The world wide web offers its audience a virtually unfiltered, unlimited supply of content. Each day, the internet grows massively, much larger than any person or group of persons could consume:

• 60,000 new web sites
• 2 million videos on  Youtube
• 1.6 million Blog Posts
• 1.5 Facebook status updates
• 10 Million Tumbler Posts
• 5 Million photos on Flickr
• 140 million tweets

- Contently, 2011

Before the digital revolution, content sent to mass audiences was filtered by gatekeepers – newspaper editors, television programming departments, radio executives, and book publishers. But the internet removed these gatekeepers, and it has only been recently that the internet audience has shown the fatigue of having too much data on hand.

Enter the Search Engines. In an effort to meet consumer expectations and to help combat content fatigue, Google and other search engines began a process of personalization – using viewing history, comments, and other digital material such as likes and follows – to try and tailor search results to match a consumer’s expectations. Tied closely to this is localization – using your IP address to givegreater relevance to results closer to your geographical location.

While a great tool for commerce – limiting search results to material that matches a user’s preference – the privatization of search results had an unseen effect, called the Filter Bubble. In a TED talk in May 2011, Eli Pariser discussed the concept and what unintended consequences it can bring about:

As you can see, this can become an issue quite quickly for businesses, especially start-ups and small business that have to break through the filter bubble to reach new customers. In health care, it could prove a challenge for public awareness and health campaigns trying to reach audiences with strongly held views or communities built around a behavior or practice that is trying to be changed or influenced.

Filter Bubble and SEO

For web developers, this creates a real issue for tweaking SEO. Two individuals looking at the same search term in Google can get radically different search results, dependent on their Filter Bubble. When I look up issues related to Web Accessibility and Social Media tools, my results often pull up the blogs of friends and colleagues I have in higher education who have written on the subject.  But if someone else was to perform the same search, those blog posts would be much lower in rankings – sometimes on entirely different pages of search results.

As you can imagine, this can lead to a lot of confusion, especially when you, as the web content editor, are discussing the search engine placement of your site with an individual by email or phone. They might see a problem with the site’s ranking that you cannot; to you, your site is number one in search results, while in reality it might be a very different picture. It is essential to know this and understand the effect so you can communicate this to your clients and administrators of a department.

How do you Account for the Filter Bubble?

Google Verbatim

Last Month, Google introduced a new search tool called Verbatim. The Verbatim tool ignores the spelling auto-correcting tool, the synonyms tool, and, most importantly, your personalized filters to give you the unadulterated results of exactly what you typed. It’s the cleanest unskewed search result you can get from Google.

To use Verbatim, type in your search as you normally would using the Google home page. On the results page, select Show search tools at the bottom of the left hand column. Verbatim will be the last option on the Search tools list. Click it and the search results will be modified to remove Google’s influences on the results.

Correcting in Normal Search Results

There’s no sure-fire way to do this, but the following are practices to follow to help keep the filter bubble from skewing your results. Doing so will affect your web browsing preferences and experience, so use only those you are comfortable with.

• Clean out your cache
• Erase your web browsing history
• Clean out your cookies
• Set your Facebook privacy settings to maximum
• Don’t be signed into a Google account when making a search
• Hide your birth date on Facebook
• Opt out of targeted ads whenever possible
• Use the private browsing option available in newer browsers, or use an anonymizer for browsing the web

Eli Pariser talks about the steps of doing these methods with most modern browsers on his web site.

Accounting for Localization

You can account for localization bias in searching Google using one the country-specific Googles [i.e. use google.co.uk instead of google.com] or by changing location with the left sidebar tools to specify a different U.S. city or zip code.

Photo of Bubbles by Jeff Kubina, Flickr

Google Analytics now provides a set of reports for tracking social interaction with a website’s content. This new metric will measure how often users use social share buttons – for example, a Facebook Like, and Google +1, a LinkedIn share, or a tweet. This information can be used to analyze how your users share information and whether your sharing audiences use your website in different ways than those who do not.

Social Engagement Report

The Social Engagement Report shows site behavior changes for visits that include clicks on any social sharing actions. In the example below, we see that of 7,777 visits to this website in the last 30 days, only two visitors were socially engaged.

In the example above, we should also note that this site is only pulling in data from Google+, which is built in to the Google Analytics application. In order to measure engagement from other social media networks, it’s necessary to modify the scripts that control each of the social media engagement tools, which we’ll discuss in a moment.

Social Actions Report

The Social Actions report displays the number of social actions (+1 clicks, Tweets, Facebook Likes, etc) that have taken place on the site. This information can be important to determine which networks are more preferred by users, which can lead to making that button larger, changing its position on the screen, etc.

Social Pages

The Social Pages report shows the specific pages on your site that are driving social actions. This is very useful for content providers to determine what content on their site is of most interest to your audience, what they are most likely to share with their friends, and what has the greatest potential to become viral content.

The Tracking Code

In order to track the other social media tools, Google has created Social Interaction Tracking. This tracking has to be integrated into each social network button.

The syntax for this tracking is as follows:

_trackSocial(network, socialAction, opt_target, opt_pagePath)

Network

Name of the social network: Google, Facebook, Twitter, Digg, etc.

SocialAction

Type of action: Like, Tweet, Send, Stumble, etc.

opt_target

This optional field allows the user to manually set the Subject of the action being taken. If it is not used, this defaults to the URL of the web page being shared. If used, it can be used to name a different web address, the name of a product or an article, or the content ID of the material, for example.

opt_pagePath

This optional field can be used to change the URL where the sharing took place (for example, for listing a virtual page name instead).

More Information

• For a more technical overview on how to implement this tag for Facebook and Twitter visit the code site article.
• To see a working example of how to integrate Analytics with Facebook and Twitter buttons check out the Social Analytics sample code.

Google Analytics and Social Interaction Part 2: In the next installment of this post, we’ll discuss methods for tracking incoming links from social networks and recording these in Google Analytics.

Let’s paint a very simple picture: You’re getting ready to announce to the world the release of your new online magazine/newsletter. It’s filled with dedicated articles on several topics and you’ve decided that a simple email will be the way you announce it’s online release.

So, you draft up an email message and hit Send All to your list. Let’s pretend your email message looks something like:

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec posuere ipsum ac ante fermentum condimentum. Morbi vehicula fermentum nisl, venenatis condimentum quam auctor a. Nunc aliquet dui at massa iaculis sed ultrices justo ullamcorper. Suspendisse nulla ante, http://www.ournewsletter.com bibendum vitae auctor non, faucibus vitae libero.

 

Nam libero nibh, tincidunt quis porta id, venenatis ultrices est. Aliquam interdum diam et nunc volutpat quis condimentum nunc dictum.

 

-Magazine

…the Latin is on purpose, but you get the idea: it’s just a small paragraph or two in a narrative style inviting users to read your new newsletter.

How about instead of a couple of paragraphs, you create something that looks more like this:

It’s not a revolutionary design. In-fact it’s down right ancient. A small thumbnail, a headline that behaves as a direct link to the article, and a brief excerpt about what the article is about. However, what this little bit of extra work resulted in was impressive:

• Overall traffic was unchanged by the format change. That is, just as many users in total visited the homepage of the website from the simple style of announcement as compared to the expanded style, so there’s not much change here, however…
• Individual article readership increased by nearly 200%! This is not a typo. Each individual article’s statistics soared by simply providing a basic thumbnail and paragraph teaser to the article itself.

Here’s a rough breakdown of the old vs. new:

The data speaks for itself, but it paints a simple picture: while people like to click links in emails, they really like to be teased ahead of time to know just what it is they’re reading about.

The Real-Time traffic monitoring software doesn’t require existing users to install anything: Your existing analytics insertion codes will work just fine.

The new service will allow you to view a “Real-Time” dashboard from within the “new” version of the GA dashboard (the same version they’ve been advertising via a red link in the header for a very long time).

The timing couldn’t have been better for us; we have been evaluating several real-time traffic monitoring tools for just this purpose. Combinations of TOP, server-status, and future software yet-installed will still play a role in our ability appreciate server resources, but the Google Analytics Real-Time traffic monitoring tool will allow us to very quickly appreciate where users are hitting across our enterprise WordPress system.

Use Cases:

For our immediate purpose, the tool will allow us to appreciate when our servers are getting hit, and where these users are coming from and what content they are interested in.

The real-time information, however, could influence how and when you pull in social media channels and allow you to measure real time the traffic generated by those channels.

Imagine posting content on FB or Twitter and watching how long it takes for that traffic to make its way to your site, adjusting your message real-time as you track their behavior.

Or perhaps discover a portion of your website that suddenly became very popular, allowing you to respond to a potential crisis (or a success) while it is still relevant to your users; not the next day, after the dust has settled.

Preview Video (full screen recommended):

Improvements:

Right now the monitoring software looks pretty bare bones. My major two gripes so far are:

• Does not seem to leverage the same _gaq.push variables that the normal analytics software measures, (such as _setDomainName), but hopefully these variables will become fully integrated with time.
• Lacks the ability to follow a user through their individual navigation paths and watch how they weave in and out of your website (a feature available in the existing static analytical tool)

Final Verdict:

The feature is a big win for Google Analytics, and a welcome addition to their immensely useful [and free] tool suite.

To enable real time tracking on your Google Analytics account, go bug Google here.

About the Author

Carlos Morales
Manager

Carlos began designing websites in the mid-90s while being completely enamored by the Hot Dog HTML editing software and the seemingly endless supply of MIDI files that needed to be posted on the web. Soon the shiny and welcoming exterior of HTML wore off and Carlos was lost for several generations. Lured in by the appeal of FPS and MMORPG worlds, many nights were lost to these deviant technologies. Desperate to clean up his act, he found himself at the footsteps of UF. The kind souls brought him in and cleaned him up, providing shelter and nourishment. He hasn’t left since. He can sometimes be found wandering the halls still searching for red armor.

When he’s not busy fixing current websites, he acts as manager for the Web Services team, attempting to lead his troup of gifted technicians to the land of semantically correct code, beautiful and usable websites, and satisfied customers.

Find Carlos on Campus

• Communicore Building, Room C3-039
• 352.273.9027
• morales2@ufl.edu

Photo by Zen Sutherland, Flickr