Welcome to WebmasterWorld Guest from 54.166.133.84

Forum Moderators: Robert Charlton & goodroi

Message Too Old, No Replies

Google RankBrain "involved in every query"

     
4:26 pm on Jun 27, 2016 (gmt 0)

Administrator from GB 

WebmasterWorld Administrator engine is a WebmasterWorld Top Contributor of All Time 10+ Year Member Top Contributors Of The Month Best Post Of The Month

joined:May 9, 2000
posts:25900
votes: 871


In a recent interview with backchannel, Google's Jeff Dean says that RankBrain "involved in every query," and said, "probably not in every query but in a lot of queries."

Think about that for a moment, and consider how machine learning may be significantly affecting the SERPs and your search traffic. I suspect it may have been ramped up over time from when RankBrain was first announced. [webmasterworld.com] We know it has no effect on crawling, or indexing. [webmasterworld.com...]

Google is characteristically fuzzy on exactly how it improves search (something to do with the long tail? Better interpretation of ambiguous requests?) but Dean says that RankBrain is “involved in every query,” and affects the actual rankings “probably not in every query but in a lot of queries.” What’s more, it’s hugely effective. Of the hundreds of “signals” Google search uses when it calculates its rankings (a signal might be the user’s geographical location, or whether the headline on a page matches the text in the query), RankBrain is now rated as the third most useful. Google RankBrain "involved in every query" [backchannel.com]
5:37 pm on June 27, 2016 (gmt 0)

Senior Member

WebmasterWorld Senior Member aristotle is a WebmasterWorld Top Contributor of All Time 10+ Year Member Top Contributors Of The Month

joined:Aug 4, 2008
posts:3545
votes: 327


Google RankBrain "involved in every query

I already suspected this. There was some early statement about it only being involved in queries that Google has never seen before, which it claimed is 15% of all queries. That never made sense to me.
8:01 pm on June 28, 2016 (gmt 0)

Senior Member from US 

WebmasterWorld Senior Member keyplyr is a WebmasterWorld Top Contributor of All Time 10+ Year Member Top Contributors Of The Month

joined:Sept 26, 2001
posts:12913
votes: 891


RankBrain
I seem to remember being called that as a schoolboy.
2:39 am on June 30, 2016 (gmt 0)

Senior Member

WebmasterWorld Senior Member 10+ Year Member

joined:June 15, 2003
posts:2412
votes: 5


So, engineers at Google invent a new toy, and SEO is dead? I think not...

I don't think Google-engineers think so either... I heard somewhere that they want us to design chat bots now... perhaps their AI is a bit lonely... "Brain the size of a planet", and all that... it may need some other "brains" to talk to and hang out with? :)

Sorry, it's late here. Or, early :)
10:37 am on June 30, 2016 (gmt 0)

Junior Member

Top Contributors Of The Month

joined:July 15, 2015
posts:117
votes: 43


"or whether the headline on a page matches the text in the query"

Wow absolute genius! I bet in the 90s nobody would have even imagined something as genius as this.
1:16 pm on June 30, 2016 (gmt 0)

Preferred Member

5+ Year Member Top Contributors Of The Month

joined:June 26, 2013
posts:454
votes: 69


RankBrain being involved in many queries may be why some of us are seeing a concentration of impossible to convert traffic from Google, though I'm still skeptical of this theory. I recall seeing that RankBrain does get taken offline periodically for learning. Maybe when RankBrain is offline is the reason why some of us see converting traffic restored temporarily for a day or so?

Does anyone know where RankBrain gets the data it is supposed to learn from? Possibly data points from Chrome, Analytics accounts, embedded Adsense code or other Google services? The reason why I am asking this is because without refined user data I don't think RankBrain can be very effective at learning. If RankBrain does not have all the info to learn, it may be years (if ever) before normal traffic patterns return to some of us from Google.
4:13 pm on June 30, 2016 (gmt 0)

Senior Member

WebmasterWorld Senior Member 5+ Year Member Top Contributors Of The Month

joined:Aug 5, 2009
posts:1682
votes: 337


This might explain the "wild goose chase" searches. I look for something specific and the word is nowhere to be found on the page. Rinse. Repeat. I guess I can pin this on rankbrain. It's still stupid?
6:46 pm on June 30, 2016 (gmt 0)

Senior Member

WebmasterWorld Senior Member aristotle is a WebmasterWorld Top Contributor of All Time 10+ Year Member Top Contributors Of The Month

joined:Aug 4, 2008
posts:3545
votes: 327


Maybe RankBrain is having trouble recognizing the difference between informational sites and ecommerce sites. If so, it could be sending a lot of information seekers to ecommerce sites. In other words, mis-matched traffic.
8:11 pm on June 30, 2016 (gmt 0)

Preferred Member

5+ Year Member Top Contributors Of The Month

joined:June 26, 2013
posts:454
votes: 69


Aristotle, you are correct that sending information seekers to ecommerce sites might lead to mis-matched traffic. But RankBrain is supposed to be taken offline to learn. Where is Google getting this data from to teach RankBrain? I see many one page quick bounces from Analytics and find it hard to believe RankBrain would not have learned something over three quarters of a year. If the heavy amount of mis-matched traffic is because of RankBrain, I'm thinking Google is not feeding it enough or possibly the wrong data. But as an ecommerce store owner, I'd sure like to find out if there is anything I can do from my side to feed RankBrain the data I am seeing because it is not pretty. Maybe it would compel RankBrain to make adjustments?
9:48 pm on June 30, 2016 (gmt 0)

Senior Member from US 

WebmasterWorld Senior Member keyplyr is a WebmasterWorld Top Contributor of All Time 10+ Year Member Top Contributors Of The Month

joined:Sept 26, 2001
posts:12913
votes: 891


I'd sure like to find out if there is anything I can do from my side to feed RankBrain the data
Try structured data to identify/create rich rich snippets & rich cards, explained in the Google Search Console.
12:44 am on July 1, 2016 (gmt 0)

Senior Member from GB 

WebmasterWorld Senior Member 10+ Year Member Top Contributors Of The Month

joined:Apr 29, 2005
posts:2102
votes: 118


Try structured data to identify/create rich rich snippets & rich cards, explained in the Google Search Console.


That appears to be very sound advice.

Do you have any particular reasons or background information sources that lead you to that conclusion?
1:35 am on July 1, 2016 (gmt 0)

Senior Member from US 

WebmasterWorld Senior Member keyplyr is a WebmasterWorld Top Contributor of All Time 10+ Year Member Top Contributors Of The Month

joined:Sept 26, 2001
posts:12913
votes: 891


As said, the info is all in GSC. Not saying that is the only approach, but certainly it is one Google recommends.
6:02 am on July 7, 2016 (gmt 0)

Preferred Member from BG 

Top Contributors Of The Month

joined:Aug 11, 2014
posts:546
votes: 173


I have not observed RankBrain activity for generic one keyword queries. In huge money keywords like "gold" you will often see a crappy landing page with not text and only two graphs filled with "buy gold" menu pages and "trick banners" . The situation gets worse when you go outside the .com and US queries and go like .de or .fr or any other ccTLD.

I have seen shake-ups in the long tail money queries though, mainly due to personalized search and news feed pushing the organic links down, shrinking the traffic.

Long story short if you want to do SEO in the "rank brain era" you need to make sure you can funnel as much traffic (paid or otherwise) and remarket them. The more times people visit your site one way or the other, the more your site will appear in the "rank brained" query.

Yes, you thought right. Adwords does make your organic traffic higher when rank brain is involved. Welcome to Google Paid Search.
7:37 pm on July 7, 2016 (gmt 0)

Junior Member from US 

10+ Year Member

joined:Jan 11, 2007
posts:152
votes: 1


@Nutterum - i can echo that news feeds have been shaking up some searches i monitor. They usually last a couple weeks before the query goes back to normal.

I think we can all agree that it would be unethical to see correlation to rankings and adwords spend.
6:19 am on July 8, 2016 (gmt 0)

Preferred Member from BG 

Top Contributors Of The Month

joined:Aug 11, 2014
posts:546
votes: 173


I am not saying I like it. I am saying only what I observe. Rank Brain will show up your website in the results page the more your visitors interact with that same page, regardless of how. Do you think Google auto-magically filters users from organic search when you funnel traffic from FB paid ads to your website and then remarket them with adwords? No, what Rank Brain is seing is people going from social media to your website and if they interact well enough (as in low bounce rate, average or above average time spent on site, etc, etc) they will deem these users as interested in your service and will place your website in their personalized search bubbles. It's how things work, plain and simple.

This year I have observed three different companies with more than 50k digital advertisement budget (per month) do the funnel-remarket-organic-re-engage limbo on their viewers to great success.

Now I need to state the obvious that with crappy website, poor service, low budget, poor UX, poor demand generation strategy or combination of these factors can and will lead to different (read worse) results. Yet, it was blatantly obvious that adwords remarketing, doubleclick banners on big websites and FB ads led to substantial increase in authority and organic search. And the correlation was visible with the increase or decrease of the demand generation budget.

Do not worry though, unless your business has big lump sum of cash to spend on advertising you will not feel these effects, nor will your competitors. Does not mean the "authority websites" won't eat your top SERP spots should they decided to move in your niche.
7:02 pm on July 8, 2016 (gmt 0)

Moderator from US 

WebmasterWorld Administrator martinibuster is a WebmasterWorld Top Contributor of All Time 10+ Year Member Top Contributors Of The Month

joined:Apr 13, 2002
posts:14886
votes: 484


what Rank Brain is seing is people going from social media to your website


Google does not "see" traffic from Facebook to your site, even if you're using Google Analytics. Google Analytics data is not used for ranking purposes, it is only used in the aggregate.

Rank Brain will show up your website in the results page the more your visitors interact with that same page, regardless of how.


That's not how CTR log data is used. Google does not evaluate every single user query user experience and adjust rankings in real-time.
Every kind of search query exhibits different click through patterns. Many queries have multiple user intent. For Example Query A there may be three user intents. This means the top ten of the SERPs must be divided between three user intents. These intents are ordered according to the popularity of the intent, i.e. what most of the users want to see. This is the part that screws with people's heads because they see a site at the top that the site publisher feels doesn't answer the question while there's is at position 6. The difference is that most people prefer the user intent of position 1 over that in position 6.

User interaction data within click logs is used to study and understand user intent. Here is the part that may be new to you: Search engines can predict what user interaction may be based on that past data it studied, it can re-rank a site from position 15 and place it in position 5 because it can predict a better user experience.

However, Google takes into consideration your previous searches and can re-rank the SERPs so that it matches your user intent.

Site interaction/UX is not a ranking factor. I wrote a 4,000 word article about UX on SearchEngineJournal [searchenginejournal.com], you may wish to review that because it has links to authoritative sources of information on the topic of user interactions and how that affects rankings.
7:13 am on July 11, 2016 (gmt 0)

Preferred Member from BG 

Top Contributors Of The Month

joined:Aug 11, 2014
posts:546
votes: 173


@Martinibuster,

I am well aware of your article and I do read most of your content as it is quite interesting and insightful. That being said, what Google say and what Google does are two different things. With the correct approach, demand generation and PR (and budget!) you can create searches that will supplement a money-related topic you want to rank your website for. I have seen quite interesting shifts in users behavior and background (using demand base data linked to my GAP account) pre and post a big campaign. Trust me when I tell you that those social interactions and social shares and content distributions do play a good part at pushing the not-in-the-focus landing pages targeting money queries like "non branded product demo". Now I do not pretend to know what is under the hood of the Google Algo. As I mentioned in my post, I just observe how making a big campaign with lots of social media and remarketing advertisment via Double Click, increase the rankings for the money keyword landing pages and as soon as the campaign finished the ranking dropped to their previous positions.
3:52 pm on July 11, 2016 (gmt 0)

Moderator from US 

WebmasterWorld Administrator martinibuster is a WebmasterWorld Top Contributor of All Time 10+ Year Member Top Contributors Of The Month

joined:Apr 13, 2002
posts:14886
votes: 484


If your rankings are temporary and stops when social media attention drops then you may wish to consider the possibility that it's a trending topic/current event boost, a boost created for documents matching a trending topic. That's not rank brain, afaik. I don't read every rank brain article out there because most of them are speculative or guesswork. But I would be most interested in a link to anything authoritative that explicitly says RankBrain looks at social media shares and gives boosts to trending topics. But that's not how I understand rankbrain to work.

This is however an interesting direction for study, and not just limited to social media, too!

Good luck,
;)

mb
6:10 am on July 12, 2016 (gmt 0)

Preferred Member from BG 

Top Contributors Of The Month

joined:Aug 11, 2014
posts:546
votes: 173


Maybe you are correct and this has nothing to do with Rankbrain, but I can't shake the feeling that the the more people "stumble" upon a campaign, the more they maybe will start searching for details, making a Google trend and ultimately falling in a "rankbrained" personalized search bubble. Even if the connection is not that direct, I have seen it working pretty good, which is bad, because it would mean that with enough money to burn on a campaign, you can force yourself in to peoples search results, something that is against Googles policy afaik.

Still though, thank you for the discussion, interesting topic to delve into for sure.

Best,

Ntt