Blind Five Year Old

Last week Google made a major change to the +1 button, allowing users to share a +1 with Circles on Google+. Essentially, the +1 button now has similar functionality to Facebook’s Like button. I’d been procrastinating on Like button optimization so this was the impetus I needed to finally do something about it.

Here’s what I did to optimize the Google +1 button. Let me state up front that I am not a coding guru. But I know enough to be dangerous, enjoy tinkering and am not afraid to break my site as I experiment.

I recommend this only for those using a self-hosted WordPress blog.

Speed

When the Google +1 button first came out it was s l o w. It was the last of my buttons to load. But that changed when Google released asynchronous code. If you haven’t made the switch yet, go and get the new code.

I use the Smart Sharing plugin for my floating share bar. So I just dropped the appropriate code snippets into the ‘Custom Codes’ area. This is really easy. After doing this, the +1 button is often the very first to load.

Placement

Where you put the Google +1 button matters. The floating share bar on this blog, powered by the Smart Sharing plugin, did require a fair amount of CSS customization to display properly. (Thank goodness for Firebug and other sites who use a similar display.) But the customization has been worth it!

It’s not just about where but how many buttons you present. A huge row or drop down full of buttons can cause indecision. Too many choices are bad. So pick the buttons that are meaningful to your audience and use them.

I get far more social engagement with a handful of  floating buttons than a slew of buttons at the top or bottom of the post.

Snippet

Now that speed and placement are taken care of it’s time to optimize the snippet. The snippet is the title, description and image displayed when someone shares your content through a Like or Google +1 button. Controlling what this looks like is important.

Here’s an optimized snippet from this blog.

And here’s a non-optimized snippet from the Official Google Blog.

Now, maybe Google can get away with a poor snippet but the rest of us probably can’t.

Structured Markup

Optimizing your snippet means adding some structured markup to your page. This is a lot less complicated than it sounds. In fact, both Facebook and Google are omnivorous and will parse your site looking for any way to cobble together the best snippet.

Google recommends using the new Schema.org microdata. But they seem happy to fall back on the more ubiquitous Open Graph protocol (funny how they don’t mention the word Facebook) or normal meta title and description tags. In fact, the only method they don’t recommend is having absolutely no structured markup at all.

Don’t obsess about implementing microdata because Google certainly isn’t.

Most of you probably have the title and description taken care of, but it’s the image that might be causing you some heartburn. In many ways though, that might be the most important part of the snippet. People scan content and a good image is like a stop sign for the eyes.

For a long time Facebook was using my Feedburner chicklet for my snippet image. I got tired of seeing this and created a default image for all of my posts. It wasn’t optimal but at least I didn’t look like a buffoon anymore. Read on to see how I took things to the next level.

Plugins

Like millions of others, I rely on a number of plugins to add features and functionality to my WordPress blog. It’s important that you understand the plugins I’m using since it will impact how you might implement snippet optimization with your own WordPress configuration.

I use the Like plugin to generate my Open Graph tags. It’s worked well, though I may think about doing it myself after going through this process. In general, this plugin is easy to use and efficient. It’s this plugin that allowed me to hard code a default image for my snippet.

I use All In One SEO to optimize title and meta description, among other things. I hear good things about WordPress SEO by Yoast, but haven’t had any real reason to switch. I may eventually, but I use other plugins that may have to be deactivated to take advantage of the full feature set offered.

I tried the Schema for WordPress plugin and admire the effort, but find it too complex and onerous. In addition, I don’t believe it supports the Schema meta tags which is what the Google +1 button seems to want.

Frankly, this is part of the problem with the microdata implementation. Do I put it in the head or the body? Or both? It’s almost too flexible.

Custom Field

So, with all those Plugins in place, my main task was to find a way to generate a different image for each post. That’s where a custom field comes in handy. Custom fields sound complicated but they’re actually pretty straight forward. I created a custom field that would contain the URL of the snippet image.

You’ll see the Custom Fields option in your normal Edit Post window.

Click ‘Enter new’ and you’ll now be presented with an entry field.

Just enter the name of your custom field (I used og_img but you could name is something different) and the value you want to give that custom field for this post and click the ‘Add Custom Field’ button. In this case enter the full URL path of the image you want to use for this post’s snippet in the value field.

This creates that custom field in the database and associates the URL for that custom field for this post. This does not mean that you are assigning this value (this URL) to every post. This is just the way you get the custom field ball rolling.

Now you’ve got the custom field in place for that post and you can select that custom field from the drop down menu for all other posts from now on.

In fact, that’s how you’ll specify the image snippet for each post moving forward. You only use the ‘Enter new’ to create the custom field. Once created, always select it from the Name drop down menu.

Yes, this means that you’ll be taking one extra step each time you publish a blog post. I see this as a small price to pay for an optimized image snippet and lump it in with writing an optimized title and description. They’re simply part of the publishing process.

Image

The image you use can make or break your snippet. It’s not just about the content of that image but the size of that image. You could have the best, funniest image ever but if it doesn’t translate on social platforms then it’s all for naught. In this case size actually refers to the dimensions of the image.

A square image (e.g. – 300×300) is a safe bet. If you’re not using a square or squarish image then you want to pick vertical rectangles over horizontal rectangles. Obviously, really long vertical rectangles won’t work either. Use commonsense, test and learn as you go.

One other thing to note is that the image you use doesn’t actually have to be in the blog post. I personally think it should, but you can upload any image to use as your image snippet. This might come in handy if you don’t have any images in a post (not recommended) or images that won’t translate onto social platforms (such as a horizontal code screen capture.)

Header Code

Creating the Custom Field only gets you half way there. Then you have to use it within your theme. That means you’ll need to tinker with your header.php file by navigating to Appearance -> Editor -> Header (header.php)

First things first. Copy your entire header.php file into a text file just in case you really screw something up and need to revert back. Just copy everything in that window into your favorite text application.

Next, you want to place the following code into your header.php file between the head tags (<head>, </head>). I recommend placing the code in just before the close (</head>) tag.

<!-- Facebook Open Graph Image -->
<?php if(get_post_meta($post->ID, 'og_img')){ ?>
<meta property="og:image" content="<?php echo get_post_meta($post->ID, 'og_img', $single = true); ?>" />
<?php } else { ?>
<meta property="og:image" content="https://www.blindfiveyearold.com/wp-content/uploads/2008/09/blind-five-year-old-150x150.png" />
<?php } ?>

This code creates an og:image tag using the image in your custom field (og_img) or, if there is no custom field for that post, will use the default image specified instead. This is essentially the code that starfly describes in this WordPress forum post.

Make sure you change the custom field name appropriately (if you didn’t use og_img as I did) and that you use your own default image instead of mine. The first line of this code is non-functional and is there only to help you identify it amid all the other code.

After you’ve entered this code, click ‘update the file’ and you’re done!

Validation

To test whether it works you can first look at your blog post and view source. Look for the og:image tag and make sure that it’s populating correctly. If that’s not your cup of tea, you can also run your blog post through Facebook’s URL Linter Debugger.

Finally, try sharing your post on using the Google +1 and Like buttons and make sure it’s optimized to your liking.

Where’s the Microdata?

I intended to show you how to cover your bases and implement an additional set of microdata in your header. It was going to be pretty darn clever. In fact, I spent two days figuring it all out, cursing PHP and breaking my site for hours on end. Trust me, I want to show it off!

But as I did due diligence on this post the microdata seemed to confuse Facebook and was, overall, just too brittle to use at this point. Honestly, do you really want to maintain two different types of structured markup?

I can see using microdata in other ways, but not for snippet optimization.

Google +1 Button Optimization

So, in short, I’m using All In One SEO and the Like plugin along with a Custom Field to generate Open Graph tags that are then used by both Facebook and Google in their button snippets. Two birds with one stone!

I welcome questions on my implementation as well as feedback from others who can improve my code or have found other ways to optimize these snippets.

Over the last two and a half years I’ve become convinced that Google performs rich categorization of sites and queries. The signs are plain as day and the impact is substantial. Categorization allows Google to apply algorithmic changes by category as well as deliver more relevant and diverse search results.

Site Categorization

In late 2008 I wrote about how taxonomic search could be used to create diversity in search engine results. Later, In May 2009, Google launched search options that included the ability to filter your results by Forums and Reviews.

Rich Snippets stole the show during this launch but it should have been crystal clear that Google was able to distinguish site or page type. How else would it deliver results in these filters?

Google made this even more transparent when they tested the ability to see more or fewer shopping sites. Obviously Google was able to tell (with relative accuracy) which sites were eCommerce enabled.

You may also come across search results for Q&A sites or forums that include a pseudo-rich snippet which includes the number of posts and authors.

This is not structured markup. This isn’t RDFa or microdata. Instead, Google has identified patterns in these sites, extracting the relevant information and creating their own rich snippet. More recently, Google demonstrated that it could determine if a page contained search results.

These are interesting applications but it’s what is at the heart of it all that is really remarkable – the ability to algorithmically identify and categorize a page or site.

Query Categorization

If it were a snake it would have bit you already. Of course Google categorizes queries! You see this when local results are triggered, or when the Onebox is presented. Enter a query that is product based and the Shopping Onebox is likely to appear. Enter a query that is health related and the Health Onebox may appear.

It’s not a matter of if, but how deeply Google categorizes queries.

Matching Queries to Categories

Matching queries to categories can help deliver the right information to users. You can already see this happening as Google changes the left hand search options based on the query.

You won’t get a Recipes option if you search for ‘baked new jersey’.

You won’t get a Books option if you search for ‘crying of lot 23’.

These examples could be delivered by simply accessing relevant databases of recipes or books, but I’m guessing the relation is far more dynamic.

Matching Documents to Queries

The question then becomes how Google matches web pages (aka documents) to queries. What documents (or results) does Google return based on the query categorization?

A recent patent dissected by Bill Slawski shows how Google might think about matching documents, queries, keywords and categories. It indicates that what Google thinks your site is about could influence how and when it is returned for certain queries.

If you were to take all of that information that Google provides for your site, and try to guess at a category or categories that Google might assign for your site, could you?

That is an interesting question.

Google Ad Planner

One of the unconventional places I’ve looked at recently to give me a sense of what Google might think a site is about is Google Ad Planner.

The content categories are usually a mixture of accurate and bizarre, particularly around geography. Now, I’m not saying that the content categories here influence Google’s categorization. But I have found it illuminating to look at Ad Planner results when comparing competitors.

Beyond the content categories are affinity scores for sites also visited and audience interests. The latter is another interesting data point when thinking about how Google might categorize sites.

The good news is Google has a compelling reason for this data to be accurate (it’s attached to advertising) and the data seems to be updated frequently. As an example, the Ad Planner thumbnail for this blog shows my most recent blog post.

Why Google Ad Planner thinks I’m in Canada is a bit of a mystery. I did do a blog posts about Hockey Memories and The Flyers 1987 Stanley Cup Playoffs, but I’m not in Canada (as lovely a country as it may be.)

After verifying site ownership you can change your categories, which reveals a fairly robust taxonomy. Is this the same taxonomy Google uses in their algorithm? Probably not. But it might help inform or update the one used for search. So, I’ve gone ahead and changed my categories and description.

It’s in my can’t-hurt-and-might-help category.

TL;DR

Google continues to work on matching categories to queries and queries to web pages to improve search quality. Understanding how your site might be perceived by Google is an important new step in search engine optimization.

I had the opportunity to attend last night’s Churchill Club event held at the Computer History Museum that brought together Amit Singhal, Ben Gomes and Matt Cutts.

Inside Google’s Search Office marked the first time the three have appeared together. If that wasn’t enough, the event was moderated by Danny Sullivan who is both incredibly smart about search and a gifted moderator.

For someone who follows search closely, this was a can’t miss event. Thanks to Rick Bucich (who also helped this introvert mingle) for the heads up on the event.

This was not an SEO conference or event. Very few in the audience were hoping to capture that ‘gotcha’ moment. Instead, we were given a more human look at how these three Googlers think and feel about their work.

Passion

All clearly have a passion for search. Amit, in particular, was animated and visibly enthusiastic about the subject matter, at times perched on the edge of his seat as he told a story or answered a question. Matt was his usual self, mixing down-to-earth advice with a goofy sense of humor. Ben, the most reserved of the bunch, lit up when talking about specific challenges faced and overcome.

I saw Amit bristle at the idea that AdWords spending influences organic search. I watched Matt grimace at the mention of Eli Pariser’s Filter Bubble and the conclusion that Google was doing something untoward in personalizing results. These things generate a response because they care.

What they care most about is doing right by the user. This often sounds trite or contrived. Some claim that this altruistic goal is just a facade for greed. But I don’t think any of them would dispute that Google prospers when they do right by the user. They strongly believe that one leads to the other.

As an aside, I’m not sure who would win a gesticulation war between Matt and Amit.

Testing

We all know that Google does a tremendous amount of testing. Amit described the process Google goes through in approving an algorithm update. It reminded me a bit of Schoolhouse Rock’s How a Bill Becomes a Law.

Many changes are tested but only a few make it through the entire process and result in an algorithm update. Based on the ballpark numbers mentioned (20,000 changes and 500 updates), less than 3% of the tests result in an algorithm change.

Relevance

One of the things I’ve written about previously is the subjectivity of relevance and quality. One man’s trash is another man’s treasure. This was the first time I heard Google address this issue head on. While they believe that their scientific method allows them to come closer to an objective relevance they know it won’t be perfect.

Matt talked about the fact that each algorithm update will make some certain amount of queries better but some much smaller amount of queries worse. The idea that each algorithm update will only have beneficial results is … fantasy.

I’ll paraphrase another quote referenced by Danny and Ben. Two humans will agree on relevance only 80% of the time. If you looked at that same result a year later, you may not agree with yourself, let alone someone else. The implication (one I happen to agree with) is that relevance is a moving target.

These statements showed a humility that I had previously found lacking. Google should rightly be proud of having the best search algorithm on the planet, but it was refreshing to hear them admit that it was an imperfect work-in-progress. Google is on the cutting edge of this field and they continue to work (hard) at getting it right.

Speed

Speed was a central theme in the discussion of search innovation. It got me thinking more about measuring quality and relevance based on the amount of time it took from query to click.

Obviously Google has worked to reduce that time through Google Instant and, most recently, Instant Pages. But might Google score search results on how quickly they satisfied query intent? That doesn’t mean the user clicked on the first result, it simply means that the results presented yielded the best user result as measured in time to productive click.

Think about that for a moment. Is Google less concerned with getting rank right and more concerned with optimizing for quick knowledge transfer? Clearly having the right listings at or near the top matters, but perhaps a combination of listings is most effective. We focus on the percentage of clicks by rank, but that may not be the dominant metric Google is using.

Social

Google seems both excited and cautious about how it integrates and uses social signals. Amit reduced social to ‘who knows who and about what.’ But there’s a lot more to be done here and identity is clearly on their mind as a way to strengthen this signal.

In addition, I believe Google is a bit stunned at the quick adoption rate of Google+. I’m not saying they didn’t think it would be successful, but it feels like they didn’t expect these types of numbers. At this stage, I think they’d like G+ users to take a breath and provide them some time to iterate.

Googlers

This was also a fun event for me because I met a few more Googlers. I got to talk with Sagar Kamgar and Jeremy Hylton about rel=”author” and how it may change moving forward, met Jim Haljun and chatted briefly with both Amit and Ben. And I was happy to reconnect with Matt after missing him at this year’s SMX West.

I’m not a Google apologist. I think my writing makes that clear. But I find it difficult to stomach claims that Googlers are elitist liars who feed the SEO community misinformation. This irritates me because nearly all the Googlers I’ve met (online and in person) have been smart, friendly and very generous with their time.

Future

Amit and Matt both shared their visions for the future of search. In five years Amit would like to see search behaving more like the ship’s computer on Star Trek. He’s keen on voice search and the growing impact of mobile devices.

Matt’s vision was limited to the next few years where the use of geo-location, mobile devices and other contextual data could allow Google to prompt users. Matt would never forget to pick up cat food again! He’d also like the voice to text ability to be refined so he can respond to email on his drive home. “That shouldn’t be that hard. I’ll file a bug when I get home.”

Beyond these predictions I was caught up in their passion and reminded of why I love search: it’s full of amazingly talented people with inquisitive minds working on something that is constantly evolving.

Googlebot does not laugh. It is not the lovable Johnny 5 and is probably closer to HAL in disposition.

Humor Signal

Does Google use humor as a search algorithm signal? My first instinct is to state, without hesitation, that there is no humor signal. There’s no analysis of puns or witticisms or double-entendres or other grammatical humor.

The clever pun you use as your browser title will be lost on Googlebot and, truth be told, the majority of your readers as well. Googlebot doesn’t have a funny bone.

Backhanded Signal

Yet the algorithm will reward humor through other signals. Something that’s really funny will attract links and generate a high number of Likes, Tweets and other social gestures.

So humor is reflected in the algorithm through the link graph and social signals.

Research shows that our sharing behavior is heavily informed by humor. We share to get a laugh. Just ask Ben Huh, CEO of I Can Has Cheezburger?

But that’s an indirect signal, sort of like a backhanded compliment. Google doesn’t really know it’s funny, it just knows it’s popular and engaging.

Computational Humor

Google might be able to determine whether a document is humorous or not. Computational humor, a relatively new field of study, has shown a moderate amount of promise.

Early computational humor research focused on producing jokes, resulting in the Joke Analysis and Production Engine (JAPE) and HAHAcronym, which does just what it says. More recently, there’s been a greater emphasis on identifying humor using natural language processing and machine learning (pdf).

Researchers were able to train an algorithm, with high accuracy, to identify humorous one-liners from standard headlines and even proverbs, which had similar sentence construction.

Google Humor

Is Google interested in humor? It’s tough to tell. But a recent patent, brought to my attention by (who else) Bill Slawski, titled Embedded Communication of Link Information contains an interesting idea.

The improved link tags may allow the publishers to communicate additional information, such as opinions, about the content locations and/or document locations. The additional information may be along one or more dimensions. Therefore, different information may be conveyed at the same time. For example, one dimension may indicate that a content location and/or a document location is offensive as well as funny.

For example, “offensive=very” or “funny=somewhat.”

Coincidence? Probably. Nevertheless, I think it’s interesting that humor was one of the examples selected. And most of the Googlers I know have a keen wit.

Humor and Quality

At the end of the day the real question is whether humor and quality are related? Does the incidence of certain types of puns or other humor signal quality? Could quality be predicted based on Monty Python references? Or the presence of alliteration?

Could you construct a database of humor? Monty Python phrases. Caddy Shack quotes. Famous puns. Palindromes. Internet memes. Knock knock jokes. Using that database, could each type of humor be given a score based on how often it appears in what quality of content?

Those scores in place, you can begin to use humor as a signal to help ‘turn up’ high quality content. Perhaps … to 11.

Perhaps the relationship between content quality and humor doesn’t exist. There’s a tremendous amount of personal taste involved and the cultural differences are also substantial. In addition, much of our humor is visual in nature and for all the advances in image search Googlebot still can’t really see.

Yet, humor is clearly a compelling part of the content puzzle. And there’s a part of me that believes that the mere presence of humor indicates a higher level of care and passion, which dovetails nicely with Google’s recent guidance on high quality sites. I doubt you’re going to see humor (at least intentional humor) on splogs or MFA sites.

I still don’t think there’s a humor signal. But maybe there should be. It would be nice to reward writers for their spelling, grammar and creativity.

Think you got a bump from Panda 2.3? Not so fast.

Image Search Analytics

In looking at a number of client sites I notice that image search traffic, tracked under referring traffic (google.com / referral) with the referral path of imgres, fell off a cliff as of July 23rd.

Where did that image traffic go? Organic.

So if you thought you’d been the beneficiary of Panda 2.3 (launched late last week), you might want to make sure it’s not a phantom image search bump.

The Definition of Organic

At present I can’t find an easy way within Google Analytics to distinguish between organic traffic that is search based versus image based. That strikes me as a step back since these forms of traffic are not homogeneous in nature. Lumping image search in with organic is like smearing vaseline on your windshield. I can still see, just not as well as I could before.

There’s probably a hack you can put together via filters, but most users won’t make that effort.

Where’s Image Search?

This isn’t the first time Google has played Where’s Waldo with image search. On May 6th, 2010 Google moved image search traffic from images.google.com to google.com.

At least that time you could wander around Google Analytics and spot the new source/medium that would provide the same level of specificity. Oddly, you’d still see some stray images.google.com traffic after this change. I always meant to track that down but never got around to it. This new update seems to finish the job and eliminate the remaining images.google.com traffic that had been trickling in.

New Dimension Please

I am hoping that this is just evidence that Google Analytics will launch a new dimension so we can separate these two different types of search traffic. Yet, you’d think they’d launch the dimension before migrating the traffic.

For a long time I figured that these changes were an indication that image search was the ugly duckling of the bunch. But recent events make me believe that Google is very invested in image search, so why the lack viable reporting? No, ‘it’s free’ is not the right answer.

I’m waiting to hear from a few Google sources and will update this post if I get any type of insight or confirmation. Until then, how do you feel about this change?

Google+ has seen unprecedented adoption over the last month and it is clearly the best social effort Google has produced to date.

But why is Google pursuing social in the first place?

G+

Some see G+ as proof that Google is chasing tail lights, while others see it as a natural extension of Google’s mission, which is ‘to organize the world’s information and make it universally accessible and useful.’ In this instance, that information is social in nature.

I see two compelling reasons behind G+: search and advertising.

G+ provides Google with a primary source of social data to mine and use for search purposes. Can you blame Google for not wanting to rely on Twitter or Facebook for social signals?

G+ also seeks to break the stranglehold Facebook has on attention. This is important because the ad business runs on attention. If you’re on Facebook for 55 minutes a day, that leaves less time for other sites. G+ doesn’t have to beat Facebook it simply needs to fracture its monopoly on attention.

Identity

One of the more controversial aspects of Google+ has been Google’s insistence on people using real names.

Google Profiles is a product that works best in the identified state. This way you can be certain you’re connecting with the right person, and others will have confidence knowing that there is someone real behind the profile they’re checking out. For this reason, Google Profiles requires you to use the name that you commonly go by in daily life.

I probably won’t make many friends by saying I completely agree with this policy.

Some believe that sacrificing anonymity will sacrifice risk-taking, honesty and whistle-blowing. There may be less of that, but it won’t stop. And is it so bad to encourage people to take risks and be honest without the crutch of anonymity? For those of you who think I’m not personally aware of the issues here, let me disabuse you of that notion.

Back in 2008 I dashed off a quick email to the editor of The Contra Costa Times. It wound up in the paper and was attributed with my name and city. About a week later I got this in the mail.

This person found my address (which was not in the paper) and sent me this erudite note. It was unsigned and did not have a return address. It was a bit chilling. Yet, the anonymity used belies a cowardly nature. This person also can’t seem to get my name right, but who cares about details like that when you just want to call someone names.

I’m not saying all anonymity leads to this type of behavior. Nor do I downplay the need for anonymity in certain situations. Speaking out in other countries could have dire consequences. But one of the reasons people embraced Facebook’s new commenting system was the fact that identity would reduce spam and needless flame wars. No more anonymous pot shots from the peanut gallery.

Could Google be doing a better job in how they handle the policy? Yes. But I believe they’re far more concerned with ensuring that spam doesn’t enter this new ecosystem, both for user experience and data integrity.

What’s In A Name?

I don’t think Google is saying you have to use your true name. I’m assuming that Mark Twain could sign up, but so could Samuel Clemens. And it’s not that hard to set up a new persona online. Maybe we’d even see profiles for both Neil Caffrey and Nick Halden.

But why does Google care so much about identity, particularly on G+? What does it really do for them? Here’s where I think the rubber really meets the road.

Identity allows Google to map the web of people.

Since Google launched, they’ve indexed the web using links. That’s worked amazingly well, despite what many competitors and faux-academics have contended. Mind you, there’s still a lot of room for improvement.

Today we know that there are people behind sites and documents. Facebook understood that people were ‘hiding’ behind websites and they wanted to reveal who was visiting (and Liking) those sites. Google, on the other hand, seems more interested in who created those websites and documents. They’re looking to augment the web of links with the web of people. That’s why identity is so important.

Without identity Google can’t create a web of people. They can’t mine the interactions on G+ to determine trust and influence. They need a primary key for establishing a new social database and name makes a whole lot of sense.

Yes, Google will want alternate email addresses, pseudonyms and even usernames, but they need something to pivot around. Mapping the social graph necessitates this type of approach. Google’s already pretty good at this, which is why you may be presented with suggested links in your Google Profile. One look at your Social Connections should give you an idea of how much Google knows. (Really, go look at this today!)

SEO

The carrot that Google can wave to encourage identity is search. Rel=”author” is the first overt way that Google is rewarding identity in search results. But that’s just the tip of the iceberg as far as I can tell.

Getting your smiling mug on search results will certainly help your click-through rate and boost your ego, but long term Google can begin to map all of your social interactions. They’ll encourage you to link more and more of your digital self so they can confidently perform this analysis.

I’m a firm believer that Google has a GIGO problem. The explosion of digital content has made everyone an author, an expert and pundit. Jason Calacanis certainly doesn’t suffer fools gladly.

There are a lot of stupid people out there … and stupid people shouldn’t write.

I don’t agree with that statement but I do with what Jason said next.

There needs to be a better system for tuning down the stupid people and tuning up the smart people.

Google needs a better way to sift through and determine quality in an age where content can be produced and distributed with such ease.

Instead of going from the document level and building up, what if you used an ‘author’ filter? Google’s Panda update created an A and B pile for site quality. Google could do the same for author quality. Yes, it’s a scary thought that Google might pass judgement on your corpus of content, but in the end that’s pretty much their job.

This doesn’t mean that B pile authored content wouldn’t rank, but it may need to exhibit other signals to outrank A pile authored content.

AuthorRank

AuthorRank will help augment the rickety link graph. The reliance on a Google Profile, insistence on a real name, the implementation of rel=”author” and the whole of G+ makes me believe that authorship is a vital part of how Google wants to measure the web. I’m hopeful that Google can determine a better way to measure these social gestures, beyond what Klout, PeerIndex or others currently provide.

Could Google transform trust and influence from a popularity contest to a meritocracy?

Google could look at the speed of sharing. If the time between my interaction with that content and my sharing of that content is too short, I’d like that to count less. This may eliminate much of the robo-syndication that persists today and even reduce the sheep and ego mentality of promoting luminary content.

In addition, comments are a rich source of data, both as meta data on the shared piece of content but as a signal of depth and breadth of engagement. You’ll have noted that some of the Internet famous get a slew of comments on G+. But how many of them are in the ‘brilliant post’, ‘totally agree’ and ‘Yes!’ variety? What percentage of your comments are productive versus congratulatory?

AuthorRank could provide real value to search right out of the gate. Imagine if blogs without real identity were given less prominence? Suddenly splogs would disappear as would many of the exact match keyword domains with MFA (Made for AdSense or Made for Amazon) content.

TL;DR

Google+ requires identity so it can confidently map a web of people and use authorship to better refine search results. In the process, and as an added bonus, they may also break Facebook’s monopoly on attention.

I’ve been thinking a lot about Google’s Panda update. Who in the SEO community hasn’t, right? But there’s one thing in particular that continues to bug me.

Why is Panda applied at the site level?

A site wide quality metric seems very un-Googly. It was one of my major complaints when Panda (then called Farmer) was rolled out. It treats lousy content the same as great content. This seems to run contrary to Google’s mission to return the best and most relevant search results.

You might argue that the great content will continue to rank well based on other signals, but there’s little doubt that it will be negatively impacted. And the content that now outranks it may not be better at the page level.

Panda Mechanics

At this point I believe we have a fairly good idea of how Panda is applied. Bill Slawski (here and here), Danny Sullivan and Eric Enge have all provided great insight into how Panda might have been constructed and implemented.

In general, the conclusion seems to be that Panda acts as a type of quality filter that sits on top of the algorithm, placing a penalty of sorts on those sites it deems to be unworthy.

My Panda Theory

It seems probable that Panda is a document based classifier that evaluates and scores the quality of a page. But why not integrate that page score as a true signal so lousy content would be demoted and quality content would rise? Let each piece of content compete on its own merits.

Could it be that the confidence interval for the Panda classifier isn’t high enough on any single document?

But if you sample enough pages from a site the Panda classifier reaches an acceptable confidence level, allowing Google to pass judgement on the entire site as a whole, but not on an individual URL basis.

If the Panda score is on a scale of 100 you might wind up with something like this. So while document three scores very well with a 94, the rest of the documents on the site drag the aggregate Panda score down.

This could explain why removing certain thin content pages might impact your Panda status, since the aggregate score might rise substantially.

It could also explain why many saw sites with a small content corpus escape the wrath of Panda, since the lack of a viable sample made the aggregate Panda score invalid.

Panda and Big Data

Part of this theory draws from Mike Cohen’s presentation at the Inside Search event. He stated that the way in which Google was improving the accuracy of voice search was through “massive amounts of data.”

And if you read In The Plex (which you should) you also come away with the feeling that it is sometimes less about tweaking the algorithm and more about feeding that algorithm more data. Machine learning requires big data.

Danny Sullivan reported from SMX Advanced that  “the Panda filter isn’t running all the time. Right now, it’s too much computing power to be running this particular analysis of pages.”

Again, this seems to indicate that Panda isn’t part of the normal evaluation process and that it requires a substantial effort to recompute, even by Google’s standards.

Constraint or On Purpose?

Perhaps the iterative nature of the Panda updates will result in a more accurate document classifier that could be applied on the document level. Or maybe Google simply believes that a site’s overall content corpus should have an impact on all of the content on that site.

Maybe Google could apply Panda on the document level but instead believes site level application is more expedient.

There’s a small bit of logic there. If I buy products from a store and have them break again and again, I might not want to patronize that store even if a few of their other products were well-crafted and solid. So, the store (aka site) develops a reputation and the once bitten, twice shy adage kicks in. This dovetails nicely into the idea that your brand equity can have an impact on perceived relevance.

Of course, none of this could be even remotely true because I’m not a data scientist or statistician. I’m just a guy who reads a lot, experiments and enjoys uncovering patterns. And it doesn’t change the facts nor how to get out of Panda Jail.

What do you think? Is Panda’s site level application a product of constraint or done by design?

Overshadowed by the Google+ launch was the implementation of the rel=author markup in search results. Once implemented, authors are given a very prominent treatment on search results.

It doesn’t reorder the results (yet) but it certainly highlights that result and likely drives a much higher click through rate. I was already interested in rel=author, but this was enough to get me off the proverbial couch and try it out myself.

Unfortunately the authorship directions provided by Google, while probably comprehensive, are confusing.

Thankfully, Louis Gray got me into Google+ and it was there that I put out the bat signal for a rel=author expert. Three Google employees quickly responded and set me straight on how exactly to implement rel=author.

A big thank you to Googlers Pierre Far, Daniel Dulitz and Jeremy Hylton for their assistance. Here’s what I learned from them.

Three Link Monte

The TL;DR version for implementing rel=”author” is that it requires three specific links.

• A link from your blog post or article to your author page using rel=”author”
• A link from your author page to your Google profile page using rel=”me”
• A link from your Google profile page to your author page using rel=”me”

Read on for specific directions on how to get rel=”author” up and running on your own site or blog.

[Update 12/15/2011] While I still prefer the method described in this post, Google does allow you to verify authorship via an email address. Directions for this method can be found on the new Authorship home page.

Blog Post

The first link is from your blog post to an author page on the same domain. This is essentially a link that tells Google about the authorship of the posts on that domain. That’s why you use rel=”author” on this link. A blog with multiple authors will have multiple author pages, with the posts each author has written pointing to their own author page using rel=”author”.

But the author page does not have to be an actual author page. For a solo blogger, you can simply use your about page because that is about the author of the site.

Most templates will have the author of the post in the byline. This is where you want to place the rel=”author” link.

Here’s what I did. In WordPress I navigated to Appearance > Editor and then chose to edit my Single Post file. I then looked for the byline section and updated where it was going and added the rel=”author” attribute.  (Use the quotes!)

Now every one of my posts will have a link in the byline from my name to my about page using the rel=”author” attribute.

[Update] You can actually use the root domain as your author page if it a) has a link, using rel=”me” to your Google profile and b) if it is not on a free host domain such as WordPress or Blogger. (Credit: @pedrodias)

Frankly, I think this makes it a bit more confusing but it is another option if you really don’t have a true author or about page.

Author Page

The second link is from the author page to your Google profile page. This link tells Google that the author of that domain is the same person as the one in the Google profile. You’re essentially claiming that Google profile as your own, which is why you use rel=”me” on this link.

The best practice is to link to the base URL of your Google profile.

https://plus.google.com/115106448444522478339

This might be a bit confusing because the base URL will default to a /posts suffix. It is further complicated by the implementation of Google+ which changes the subdomain from profiles to plus.

Don’t worry. If you’re not using Google+ yet use your current Google profile URL. Google will put in the proper 301 redirect from the old profile to the new once you’re using Google+.

[Update] Another way you can link to your Google profile is by creating a G+ button. Just make sure you select the ‘author’ option when generating the code and it will insert the rel=”me” attribute. (Credit: @pedrodias)

Google Profile

The third link is from your Google profile to your author page. A link to your domain is not going to cut it.

Linking to the actual author page makes sense in light of multi-author sites and blogs. You might not be the author of all the content on that domain, but you want to show that you’re the author of those few guest posts.

Go to your Google profile and select edit profile. Then click on the Links section and click Add custom link (it’s at the bottom). Then enter your label and author page URL and make sure you check the ‘This page is specifically about me.’ box. That will put a rel=”me” on this link.

Now you have a rel=”me” attribute pointing from your Google profile to your author page. Add a link for each author page. This would include your own blog but also the author pages for any sites or blogs to which you’ve contributed content.

My OCD kicks in here since every other link I have is a very clean link to my places on the web. In the future I’m hoping Google could create a separate link list for these author pages.

[Update] Ask and you shall receive! Google has created a separate set of links labeled ‘Contributor to’ for authorship purposes. It is now recommended to use this section to complete the authorship loop.

This new option is displayed when you edit links on your profile. The one major change here is that you’ll no longer see the ‘This page is specifically about me’ box.

Link to your author page if you’re using the three-link method, particularly for work on a multi-author site or blog. Link to the home page if you’re using the two-link method which I haven’t described here (yet). There is some indication that a home page link here might work for either method, but I’d err on the side of caution until that is confirmed.

How To Check Your Work

The last step is checking your work. To see if you’ve done everything correctly, run a sample blog post or article through the rich snippet testing tool. You can use my time saving Rich Snippets Testing Tool Bookmarklet for this task.

In my first attempt, I kept getting “verified = Author link is not verified.”

What did that mean?

Was Google just waiting to verify the links? No. It meant I’d screwed up the implementation. This is what you want to see instead.

If this isn’t what you’re seeing go back and check your work again. A missed quote or placing it outside of the link element could be the culprit. In addition, you will not see the author image in the tool. Nor will you see the image immediately in search results. That might be disconcerting but it’s expected and nothing to worry about.

[Update] There’s been some chatter about whether you need to submit your site via Authorship Request Form or Rich Snippets Interest Form to enable rel=”author”.

The answer is NO. Here’s what Google Webmaster Trends Analyst John Mueller confirmed.

We will pick up the authorship information automatically as we recrawl and reindex the pages involved, but this can take a bit of time until it’s visible. You do not need to submit either of these forms for authorship information. That said, the form linked from the help center article is useful to fill out, since it gives our team a contact person on your side should we notice something amiss with the markup on your side.

[Update] Google Product Manager Sagar Kamdar reports that there is a bug in the Rich Snippet Testing Tool for those trying to verify authorship mark-up, particularly for those using the alternate two-link method. This method can be used for blogs with one author and entails a rel=”author” link from the home page to the Google Profile with a rel=”me” link from the Google Profile back to the home page.

The bug means that you’ll get a negative response even when you’ve set it up properly. Google is working quickly to fix this bug, hoping to have it deployed by next week.

I’ll update this post once more when I know how long it takes between implementation and having the profile image displayed in search results.

[Update] On July 7th, about one week after I implemented rel=”author” on this blog, my smiling mug is being displayed on Google search results. (Thanks to the JoshMeister for the heads up.)

The week between implementation and display is only for those URLs that have been recrawled recently. So this post, and my recent Google+ review both have rel=”author” display but an older post on SEO and UX does not. To be clear, it’s not about the age of the URL, it’s simply what Google is crawling again. I have older posts, such as my Facebook SEO post that do have the rel=”author” in place.

So your mileage may vary depending on how often your site and individual URLs are crawled.

[Update 12/15/2011] You can now also check Author stats in Google Webmaster Tools to see statistics for pages for which you are the verified author.

That’s how I implemented rel=author. Let me know if it works for you and if you’ve found other ways to implement it on WordPress or other platforms like Blogger.

Did Google put you in Panda Jail?

Image credit: Alex Pilchin

Many of the hundreds of Panda blog posts contain theories and advice on how to get out of Panda Jail. I’m going to review some of the more popular recommendations to show you why they’re both completely right but absolutely pointless at the same time.

Noindex Low Quality Content

Remove the cancerous content and you’ll escape from Panda Jail, right? Heck, even Google suggests you noindex duplicate and thin content pages.

The process is pretty straight forward. Look for the offending content using certain metrics. Isolate and noindex the pages on your site which have a high bounce rate, a high exit rate (careful with that one) or haven’t received any search traffic over a long period of time.

This isn’t a bad thing to do, but it won’t get you out of Panda Jail.

Instead: Grade your content corpus like your high school English teacher would.

Lower Your Bounce Rate

One of the more popular theories is that it’s all about bounce rate or pogosticking. In essence, reducing your bounce rate is a signal of user satisfaction. Sounds good right? But it’s really easy to artificially reduce bounce rate through some clever user interaction design.

In addition, bounce rate is often not a signal of user satisfaction. The bounce rate for a Q&A site is going to be very different from an eCommerce site. Do you believe that Google measures all sites using the same benchmark? Not a chance.

The goal isn’t to lower your bounce rate but is to increase user satisfaction. I can easily see a situation where user satisfaction would go up, but so would bounce rate.

Instead: Make sure your pages match query intent with relevance and value.

Lower Your Ad to Text Ratio

I’m sure many of you have fired up the Google Browser Size Tool and applied it against your website. The idea is that Panda is tripped if there is more advertising then content on pages. Google seems to understand what is content and what is advertising or chrome (e.g. – navigation and masthead).

It should be a red flag if the ads on a page actually make it difficult to read the content. But focusing on the actual percentage and trying to figure out when you cross some magical algorithmic line will be a waste of time.

Instead: Make sure your site passes some basic usability and readability tests.

Fix Your Link Profile

Do you have a lot of links from low quality sites or sites in Panda Jail? Some believe that an over abundance of links from these sites could put you in Panda Jail by association. It’s more likely that those links simply got neutralized and don’t pass as much trust and authority as they used to.

Trying to shape your anchor text or removing yourself from bad neighborhoods won’t do you much good. Mind you, make sure you’re not in a ring of porn sites but overall this isn’t why you’re in Panda Jail. In fact, where you link out to is far more important.

Instead: Grow your links organically by building your reputation and expertise.

How To Get Out of Panda Jail

Too many people are putting the horse before the cart and focusing on the tree and not the forest.

Image credit: Tom Stanley Janca

I love numbers and metrics but getting out of Panda Jail is not about optimizing for each specific metric. The lock on Panda Jail isn’t picked though a combination of simple numbers.

Instead, it’s about changing those numbers by understanding query intent and matching it with relevance and value. It’s about evaluating your site for usability and readability. It’s about delivering quality content which should not be confused with keyword matched content or a lot of content.

Getting out of Panda Jail requires you to understand mental models, information architecture, user experience, interaction design and conversion rate optimization.

The numbers will change, but they’ll change for the right reasons. And that just might get you paroled from Panda Jail.

What do Matt Cutts and Black Hat SEO have in common?

They share a C Block IP address.

OMG! Is this bad or dangerous? Here’s a fully non-technical explanation of C Block IP addresses and why you should or shouldn’t care about them.

What is an IP address?

An Internet Protocol (IP) address is essentially the computer address of a machine, in this case your site. It’s expressed as a series of four numbers (each between 0 and 255) separated by dots.

It might look something like 64.62.209.175

This is where you live. It’s home. Your little corner of the Internet.

What’s a C Block IP address?

You might get some conflicting results if you search for information on C Block IP because it often is confused with C Class IP addresses. For our purposes, the IP address is split into lettered sections (or blocks).

AAA.BBB.CCC.DDD

64.62.209.175 is in the same C Block as 64.62.209.10

The C Block is your neighborhood and those in the same C Block are your neighbors.

C Class IP addresses is a reference to the assignment of IP addresses. I’m not going to explain it further here because it just confuses the issue at hand. (If interested, here’s some basic information on IP Classes.)

C Block Links

The question that often comes up is whether links from the same C Block are some sort of red flag. For the most part, links from the same C Block aren’t a problem. It’s okay to get links from your neighbors.

But if you’re only getting links from your neighbors, things start to look a little fishy. Google might suspect that those neighbors aren’t entirely legitimate. A nefarious type might set up a slew of domains and have them all link to each other.

Links from 50 domains but just one C Block will look very strange. Links from 50 domains from 45 C Blocks will look just fine.

Bad C Blocks

Sometimes a bad element can move into your neighborhood. If you’re using a shared hosting provider you’re essentially living in an apartment building. Other people live at that address. If every one of them except for you is a notorious porn site, it might not look so great.

This could even happen on a larger level where an entire C Block has been used for some unsavory purpose. In this instance it’s not just one building that’s dilapidated, it’s the entire neighborhood.

Here’s what Matt Cutts had to say about it recently.

Check Your Neighborhood

Use the Majestic SEO Neighbourhood Checker if you’re moving into a new neighborhood or if you’re just interested in checking up on your neighbors. That’s how I stumbled upon this SEO C Block.

Sure enough mattcutts.com shares a C Block with seoblackhat.com, home of QuadsZilla. C Block neighbors also include Sphinn, Search Engine Land, Search Engine Journal, Wolf-Howl, Search Marketing Expo and Daggle.

None of these sites are getting dinged by association or even by the fact that they link to each other quite frequently.

TL;DR

An IP address is your home. The C Block is your neighborhood. C Block IP is really only important for SEO if you find out your IP is in a C Block slum.