joi, 8 noiembrie 2012

Algo Hunters - An Interview with Barry Schwartz

Algo Hunters - An Interview with Barry Schwartz


Algo Hunters - An Interview with Barry Schwartz

Posted: 07 Nov 2012 06:51 PM PST

Posted by Dr. Pete

Barry SchwartzIf you’ve spent any time in search, you’ve probably heard the name Barry Schwartz (aka RustyBrick). Barry is the founder of Search Engine Roundtable, News Editor for Search Engine Land, and a driving force in the hunt to understand the Google algorithm.

I was formally introduced to Barry while I was building the MozCast project, and he’s been very generous in helping me to understand the methods we use to track algorithm updates. Many people don’t realize that it’s a manual and often painstaking process. A very, very small number of people (like Barry, Danny Sullivan, and Ted Ulle) actively monitor “chatter” – webmasters talking about changes on forums like Webmaster World – and try to interpret whether that chatter indicates an event. When they’re reasonably sure something important has happened, they report it and try to get Google to answer the million-dollar question: “Was there an algorithm update?” Barry is one of the few people who has been able to consistently get Google to answer that question.

Given our mutual interest in the algorithm, Barry and I thought it might be fun (and hopefully informative) to do something a little different. So, we interviewed each other, round-robin style. I got to lead things off…

Dr. Pete:

(Q) How did you find yourself becoming a liaison between the search community and Google? Not many people have a direct line to Matt Cutts, and I'm guessing you didn't set out to be the messenger for algorithm updates.

Barry:

(A) I think it just happened naturally. It started when I started covering the search industry in 2003. I just began reporting what I saw the search community talking about. I assumed a lot of that was misinformation that the search engines wanted to correct, so they started reaching out. I wasn't a fan of asking for information early (i.e. during news embargos or pre-release). I liked to find the information through leaks and small tests the search engines would run. But as I started at Search Engine Watch with Danny, now Search Engine Land, it became more official.

(Q) You know, there are many times I honestly wish I didn't have that connection. I like the grassroots approach of the old days. A lot of what you do is analyze data patterns to detect changes without having an official connection to Google. I bet it is incredibly satisfying when the data is right. I feel that way when I break updates before they are official. How do you feel?

Dr. Pete:

(A) To be completely honest, I'm having a lot of fun right now. Data science in our industry is still pretty new, and it's wide open. For me, it's like being one of the first people to land on a new continent, and I'm discovering things about the algorithm every day. Since I'm not getting my information from Google representatives, some people seem to paint me as anti-Google, but that's a pretty simplistic view. To me, you and I are both looking at the same problem from different angles, and the direct line to Google is a critical piece of the puzzle.

(Q) I am starting to see that darker side, though - my post about the EMD update is at 243 comments and counting, and many people are angry. I've worked with a lot of SMBs who have been ruined by algo updates, and I get their anger, but I don't get the "shoot the messenger" mentality. Why do you think people are so quick to lash out against people like you and Danny Sullivan?

Barry:

(A) They are quick to lash out because they need to. Imagine, you have a site that is making a modest living, and you are paying your bills for the past several years through your e-commerce store. Then one day Google decides to update their algorithm and your site drops off and your sales drop 90%.  That is huge and I could totally understand that they need an outlet to yell at someone.

They scramble, search for things to see what changed and stumble upon your article, my article, Danny's article or whomever. They see us quoting Google, they see us explaining in a calm way what happened and why Google thinks it is a good thing. They can't yell at Google. They can't comment on the Google blogs. So they comment on our sites.

I've been threatened, yelled at, cursed at and called names. Over the years, you learn to grow thick skin. Honestly, I am surprised there has never been a documented incident of someone actually getting hurt. I am surprised Matt Cutts can walk around at a conference without a bodyguard. Some might think that is ridiculous, but honestly, I see things from the inside and it is nasty out there.

(Q) How are you dealing with it? Do you remember the first time you had one of these incidents?

Dr. Pete:

(A) I go back and forth, I admit. As someone who used to work with SMBs and has seen businesses fall apart and lives ruined, I sincerely sympathize. Some people just put too much in the Google basket or rested on early wins for too long and didn't really do anything wrong, and I feel for them. On the other hand, I see the daily barrage of complaints from sites that are nothing but scraped content and have profited for years off of other people's work, and I get angry at their outrage after making thousands of dollars contributing little or nothing.

On my good days, I try to remember to have empathy, both for the business owners and for Google. I'm beginning to realize just how difficult the search problem is - for every winner there's inevitably a loser. At the same time, I recognize that every "loser" isn't just a number on a rank-tracking site - it's a real person with real skin in the game. So, like you, I try not to take it personally. I think you raise an excellent point - people lash out at the messengers because those are the only points of contact they have.

(Q) Every angry webmaster has some theory about how Google is just giving favors to big brands or trying to make us click on more ads, and yet there are real people at Google working full-time for a team dedicated to search quality. So, let's ask a hard question. Do you think Google cares about search quality? I have my own opinions, but I'd like to hear your take.

Barry:

(A) Yes, I think they really do care about search quality. I think it runs through Matt Cutts’ veins. I think Google won't hire people to the search quality team if they don't. People do burn out and they leave, but the core people who manage search at Google care deeply about search quality.

(Q) I guess you disagree?

Dr. Pete:

(A) No, not at all. I didn't want to bias your answer :) I actually agree – I think Google's idea of "quality" may often not sync with our own individual ideas, but I absolutely believe they care about search quality. Here's the thing: people always say that Google only cares about advertising, but then they run off and assume this means sacrificing organic search for paid ads. Organic search is the portal to paid ads. If people lose enough confidence in search results, they'll stop going to Google, and the ad revenue will dry up. I can't speak to any one person's motivation, but there's a multi-billion dollar incentive for Google to care about search quality.

(Q) Of course, the trick is that Google serves search visitors, not webmasters. How do you think that being SEOs and website owners blinds us to Google's broader mission? Are we the ones who have forgotten about search quality, or are we just stuck between two competing perspectives?

Barry:

(A) Well, since everything I publish to the web is better than anything else already published on the web and anything that will be published to the web in the future, all my content should be ranked number one in Google. There are two things here:

  1. Owner bias, i.e. my content is mine and thus the best.
  2. We understand search and think we know what search quality is. But you and I don't study it in the same way Google does.

We need to step into their shoes and outside of ours.

(Q) …but that is where you differ. You see the data – not as much data as Google, but you see more than a site owner whose single site tanked in Google.

Don't you think what Google felt was quality two years ago is different from what they would say it is today?

Dr. Pete:

(A) It's tricky. As humans, I think we have a decent gauge of quality and Google's quality raters play into that, but ultimately they have to translate quality into algorithmic signals, and that's where it gets tough. Google is a coder culture, and they want elegant, algorithmic solutions, not to flag domains and manually penalize people. That's just not scalable. So, I don't think their definition of quality has changed so much as their ability to translate quality into something a computer can quantify and evaluate has.

There's also the arms race, and this is where we SEOs share some blame. After the EMD update, some people posted that Matt was being inconsistent, because he said a couple of years back that exact-match domains were a quality signal. I don't think Matt was inconsistent at all - in the past, EMDs often were a legitimate signal of relevance. Then, EMDs got spammed to death, and now they aren't as relevant. So, Google had to adjust.

(Q) I'm sure we'll be accused in the comments of being both Google apologists and Google haters, so let's mix it up and swing back the other way. What's one thing you think Google got wrong in 2012 (could be an algo update, a policy statement, a new feature, etc.)?

Barry:

(A) Got wrong is a tough statement. Much of what they do is trial and error. I like that.  I dislike what they are doing to Google shopping making it purely a paid inclusion thing. But who am I to say they did something wrong. Let them test it and learn from those tests.  

(Q) Isn't that how all data guys like to do things?

Dr. Pete:

(A) You may be surprised to hear it, but the one thing I think they may be doing wrong lately is testing too much. Google has mountains of data and can find statistical significance for even the slightest differences. So, they roll out 7-result SERPs and determine that the faster page load improves some metric (let's say bounce rate) by 0.4% (I'm making that up), and that's significant across the millions of visitors they test it on. What if it negatively impacts something they aren't measuring, though? What if 0.4% is statistically significant but doesn't practically matter? I wonder if data is starting to overwhelm common sense, and if maybe a data-driven culture can go too far. Obviously, these are tough questions, and I'm not seeing what they're seeing, but I think you can definitely overdo even a good thing.

(Q) You've spent a lot of time watching and trying to interpret what Google is doing. Given what you know, what's the biggest change you expect Google to make in the next year? I know this is pure speculation (I doubt even Google has it charted out), but I'm sure our readers would find your point of view interesting.

Barry:

(A) On the algorithm penalty side, I think over the next year, we will see Google probably target yet another usability factor in their algorithm.  We had Panda and the page layout algorithm, we also had page speed.  I'd look at an additional usability factor to be targeted with a Google update.

In the ranking algorithm side, I'd think we would see more social (Google+) come into play not just in personalized search but in weighting certain pages – if not in 2014, then for sure by 2015.


I'll throw in my two cents - I think we've only seen the tip of the iceberg with Knowledge Graph. There's a lot of code and data behind what may look like a toy to some people, and we're seeing the real beginning of the semantic web. I expect it to have huge organic search implications in the next couple of years.

Thanks to Barry, not only for the interview, but for being so generous with his time and information over the past few months, not to mention for his contributions to the industry in general. On behalf of both of us - try not to shoot the messenger.

I'd also like to announce, for those who haven't heard, that November 1st marked my first official day working full-time with SEOmoz. I'll be expanding my role with the marketing and data science teams and putting more time into "big content" and projects like MozCast. I'd like to thank the community especially for your support over the past 5 years, and I look forward to spending more time together.


Sign up for The Moz Top 10, a semimonthly mailer updating you on the top ten hottest pieces of SEO news, tips, and rad links uncovered by the Moz team. Think of it as your exclusive digest of stuff you don't have time to hunt down but want to read!

Niciun comentariu:

Trimiteți un comentariu