joi, 14 noiembrie 2013

Future SERP: A Glimpse at Google 2014

Future SERP: A Glimpse at Google 2014

Future SERP: A Glimpse at Google 2014

Posted: 13 Nov 2013 03:10 PM PST

Posted by Dr-Pete

Watching Google change can quickly become an obsession, and it's easy to jump at every shadow when they test thousands of ideas per year (and roll out hundreds). This post is an attempt to take all of the things I've seen in the past six months and tell a story driven by real data. This is the story of how I think Google will look by the end of 2014, and what that implies about their direction and core philosophy.

Two data sources

(1) MozCast Feature Alert

In April of 2012, I launched "Project Algo Alert", a prototype that would later become MozCast. What was originally one "weather" station, designed to measure daily fluctuations in top 10 rankings across 1000 keywords, has evolved into 11 stations and three unique systems. One of those systems is Feature Alert, which was based on a simple idea â€" how could we detect when Google launched new SERP features, without any prior knowledge of what those features would be?

Feature Alert solves this problem by cataloging the basic building blocks of Google's source code, the container names and IDs in CSS. Let's say for example, that Feature Alert sees the following chunk of HTML/CSS code:

The system checks each building block against an archive, and if "ads-container c mnr-c" is a new object, it's captured and I'm alerted that something new happened. When I built Feature Alert, I thought something new might pop up a couple of times a month. As of writing this post, the system has captured 2,441 unique building blocks.

A side effect of the system is that, at large scale, it frequently catches Google in the act of testing new features and UI changes. Keep in mind that Google ran 7,018 "live traffic experiments" in 2012 â€" while we probably capture only a small number of them, these tests allow us to get a glimpse into what's coming next. While any given change may be rejected (Google launched just over 9% of the changes they tested last year), some changes appear repeatedly in testing and in different formats over time, strongly suggesting that Google is intent on launch.

(2) Mobile feature launches

Google is terrified of mobile â€" the ad landscape that drove 84% of Google's revenue in Q3 is a completely different animal on mobile, and consumer behavior is evolving rapidly. One clear pattern in 2013 is that many major UI changes hit mobile before they hit desktop. Google is designing for tomorrow's devices and is desperate to make sure that ad CTR and CPC don't fall as mobile search volume increases and new devices (like Google Glass) come onto the scene.

When we see a new feature in testing and then realize it already exists on mobile, odds are good that that change is coming to desktop soon. By combining these two data sources, we've been able to paint a picture of Google's near future. Based on the past few months, I'm going to make six predictions for 2014 and turn those six predictions into two conceptual screenshots.

Six predictions

For each of the six predictions below, I'll provide evidence from MozCast and/or mobile search, along with my confidence in the prediction. These predictions are grouped to tell a story, but are otherwise in no particular order:

(1) New Knowledge Graph â€" 98%

Since its launch, the vast majority of Google's Knowledge Graph has been built on a very few data sources (including Wikipedia, Freebase, and the CIA Factbook). The core problem is that these sources are limited and only work well for highly structured data. To expand, Google needs to extract answers from their entire index of the web. Put simply, Google needs to be able to create answers from content. Over the past few months, we've seen extensive testing of Knowledge Graph entries like this one:

Notice the "In context" section, which is the bulk of the informational content â€" this is entirely driven by third-party websites. All of the blue links are links to additional Google searches, but the light-gray links show the original sources. Put simply, Google will soon be building their Knowledge Graph on your data.

It's interesting to note that the queries we're seeing this on seem to be fairly broad and/or have a generic intent. When Google launched "in-depth articles", they made the following statement:

To understand a broad topic, sometimes you need more than a quick answer. Our research indicates perhaps 10% of people's daily information needs fit this category.

It's very likely that this new Knowledge Graph approach is an attempt to solve the same problem, and that these entries will appear on searches that previously had no Knowledge Graph data. Long story short, this isn't just a change â€" it's an expansion.

Knowledge Graph drives many variations of answer boxes, and so it's not surprising that we're also seeing new answer boxes in testing. This answer box was captured from a search for "is pneumonia contagious":

Unlike traditional answer boxes, this information is being extracted from the index (in this case, kidshealth.org) and treated more like an organic search result. Given that we've seen multiple variations of new answer boxes and multiple tests of these new Knowledge Graph entries, my confidence is very high that some version of these features will roll out in the next few months.

(2) Revamped advertising format â€" 95%

Recently, we've seen Google aggressively testing a new and somewhat surprising advertising format (and outside sources have confirmed these tests). It looks something like this (the search was "paella recipe"):

I've added the (...) in the middle section, which is more organic results, but the divider marks the top and bottom AdWords blocks. Essentially, Google has added the marker [Ad] to each advertisement, but they've removed the current background color and have formatted everything else to be nearly identical to an organic listing.

It may seem surprising that Google would visibly mark ads in this way. I suspect that this isn't entirely by choice, but is related to Google's ongoing battle with regulators and their pending settlement with the European Union. If this move seems unlikely to you, consider a second piece of evidence. This is a current search for "paella pans" via my mobile phone (all mobile screenshots come from Safari/iOS7 on an iPhone 5S):

This new format has been running on mobile browsers for a while now, and Google's widespread testing makes it look like a foregone conclusion for desktop search. This change will have huge implications on both organic and paid CTR in 2014, regardless of the final form. Expect Google to also test and iterate quickly when the new ad format launches.

(3) Ads outside of three blocks â€" 33%

This prediction is much more speculative and I have no clear evidence to support it, but the potential impact is big enough that I'm going to say it out loud. Once Google is individually labeling ads in the left-hand column (right-hand column ads only get one [Ads] marker at the top, at least in testing), ads will become stand-alone units. In other words, Google will no longer be constrained by fixed blocks at the top and bottom. So, what's to keep the test above from turning into something more like this (the next image is conceptual, not a captured test):

Individual ads could be interspersed in organic results, impacting the overall effectiveness of any given position in those results. Once Google has the flexibility to move ads, I see no compelling reason to believe that they won't test new options to improve ad effectiveness. I'll conservatively put the odds of this change at one in three.

(4) Loss of result count/stats â€" 80%

This one has serious implications for SEOs, but I think it's a move that makes sense for Google. Let's look at the entire screenshot for the "paella recipe" search we dug into previously:

Notice something missing? There's no light-gray result count at the top of this page ("About 3,270,000 results…"). Google has entirely removed that line of text in this test screen. Truthfully, as much as we rely on these numbers for SEO research (especially with search operators like "site:"), I suspect the additional data has almost no value for everyday search users. It's taking up prime real estate, and Google could very likely get rid of it.

Scroll back up to the mobile search for "paella pans" and you'll see that result count data is already gone from mobile. On a mobile phone, that data simply takes up too much valuable space. It's possible that Google could preserve the data for operators and certain searches, but I have no clear evidence either way. If you're a search marketer, I would be prepared to lose this data in 2014.

(5) Boxed design for #1 result â€" 90%

The current incarnation of an expanded #1 organic result has been around for a while â€" it has an indented set of site-links (up to six, on a normal search, or ten on domain searches) that each have links and short snippets. Google has been testing a number of variations on boxed designs for expanded #1 results, such as the following:

Notice that the entire entry is boxes, as are the individual site-links. The main link is in a larger-than-normal font, and some of the site-links have arrows that pull up related links. Google has been testing many variations on this theme for a couple of months now, but consider that one variation already exists in mobile search (this is a search for our own brand, as Ra Sushi generates local results on mobile):

While the mobile result is constrained to a single column, the overall listing is boxed, with clear dividers between the main result and site-links. Google has been testing variations on this one for a while, and they seem to be worried about getting it right, but by the end of 2014, I'm almost certain that some variation on boxed results for the #1 organic position will launch.

(6) Boxed design for entire page â€" 50%

It's easy to assume that this boxed design is purely cosmetic, but I believe it goes much deeper than that. Consider the look of another Google product, Google Now (via my iPhone 5S):

Google Now is divided into what Google calls "cards", distinct units of information that are individually boxed and can be mixed, matched, and sorted as Google sees fit. Notice how similar this format looks to Google's mobile search results. As Google expands into new formats, including wearable technology like Google Glass, and as screen sizes diverge â€" from phones to tablets to desktop and everything in between â€" it's going to be increasingly important that they can escape the constraints of a single, fixed-format display. Cards are a natural transition to a flexible and dynamic SERP, allowing Google to mix and match depending on the device you're using.

Google Now has already started appearing in personalized search results. Consider the following Knowledge Graph entry, produced when I do a brand search for "amazon" while being logged into my Google account:

Google has inserted a personalized card into the Knowledge Graph entry that indicates I have recent orders. Clicking on details produces an answer box containing yet more personalized information. Even the Knowledge Graph entry itself and current answer boxes are card-like, with clear outlines and separation from organic results. Answer boxes are tailor-made for mobile search, and so are Google Now cards.

So, what if Google took this idea to its logical conclusion and created an entire SERP that was divided into individual card-like units? This is how mobile search already looks, as you can see from multiple examples above. This would be a big change for desktop, and I have not seen an entirely card-based SERP in testing, but I'll put the odds of that development at about even (50/50).

Two SERP concepts

So, what might Google look like by the end of 2014? I've come up with two artist's conceptions (I created them, so take the phrase "artist" loosely). While some aspects of these concepts are based in reality, these are not real Google results (live or in testing). The first is based on my recent flight on Virgin America and a fictional brand search for "virgin america" (click on the image for a full-sized version):

Google 2014 SERP Concept v1

Here we have a completely boxed (card-style) SERP, with the new ad format, Google Now in the Knowledge Graph, and the redesigned #1 organic result with site-links. I've added the mobile background color and removed result count data. While I don't think Google will adopt this exact look and feel, it combines many of the data-driven predictions in this post.

Just for fun, let's look at a second variation. Here we have a Knowledge Graph result using "In context" data from 3rd-party sources, plus an answer box parsing information from a 3rd-party source. The first ad is placed after the answer box, and a second ad has been inserted after the top two organic results. Again, this is purely conjecture:

Google 2014 SERP Concept v2

While I don't expect Google to look exactly like either of these concepts by the end of 2014, I think the data strongly suggests that many of these concepts will be in play, and Google will have shifted strongly toward a more card-based design. Add in the expansion of Knowledge Graph and Google's rush to get mobile right, and I expect significant changes to SEO in the next year. The best we can do is keep our eyes open.

This post was adapted from a presentation at ThenSome San Francisco, called "Future SERP: The Face of Google in 2014" (available on SlideShare). Thanks to SEOGadget for hosting the event and to the audience for a great discussion that helped me vet and develop some of these ideas.


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!
Publicat de 1 la 09:33

Niciun comentariu:

Trimiteți un comentariu

Abonați-vă la: Postare comentarii (Atom)