#Pubcon Liveblog: SEO Beyond the Filter Bubble

Strategic Considerations for Getting Past Google’s Personalized Results

Chris Boggs @boggles, Internet Marketing Ninjas

What Is the Filter Bubble? The filter bubble is users getting less exposure to conflicting view points to where they become isolated intellectually in their own information bubble. Eli Pariser an Internet activist defined it this way. Some people feel the filter bubble is negative because of potential manipulation of the kind of info a user may see. We can’t know what we’re missing because of the filter bubble. But we’re agreeing to that when we use Google who is in their own mind returning the best set of results to us.

Disclaimer: Content is still king. Google isn’t always going to get it right or have the inventory to deliver a result the user wants.

There are some areas we can reinvent how we’re looking at SEO that optimizes for the filter bubble.

1. Brand Search Implications

Notice that there are phone numbers and sitelinks on the SERP when Chris is signed in because Google knows he’s been interacting with the brand lately.

Consideration 1: Sitelinks — Sitelinks are influenced slightly by past behavior, but for many brand searches tested, the 6-8 remain the same except for positioning
Consideration 2: Search history — Hummingbird driven results maybe cause changes in the way that Sitelinks show up for longer-tail brand searches. If you search “seomoz algorithm updates” and then “moz algorithm updates” and back, the displayed Sitelinks differ. Brand is the primary thing to get full control over if you want to ensure the greatest likelihood of exposure beyond people’s personalized search filter bubble. (Updated from Chris’s comment below)
Consideration 3: Longer tail brand search differences. For most longer tail brand searches, little to no difference was observed
Consideration 4: Visit history: He visits Yahoo Fantasy Sports daily and personalized return differs dramatically from non-filtered
Consideration 5 and beyond: Bing and Facebook and the rest

Summary of considerations for brand searches:

2. Non-Brand Searches

He feels there’s more personalization happening for non-brand searches. We also see a shift in ads for a logged in user vs. non-logged in (perhaps a different Quality Score calculated for a logged-in user).

Consideration 1: Universal SERP layout can be dramatically different (position of Shopping results, inline Sitelinks, compressed Amazon listing).
Consideration 2: For many long-tail terms, what you see is what you get.
Consideration 2a: Conversely, for high volume head terms there is greater volatility in results
Consideration 3: Informational vs. Transactional search results tend to trend from informational to transactional as the searcher behavior/personalization comes into play, that is later in a search series

3. Local

On the left the search is done in San Francisco and on the right the search is done in New Orleans.

Consideration 1: how many blue link search result spots are available or is it filled with a local 7 pack or brand pack
Consideration 2: is there a dominant site for local terms?
Consideration 3: Sometimes a site will dominate for [brand + keyword] terms and may be better if not investing heavily
Consideration 4: Is your local listing accurate? This can be problematic for some
Consideration 5: Optimize for non-brand. Make sure your brand is associated with what you sell and Google will do the rest

Takeaways for Avoiding the Personalization Filter:

How to Get Past Google’s Personalization Layer … Into It’s Core Scoring Algorithms

Aaron Kronis @KRONis, SEO Engine

In the last year we’ve lost a lot of the keyword-driven data we used to get from Google. He quotes Pariser again and his points that you don’t know what you’re being filtered on. Google’s black box has been growing. Edward Snowden was praising Google for turning on SSL for all searches. But that same change took data away from us.

Spam filters: 1999-2004
Local filters: 2004-2011
Social Filters
Personalization: 2011-2013

How do we take back control? We need to take a closer look into what makes up the black box. Look at the inbound link flow coming off your home page.

Instead of looking at the query layer, look at the scoring layer. As the query layer grows so doe noise and confusion.

How do I get to the scoring layer? You need a functional modern search engine to open up its algos, i.e., you need a Google simulator (without all the personalization/query layer). Ask him about the Google simulator he’s built and using and seeing 50% higher ROI on SEO.

Virginia Nussey is the director of content marketing at MobileMonkey. Prior to joining this startup in 2018, Virginia was the operations and content manager at Bruce Clay Inc., having joined the company in 2008 as a writer and blogger.

See Virginia's author page for links to connect on social media.