like oh you wouldn't believe this one guy he oh man it was like the whole afternoon passed in like the the blink of an eye and if that was on a Blog I would like totally read that there's it's no longer on the web that I can find but there was there was an adult video store clerk who would write about her experience of being an adult video store and like I spent an

entire you know night just reading like you know I'm not especially interested in adult video stores but I was interested in her stories so those kinds of things developing a personality developing a voice all that stuff works really really well the other question is about citations are mentions in PDFs or text we tend to be a little bit more skeptical of that and the reason is I don't know if you guys know this but there are some spammers in the world and

the spammers they'll use these automated packages you know X rumor scrape box whatever you know and they just say I am going to make thousands and thousands of blog comments and if you'll notice they do it very opportunistically they're going to sign it with HTML they're going to sign it with BB code they're going to sign in these links in all these different formats and then they'll also just throw in like a plain text you know

hyperlink so that's why I worry about things like citations because you would see a lot of spammers who are like I'm just going to leave a PDF lying around all over the place with all my UR L and we've seen that we've seen spam ebooks on Amazon where they're just like automatically generated and they actually have spam links in them so I'm not going to take that out the table as a signal we you know we other than you

can't buy ads and it won't help your ranking we pretty much try to leave ourselves an open area as far as potential ranking signals but I would be a little skeptical about that because I'd worry about the spam implications I and the second one was related to the blog things I heard some sort of rumor that there's a limit to how often you can blog is there any truth to that uh I would I would blog as often as you well

as you want to and you reasonably can there are some sites that have very high volumes The Verge in gadget you know some some sites that basically like you you come back and there's a whole new set of you know Huffington Post this sort of stuff um but that's usually because they have a lot of writers they have a lot of contributors stuff like that I wouldn't I wouldn't overdo it you know I actually have less than a

thousand blogs on my website right so in oh Lord 8 years of blogging I've written less than a thousand articles right so it's really more about what are the right articles I have one article that's like why Google won't remove that page you don't like and everyone whoever writes in and says oh I don't like this algorithm or I want this page to go away I can send them to that one article so often times quality over quantity

actually will prevail there time for one more question right over here Paul hey man my name is Paul from promota a nice talk um I would like to ask one specific question you guys did a great job with the Google algorithms and that's basically helps me to helps a lot of our clients to educate and they basically respect a lot more what we guys what we do and for one client that that's thr into the problem for another one that's

turn into the new opportunities my question is I came from the Ukraine which is Russian speaking country would you guys want to do some changes over there because the results in the Google on those kind of markets is still spamming what is your approach with that I agree we need to do more in other other countries you know the engineers and the web spam team are primarily based in uh in Mountain View and so we speak English we have some people that

speak Romanian we have some people that speak Chinese but I totally agree that in a lot of other languages things are worse sometimes that's because there isn't as much content like in Arabic for example that sort of thing uh but sometimes it's because we need to do better we need to make sure that our internationalized algorithms work you know more effective um especially for example Russia has been a big focus in the last few months so the team that's been working on the

reducing the amount of top above the fold at heavy Pages that's actually a team in Russia and they were driven because they were like this is a problem here uh we've got people who give us complaints in you know in Turkey who give us complaints in Polish so I know we have a long way to go uh we'll continue to work on it and hopefully as people make great content and we keep continuing to focus in other countries

we'll drive that spam level down around the world one other question that came up in several of the sessions and in yesterday's uh illustrous keynote um this concept that you know we allow you guys to use our content kind of this Unwritten partnership there implicit partnership that uh you give us a great service and you can use all our content and now over the years with the the knowledge draft and other things you guys have done that that bar on the line

has continued to move down uh the number of pixels you guys take on my uh Netbook I popped open last night I one search there literally wasn't anything but Google on the top of the page you had to scroll funy results I've never heard anybody talk about that from Google you got any comment on that or yeah it's it's interesting I I grill my officemate AMT singal and he then goes and talks to the the head of the ads or the engineer

lead on the ads product area and we do keep an eye on that and we don't want to let that get out of control so one thing that we've actually worked on if you look at the the newer versions of chrome there we've done a lot of experiments with if you've got an address bar and you've got a Google search bar you've actually got a lot of screen real estate that's kind of redundant and we've played around with taking that real

estate back and sort of unifying the address the Omni bar and the Google search box a little bit and that might actually give you quite a bit of extra pixels and then that those extra pixels can be dedicated to organic rather than to ads but uh you know I certainly hear people who you know a lot of the times when people are complaining they'll pick one search that's like credit cards you know or they'll they'll pick something

that's kind of like okay it's got a lot of ads when the vast majority of searches are more like ancient Mayan art or why did the Civil War start or something like that but at the same time we are aware of that we we are keeping an eye on that and if people keep really poke us you know don't feel shy about poking on that because that gives me more material whenever I can walk around within the company and say look people

are saying that this becomes a problem um I think in general our mission statement is to provide the world with information and if we can give answers to people we you know they do want to give answers they don't want to only point to web pages if you know how old the Prime Minister of Turkey is so uh the way I've heard it expressed is you want to add enough value that you know it's not just one little freeword

factoid that's on your website you really want people to be landing there to get high quality reviews or high quality content or discussion you know to read really read about why should I like the the iPad Air or something like that where if it's like what date was the iPad Air released that's a little factoid and that you know that's a little harder to Peg the amount of value for users so look for the in general the

more value you're added for users the more we want to return you in the search results all right thank you very much I appreciate it Matt the exhibit hall om report and Andre alow would like to thank you for your attention you can get more episodes on www report.com