Machine-learning results are going the way of desktop computers, where users created search queries they thought robots could understand, says search quality senior strategist Lipattsev in a Google Hangout on Thursday with top SEOs.
As a softened way of calling current SEO outmoded, Lipattsev calls those assembled for the hangout “advanced users” of Google’s search engine.
“You don’t generally type a question to Google as you would ask a real question,” Lipattsev says.
In natural language, consumers will use words like “without” and question marks. When typing in a query on a desktop computer, people tend to input what a machine would understand.
So natural language results are the change, he says.
“More and more so, people are interacting with their devices using voice,” Lipattsev says.
This puts into context how Google itself is changing as a result of consumers’ shift to mobile devices. As background for the next bit of information, consider that Google just got rid of most right-rail ads and is redesigning the AdWords interface to be far more graphic, so even B-to-B users can get the gist of their performances, at-a-glance.
“We rebuilt it several years ago for a desktop world — smartphones were only [a] year old,” Search Engine Land quotes of Paul Feng, AdWords product management director, on Monday. “Now we are in probably the biggest shift since AdWords was introduced (and I’d argue perhaps ever) with mobile. … And there is now increased demand on marketers and on AdWords as a platform — advertisers are running ads in search, display, shopping, mobile, video. Ultimately, that’s why we’re re-imagining AdWords.”
So the fact is, most articles summarizing Lipattsev’s hangout may be missing the point. Those authors are excited that Google, which may have been known for the secrecy surrounding its search engine algorithm before, isn’t so much lately. They exclaim about Lipattsev revealing that RankBrain, content and links into sites are the top three ingredients in Google’s ranking recipe.
RankBrain, the artificial intelligence that uses machine learning to help rank items on Google, isn’t always the top factor of the three, and is sometimes not even in the top three, say the articles summarizing Lipattsev’s hangout.
But what they seem to be ignoring is how much Lipattsev says RankBrain is being found lacking.
He says in addition to RankBrain, what marketers need to consider includes “content and links going into your site,” Lipattsev says in the hangout housed in Friday’s article on Search Engine Journal. “There is no order. … And the third place is a hotly contested issue.”
That’s when he got into the natural language explanation.
Meanwhile, the articles linked here are talking about how surprising it is that Google doesn’t use click data for ranking, when their studies show “pogosticking” — or what Search Engine Roundtable calls hopping onto a site and straight back to results pages — absolutely does matter to ranking.
Perhaps Lipattsev is seeing the future, where he envisions natural language will only reveal relevant results that negate the need for pogosticking.
What do marketers think?
Please respond in the comments section below.
Related story: Google’s Penguin Xmas SEO Gift