Nov 2015 Phantom updates: User-Engagement factors + Panda

[Robert Charlton]

Back in July we discussed Glenn Gabe's observations which suggest that Phantom 2 might be an attempt to automate Panda....

Is Phantom 2 actually Panda moving to real time?
Jul 21, 2015
https://www.webmasterworld.com/google/4758405.htm [webmasterworld.com]

In the mid-July article we reference, Glenn uses the sites involved in the ongoing "tremors" he's seen to connect the earlier Phantom updates...

Documenting Important Phantom Dates...
Original Phantom 2 Rollout: 4/29/15
Phantom Tremor: 5/27/15
Phantom Tremor: 6/8/15
Phantom Tremor: 6/17/15
Phantom Tremor: 6/28/15
Phantom Tremor: 7/14/15

Glenn has now written about more recent updates, which he's observed in November...

Analysis and Findings From the November 19, 2015 Google Algorithm Update (Including Its Connection With Phantom 2)
December 2, 2015 By Glenn Gabe
[hmtweb.com...]

From this article...

...another significant unconfirmed algorithm update on November 19, 2015. And it was big. Really big.

Additionally, he noted another adjustment later in November...

Tremor on 11/28
...there is a lot of evidence that a tremor rolled out on 11/28 and many of the sites seeing impact on 11/19 saw more impact starting on 11/28. Some went up more, some fell further, while others adjusted (going up after going down or vice versa). Keep this in mind while analyzing your own trending.

He connects both of these with his earlier observations...

There were many sites surging or dropping starting on 11/19 and it was hard to overlook the connection to previous "quality updates" in 2015.....

Significantly, Glenn noticed that for all the similarities to Panda, there was a key difference...

Phantom Versus Panda – User Engagement A Key Factor in 2015 Quality Updates

I noticed many examples of content quality problems on sites pummeled by Phantom, when most people would associate those problems with Panda.

t the more sites I analyzed that were impacted by the 2/5 update, Phantom 2 in May, and the 11/19 update, the more I started to realize the slight difference between the quality updates and Panda. And it came down to user engagement.

Glen cites many specifics about the kinds of engagement problems observed. Ongoing discussion will most likely involve these. Let's be careful to avoid unfair use as we cite the article. It's a fascinating and I think an important study.

Several members here have mentioned late November changes in other threads, and I thought I'd post the above to get us started on connecting some of those dots as well.

Additionally, we've had quite a few discussions about user-experience as a factor here, and those probably ought to be tied into this one. Gary Illyes in the past has been reluctant to acknowledge UX as a ranking factor, and it may be that this will give us a different lens with which to reconcile the two. (Also, to a degree, I see a potential tie in with Zombie traffic... but maybe we should hold off on that one).

I'm looking forward to some stimulating discussion.

[webmuppet]

To be clear, there's no evidence to date of anyone with a Panda penalty recovering because of Phantom which suggests that Phantom and Panda are separate?

[MattJankulovski]

For anyone wanting a further look at the changes, Searchmetrics have published a blog which reports much the same findings as Glenn and includes examples of some sites that won/lost as a result, so you can take a look and decide for yourself just what role UX/ content may have played: [blog.searchmetrics.com...]

[doc_z]

A Phantom update aoround November 19th with similarities to Panda fits with my observations described here [webmasterworld.com]. However, I'm seeing no evidence for user-engagement factors.

[DarthVader]

My website's traffic increased a bit. I think there has really been update in past week because I can see clear changes in some of my blogs.

[Robert Charlton]

MattJankulovski, thanks for the link to the SearchMetrics article by Marcus Tober, which I think is particularly useful for the detailed content-quality comparisons of several competing pages, for a look at several other adjusted algo areas, and for the overall notes regarding the updated quality guidelines.

Here's a link to the Tober article with title, date, and author for reference....

Phantom Update III: What Google’s updated Quality Guidelines could mean for page rankings
December 2, 2015 - by Marcus Tober
[blog.searchmetrics.com...]

-----

However, I'm seeing no evidence for user-engagement factors.

doc_z, I've gone back and forth on this, not so much because I'm not seeing user-engagement factors here... but because I don't know how much of a change that is from some of the previous Panda updates. So maybe we're in semi-agreement on this update... I don't know.

As I reflect on it, user-engagement factors were built into Panda from the beginning, and "engagement" was the big word back then. What else would you call loading-speed, ad density, degree of internal content repetition, readability, backing out to the serps, etc, if not user-engagement factors? Many of my critiques in the Site Review forum in the WebmasterWorld Supporters area have been squarely focused on user engagement, in both text content and on page and interface design.

On the pages that SearchMetrics discusses, beyond text content, I'm seeing some highly nuanced design factors which are clearly design and user-engagement related. "How easy is it to use this page?" kind of things.

Glen Gabe cites factors he's seeing like Ad Deception, Excessive Pagination, Disorganized Pages, Pop-Ups, etc. These page design factors all seem consistent with where Panda started and where it's been going... beyond text content itself.

[toidi]

re the searchmetrics article, this is great news if we can get a human to review our sites who can understand the content instead of counting characters.

[martinibuster]

UX as a ranking factor

Not likely a ranking factor. It's a part of something else.

What else would you call loading-speed, ad density, degree of internal content repetition, readability, backing out to the serps, etc, if not user-engagement factors?

I prefer to think of it as a user experience factor. A slow loading page is a bad user experience. User engagement, to my mind, is about how much people enjoy a site, how often visitors recommend a site, etc. A good user experience leads to better user engagement metrics. But the two are different things.

[aristotle]

martinibuster -- But attempting to measure user experience can be tricky. For example, if a site is about a controversial social or political issues, some of the visitors could be hostile and are only looking for places to leave nasty comments.

[martinibuster]

But attempting to measure user experience can be tricky.

Right. Which is why they're very likely not ranking factors.

Now consider this: Certain factors can affect the ability of a site to appear in the SERPs yet those same factors are not ranking factors. There is no contradiction. This is a huge point of confusion with the SEO industry.

I'm pretty sure I've stated this before about these kinds of metrics but I'll restate it for you.

Various metrics used in isolation have a level of error that makes them unsuitable as standalone metrics. However when you combine metrics then the accuracy rate of predicting the probabilities rises, for example in predicting that a site is spam, that a site is useful, or that a site is authoritative.

[doc_z]

However, I'm seeing no evidence for user-engagement factors.

When I posted this I were referring to such things as bounce rate, time on site or liking pages. 

Indeed user-engagement can be seen in various ways and have more aspects. My statement wasn't about factors like design or ad density.

According to our data, the updated Quality Guidelines from Google do seem to be affecting rankings.

I agree with that SearchMetrics that the Google Quality Guidelines are a good starting point when looking for reasons.

[martinibuster]

The quality guidelines have always been a good starting point. It's been this way for years. There's nothing new here. The quality guidelines changed a little but the substance of it is the same. :)

[EditorialGuy]

Re user engagement: Google Analytics has an "Engagement" section (Audience/Behavior/Engagements) that gets very little discussion but probably should. GA's Engagement statistics show "Sessions" and "Pageviews" under two headings, "Session Duration" and "Page Depth." These make it easy to see (for example) how many people are leaving your site in 10 seconds or less, or how many of your pageviews come from sessions of 20+ pageviews.

Such data could be useful to Google not only for measuring quality, but also (when broken down by logged-in user) for personalized search results.