fbpx

Social media algorithms, artificial intelligence, and our own genetics are among the factors influencing us beyond our awareness. This raises an ancient question: do we have control over our own lives? This article is part of The Conversation’s series on the science of free will.

Have you ever watched a video or movie because YouTube or Netflix recommended it to you? Or added a friend on Facebook from the list of “people you may know”?

And how does Twitter decide which tweets to show you at the top of your feed?

These platforms are driven by algorithms, which rank and recommend content for us based on our data.

 

Hear directly from the scientists on the latest research.

As Woodrow Hartzog, a professor of law and computer science at Northeastern University, Boston, explains:

If you want to know when social media companies are trying to manipulate you into disclosing information or engaging more, the answer is always.

So if we are making decisions based on what’s shown to us by these algorithms, what does that mean for our ability to make decisions freely?

What we see is tailored for us

An algorithm is a digital recipe: a list of rules for achieving an outcome, using a set of ingredients. Usually, for tech companies, that outcome is to make money by convincing us to buy something or keeping us scrolling in order to show us more advertisements.

The ingredients used are the data we provide through our actions online – knowingly or otherwise. Every time you like a post, watch a video, or buy something, you provide data that can be used to make predictions about your next move.

These algorithms can influence us, even if we’re not aware of it. As the New York Times’ Rabbit Hole podcast explores, YouTube’s recommendation algorithms can drive viewers to increasingly extreme content, potentially leading to online radicalisation.

Facebook’s News Feed algorithm ranks content to keep us engaged on the platform. It can produce a phenomenon called “emotional contagion”, in which seeing positive posts leads us to write positive posts ourselves, and seeing negative posts means we’re more likely to craft negative posts — though this study was controversial partially because the effect sizes were small.

Also, so-called “dark patterns” are designed to trick us into sharing more, or spending more on websites like Amazon. These are tricks of website design such as hiding the unsubscribe button, or showing how many people are buying the product you’re looking at right now. They subconsciously nudge you towards actions the site would like you to take.

You are being profiled

Cambridge Analytica, the company involved in the largest known Facebook data leak to date, claimed to be able to profile your psychology based on your “likes”. These profiles could then be used to target you with political advertising.

“Cookies” are small pieces of data which track us across websites. They are records of actions you’ve taken online (such as links clicked and pages visited) that are stored in the browser. When they are combined with data from multiple sources including from large-scale hacks, this is known as “data enrichment”. It can link our personal data like email addresses to other information such as our education level.

 

These data are regularly used by tech companies like Amazon, Facebook, and others to build profiles of us and predict our future behaviour.

You are being predicted

So, how much of your behaviour can be predicted by algorithms based on your data?

Our research, published in Nature Human Behaviour last year, explored this question by looking at how much information about you is contained in the posts your friends make on social media.

Using data from Twitter, we estimated how predictable peoples’ tweets were, using only the data from their friends. We found data from eight or nine friends was enough to be able to predict someone’s tweets just as well as if we had downloaded them directly (well over 50% accuracy, see graph below). Indeed, 95% of the potential predictive accuracy that a machine learning algorithm might achieve is obtainable just from friends’ data.

file-20200622-54989-bo83l3.jpg
Average predictability from your circle of closest friends (blue line). A value of 50% means getting the next word right half of the time — no mean feat as most people have a vocabulary of around 5,000 words. The curve shows how much an AI algorithm can predict about you from your friends’ data. Roughly 8-9 friends are enough to predict your future posts as accurately as if the algorithm had access to your own data (dashed line). Bagrow, Liu, & Mitchell (2019)

Our results mean that even if you #DeleteFacebook (which trended after the Cambridge Analytica scandal in 2018), you may still be able to be profiled, due to the social ties that remain. And that’s before we consider the things about Facebook that make it so difficult to delete anyway.We also found it’s possible to build profiles of  — so-called “” — based on their contacts who are on the platform. Even if you have never used Facebook, if your friends do, there is the possibility a shadow profile could be built of you.

On social media platforms like Facebook and Twitter, privacy is no longer tied to the individual, but to the network as a whole.

No more free will? Not quite

But all hope is not lost. If you do delete your account, the information contained in your social ties with friends grows stale over time. We found predictability gradually declines to a low level, so your privacy and anonymity will eventually return.

 

While it may seem like algorithms are eroding our ability to think for ourselves, it’s not necessarily the case. The evidence on the effectiveness of psychological profiling to influence voters is thin.

Most importantly, when it comes to the role of people versus algorithms in things like spreading (mis)information, people are just as important. On Facebook, the extent of your exposure to diverse points of view is more closely related to your social groupings than to the way News Feed presents you with content. And on Twitter, while “fake news” may spread faster than facts, it is primarily people who spread it, rather than bots.

Of course, content creators exploit social media platforms’ algorithms to promote content, on YouTubeReddit and other platforms, not just the other way round.

At the end of the day, underneath all the algorithms are people. And we influence the algorithms just as much as they may influence us.

[Source: This article was published in theconversation.com By Misha Ketchell - Uploaded by the Association Member: Jay Harris]

Categorized in Search Engine

Learn key insights that will help you understand how the algorithms of Instagram, YouTube, TikTok, Twitter, and Facebook work.

Here’s an old question that gets asked every year:

How do social media algorithms work?

But, you can often uncover strategic insights by looking at an old question like this one from a different perspective.

In fact, there’s a term for this effect.

It’s called the “parallax” view.

parallax-view.png

For example, marketers often look for influencers on the social media platforms with the greatest reach.

But, influencers evaluate these same platforms based on their opportunity to grow their audience and make more money.

 

This explains why The State of Influencer Marketing 2020: Benchmark Report found that the top five social media platforms for influencer marketing are:

  • Instagram (82%).
  • YouTube (41%).
  • TikTok (23%).
  • Twitter (23%).
  • Facebook (5%).

This list made me wonder why marketers focus on the reach of their campaign’s outputs, but influencers are focused on the growth of their program’s outcomes.

Influencers want to learn how the Instagram and YouTube algorithms work, because they want their videos discovered by more people.

And influencers are interested in learning how the TikTok and Twitter algorithms work, because they are thinking about creating content for those platforms.

Facebook’s algorithm, however, doesn’t seem quite as important to today’s influencers – unless Facebook represents a significant opportunity for them to make more money.

There are a lot of strategic insights that marketers can glean from looking at how social media algorithms work from an influencer’s point of view.

How the Instagram Algorithm Works

Back in 2016, Instagram stopped using a reverse-chronological feed.

Since then, the posts in each user’s feed on the platform has been ordered according to the Instagram algorithm’s ranking signals.

According to the Instagram Help Center:

“Instagram’s technology uses different ways, or signals, to determine the order of posts in your feed. These signals are used to help determine how your feed is ordered, and may include:

  • “Likelihood you’ll be interested in the content.
  • “Date the post was shared.
  • “Previous interactions with the person posting.”

This has a profound impact on influencers – as well as the marketers who are trying to identify the right influencers, find the right engagement tactics, and measure the performance of their programs.

Relevance

The first key signal is relevance, not reach.

Is click fraud affecting your PPC campaigns?
Start a free trial now and experience your GoogleAds account in a fraud-free environment.

 

Why?

Because Instagram users are more likely to be interested in an influencer’s content if it is relevant – if it’s about what interests them.

In other words, if you’re interested in football (a.k.a., soccer), then the likelihood that you’ll be interested in content by Nabaa Al Dabbagh, aka “I Speak Football Only,” is high.

But, far too many marketers are looking for celebrities and mega-influencers who have lots of Instagram followers (a.k.a., reach), instead of looking for macro-, mid-tier, micro-, or nano-influencers who are creating relevant content that their target audience is more likely to find interesting.

i-speak-football-only.png

Recency

The second key signal is recency, or how recently a post has been shared.

This gives an advantage to influencers like Marwan Parham Al Awadhi, a.k.a., “DJ Bliss,” who post frequently.

dj-bliss.png

Unfortunately, far too many marketers are engaging influencers to create a single post during a campaign instead of building a long-term relationship with brand advocates who will generate a series of posts that recommend their brand on an ongoing basis.

Resonance

The third key signal is resonance.

In other words, how engaging are an influencer’s posts?

Do they prompt interactions such as comments, likes, reshares, and views with the influencer’s audience?

And, unfortunately, way too many marketers assume that an influencer’s post that mentions their brand has increased their brand awareness, using bogus metrics like Earned Media Value (EMV).

If they’d read, Why International Search Marketers Should Care About Brand Measurement, then they’d realize there are a variety of legitimate ways to measure the impact of an influencer marketing campaign on:

  • Brand awareness.
  • Brand frequency.
  • Brand familiarity.
  • Brand favorability.
  • Brand emotions.
  • Purchase consideration.
  • Brand preference.
  • Brand demand.

Using this parallax view, it’s easy to see that too many marketers mistakenly think influencer marketing is just like display advertising.

They’re buying posts from influencers the same way they would buy ads from publishers.

So, marketers who only look at an influencer’s reach shouldn’t be shocked, shocked to discover that some influencers are using bad practices such as fake followers, bots, and fraud to inflate their numbers.

If you use a one-dimensional view of an influencer’s influence, then you reap what you sow.

How Does the YouTube Algorithm Work?

Now, I’ve already written several articles on how the YouTube algorithm works, including:

But, these articles were written for marketers, not influencers.

So, what can we learn from looking at YouTube’s algorithm from an influencer’s point of view?

Well, according to YouTube Help:

“The goals of YouTube’s search and discovery system are twofold: to help viewers find the videos they want to watch, and to maximize long-term viewer engagement and satisfaction.”

So, YouTube influencers need to start by creating great content on discoverable topics.

Why?

Well, YouTube is one of the most-used search engines in the world.

People visit the site looking for videos about all sorts of subjects.

These viewers may not necessarily be looking for a specific influencer’s video, but they’ll discover it if it ranks well in YouTube search results or suggested videos.

Learn how to use Google Trends to find out what your audiences is looking for on YouTube.

The default results in Google Trends show “web search” interest in a search term or a topic.

But, if you click on the “web search” tab, the drop-down menu will show you that one of your other options is “YouTube search” interest.

YouTube influencers can then use what they see to inform their content strategies.

For example, you might learn that there was 31% more YouTube search interest worldwide in the topic, beauty, than in the topic, fashion.

fashion-vs-beauty.png

Or you might discover that there was 18 times more YouTube search interest worldwide in the sport, drifting, than in the sport, motorsport.

motorsport-vs-drifting.png

YouTube’s algorithm can’t watch your videos, so you need to optimize your metadata, including your titles, tags, and descriptions.

Unfortunately, most marketers don’t use this approach to find the search terms and topics on YouTube that are relevant for their brand and then identify the influencers who are creating content that ranks well for these keywords and phrases.

Now, getting your YouTube video content discovered is only half the battle.

 

Influencers also need to build long watch-time sessions for their content by organizing and featuring content on their channel, including using series playlists.

As YouTube Help explains:

“A series playlist allows you to mark your playlist as an official set of videos that should be viewed together. Adding videos to a series playlist allows other videos in the playlist to be featured and recommended when someone is viewing a video in the series. YouTube may use this info to modify how the videos are presented or discovered.”

Fortunately, one of the guest speakers for NMA’s program was Mark Wiens, one of the most famous food vloggers in the world.

His YouTube channel has more than 1.4 billion views and almost 6.7 million subscribers.

Here are examples of the playlists that he had created, including Thai food and travel guides.

mark wien

Now, marketers could also look over the playlists on the YouTube channels of influencers when they’re evaluating which ones are “right” for a campaign.

However, I strongly suspect that this only happens once in a blue moon.

 

How Does the TikTok Algorithm Work?

The TikTok Newsroom posted How TikTok recommends videos #ForYou just before I was scheduled to talk about this topic.

Hey, sometimes you get lucky.

tiktok.png

Here’s what I learned:

“When you open TikTok and land in your For You feed, you’re presented with a stream of videos curated to your interests, making it easy to find content and creators you love. This feed is powered by a recommendation system that delivers content to each user that is likely to be of interest to that particular user.”

 

 

So, how does this platform’s recommendation system work?

According to TikTok:

“Recommendations are based on a number of factors, including things like:

  • “User interactions such as the videos you like or share, accounts you follow, comments you post, and content you create.
  • “Video information, which might include details like captions, sounds, and hashtags.
  • “Device and account settings like your language preference, country setting, and device type. These factors are included to make sure the system is optimized for performance, but they receive lower weight in the recommendation system relative to other data points we measure since users don’t actively express these as preferences.”

The TikTok Newsroom adds:

“All these factors are processed by our recommendation system and weighted based on their value to a user. A strong indicator of interest, such as whether a user finishes watching a longer video from beginning to end, would receive greater weight than a weak indicator, such as whether the video’s viewer and creator are both in the same country. Videos are then ranked to determine the likelihood of a user’s interest in a piece of content, and delivered to each unique For You feed.”

TikTok cautions:

“While a video is likely to receive more views if posted by an account that has more followers, by virtue of that account having built up a larger follower base, neither follower count nor whether the account has had previous high-performing videos are direct factors in the recommendation system.”

It’s worth noting that Oracle has won the bid to acquire TikTok’s U.S. operations after ByteDance rejected a bid by Walmart and Microsoft.

Meanwhile, YouTube released YouTube Shorts, a TikTok-like feature, while Facebook recently launched Instagram Reels, which is basically a TikTok knock-off.

So, it appears that some very big players are convinced that TikTok represents a significant opportunity to make more money, or a competitive threat to the growth of their own social media platforms.

I wish that I could add more, but I’m a stranger here myself.

How Does Twitter’s Algorithm Work?

When Twitter was launched back in 2006, it had a simple timeline structure and tweets were displayed in reverse chronological order from the people you followed.

 

 

But, like other social media, Twitter started using an algorithm to show users posts that different factors indicate they’ll like.

The biggest recent change to Twitter’s algorithm took place in 2017.

According to a Twitter blog post by Nicolas Koumchatzky and Anton Andryeyev:

“Right after gathering all Tweets, each is scored by a relevance model. The model’s score predicts how interesting and engaging a Tweet would be specifically to you. A set of highest-scoring Tweets is then shown at the top of your timeline, with the remainder shown directly below.”

Their post added:

“Depending on the number of candidate Tweets we have available for you and the amount of time since your last visit, we may choose to also show you a dedicated “In case you missed it” module. This modules meant to contain only a small handful of the very most relevant Tweets ordered by their relevance score, whereas the ranked timeline contains relevant Tweets ordered by time. The intent is to let you see the best Tweets at a glance first before delving into the lengthier time-ordered sections.”

How Does Facebook’s Algorithm Work?

The biggest recent change to Facebook’s algorithm took place in January 2018.

In a Facebook post, Mark Zuckerberg announced:

“I’m changing the goal I give our product teams from focusing on helping you find relevant content to helping you have more meaningful social interactions.”

He added:

“The first changes you’ll see will be in News Feed, where you can expect to see more from your friends, family and groups. As we roll this out, you’ll see less public content like posts from businesses, brands, and media. And the public content you see more will be held to the same standard — it should encourage meaningful interactions between people.”

That same day, Adam Mosseri, who was then the head of News Feed, also wrote a Facebbok post that said:

“Today we use signals like how many people react to, comment on or share posts to determine how high they appear in News Feed. With this update, we will also prioritize posts that spark conversations and meaningful interactions between people. To do this, we will predict which posts you might want to interact with your friends about, and show these posts higher in feed. These are posts that inspire back-and-forth discussion in the comments and posts that you might want to share and react to – whether that’s a post from a friend seeking advice, a friend asking for recommendations for a trip, or a news article or video prompting lots of discussion.”

He added:

“Because space in News Feed is limited, showing more posts from friends and family and updates that spark conversation means we’ll show less public content, including videos and other posts from publishers or businesses.”

So, it isn’t surprising that influencers got the memo.

Which explains why so few believe Facebook represents a significant opportunity to make more money.

Ironically, it’s unclear that marketers got the memo.

Far too many are still cranking out Facebook posts and videos despite the fact that few people are reacting to, commenting on, or sharing them.

Or, as I wrote in Two Social Media Vanity Metrics You Need to Stop Tracking, marketers should stop tracking Facebook Page Likes and Followers because “you’re lucky if .0035% of your Fans and Followers even sees your post or tweet these days.”

new-media-academy.jpg

The Takeaway

These are just some of the strategic insights that marketers can discover by looking at how social media algorithms work from an influencer’s point of view.

If you’re a marketer, then I suggest you move most of the people and budget that you’ve dedicated to creating branded content on Facebook into influencer marketing on Instagram and YouTube.

As for TikTok and Twitter, wait until after the dust settles later this year.

[Source: This article was published in searchenginejournal.com By Greg Jarboe - Uploaded by the Association Member: Corey Parker]

Categorized in Social

Google suffered a glitch that negatively impacted search quality. This was not an update. It was a mistake of some kind.

Google’s search algorithm suffered an unprecedented glitch that affected search results.

Many in the search community believed it was an update.

The disruption in Google search was not an update.

Google’s John Mueller tweeted:

“I don’t have all the details yet, but it seems like this was a glitch on our side and has been fixed in the meantime.”

Official Explanation

Tuesday August 11, 2020 Google’s Webmasters Twitter account tweeted an explanation.

Caffeine Index Issue?

Google has a web crawling and indexing system called Caffeine.  Caffeine allowed Google to process data faster than ever before.

 

This Caffeine indexing system empowered Google to continually index the entire web in real-time.

With a fresher index, Google could then show more up to date search results.

 

Google’s Gary Illyes  tweeted an explanation of how complex a search index is, with a caveat that the list he published was only a partial list.

“The indexing system, Caffeine, does multiple things:

1. ingests fetchlogs,
2. renders and converts fetched data,
3. extracts links, meta and structured data,
4. extracts and computes some signals,
5. schedules new crawls,
6. and builds the index that is pushed to serving.”

Followed by this tweet:

“If something goes wrong with most of the things that it’s supposed to do, that will show downstream in some way. If scheduling goes awry, crawling may slow down. If rendering goes wrong, we may misunderstand the pages. If index building goes bad, ranking & serving may be affected.”

Then he concluded with this:

Google Caffeine Index?

It was kind of surprising to see the Google Caffeine system cited.

It was officially announced in 2010.

The announcement stated that it was a foundation for indexing that was meant to scale for the future.

This is what the official 2010 Caffeine announcement stated:

“We’ve built Caffeine with the future in mind.

Not only is it fresher, it’s a robust foundation that makes it possible for us to build an even faster and comprehensive search engine that scales…”

Advertisment

Google Search Glitch Was Worldwide

The Google search glitch was keenly felt in Europe as well as Asia and all English speaking countries.

Google’s search glitch appeared to affect all languages, countries, and niches.

It affected everything from local services to recipes.

Ecommerce sites reported extreme fluctuations in rankings.

Bad Search Results

Recipe Blog SEO Casey Markee tweeted a screenshot of how bad the recipe search results were.

Think you should be getting better results with your Google Ads?
Your campaign may be suffering from click fraud. Check if you need to protect your ads from competitors & bots. Simple setup. Get a free checkup today.

Google Search Glitch Created Poor Search Results

Google’s search results became incredibly bad, some to the point of being useless.

I tried searching for an article from a specific site and Google wouldn’t show it to me, even when I used the name of the site that contained the article.

It felt somewhat like in the old days when PageRank had a stronger influence.

WebmasterWorld had great real-time coverage of the glitch as it happened.

 

A member from WebmasterWorld, webdev29, noted how the big sites like Amazon seemed to dominate the SERPs.

“huge update also in France ATM, no word to describe the mess, its simply crazy ! there is no more ecommerces in my SERP (decoration) and mine has just lost everything…6 years destroyed in just one minute and the lives of several employees at stake! it’s not possible that it continues like this, in the SERP, there are only the big marketplaces (cdiscount, amazon, laredoute, aliexpress…) and some more or less recent sites without much interest…all the rest has disappeared on the deep pages of the search engine.”

 

Report from Italy

WebmasterWorld member teokolo shared:

“Seems like a big update in progress here in Italy.
Every niche I follow is messed up. Shops are gone, affiliate sites have disappeared, serps are full of Amazon, ebay and news sites.”

Google Glitch Impact in Norway

mini_007 said:

“wow insane big update here in Norway, never seen so big change.”

Massive Fluctuations in Google Search Results

Whether it was on Facebook, WebmasterWorld or Twitter, the common observation was that there were massive fluctuations in the search results.

This report from WebmasterWorld member Whoa182 is typical:

“What the hell is going on?

Just noticed my articles have gone from page 1 to page 7+

Seems to have just happened in the past few hours! Quite a few of my competitors have all disappeared from the SERPs.

Edit: Okay, it’s just massive fluctuations in page positions. One minute it’s on page 1, next it’s page 7 or whatever, and then back again.”

Google Has Not Yet Explained the Cause

Google’s Danny Sullivan is the one who typically announces updates.

 

Google Webmaster Trends Analysts, Gary Illyes and Mueller also share announcements of changes at Google as well, including glitches.

For example, at the beginning of 2020 Google suffered a glitch that caused an issue with Google’s index. It was Illyes who did the explaining.

Google suffered a massive glitch that caused the search results around the world to become less usable.

The cause of the glitch, according to Google Webmaster Trends Analyst Gary Illyes appears to be related to Google’s Caffeine indexing system or something along those lines.

[Source: This article was published in searchenginejournal.com By Roger Montti - Uploaded by the Association Member: Deborah Tannen]

Categorized in Search Engine

Pinterest aims to display a greater variety of content types in the home feed by utilizing a new ranking model.

Pinterest is introducing a new ranking model to its home feed in an effort to surface certain types of content more often.

Traditionally, Pinterest ranks content in the home feed using a click-through prediction model.

Pins that a user is most likely to click on, as determined by past activity, are prioritized in their home feed

While that model is effective at maximizing user engagement, it’s not the best model for surfacing a variety of content types.

For example, if a user never clicks on video content then they’ll never be shown pins with video in their home feed.

 

But that doesn’t necessarily mean they wouldn’t engage with video content if it were to be surfaced.

Pinterest found itself with a problem of wanting to boost more content types while still keeping content recommendations relevant.

To solve this problem, Pinterest is introducing a real-time ranking system for its home feed called “controllable distribution.”

Controllable Distribution

Pinterest describes controllable distribution as a “flexible real-time system.”

It’s not a complete algorithm overhaul. Rather, controllable distribution is only applied after the traditional home feed ranking algorithm.

Pinterest will still use its click-through prediction model to find relevant content. Then it will apply controllable distribution to diversify the types of content being displayed.

Controllable distribution makes it possible to specify a target for how many impressions a certain content type should receive.

For example, controllable distribution could be used to specify that 4% of users’ home feeds should contain video content.

This is done through a system that tracks what percentage of the feed was video in the past. Then, the system boosts or demotes content according to how close that percentage is to the specified target.

Pinterest says this can be accomplished while still respecting users’ content preferences.

What Does This Mean for Marketers?

As a real-time system, the controllable distribution model will be continuously adjusted.

On one hand, that means the home feed won’t get stale for users.

On the other hand, it’s not exactly possible to optimize for an algorithm that changes in realtime.

Perhaps the best piece of advice for Pinterest marketers to take away from this is to follow Pinterest’s lead.

Pinterest is diversifying the types of content in the home feed. If you want more opportunities to show up in peoples’ feeds then diversify the types of content you publish.

For example, if you only publish photos, then consider adding some videos or GIFs to the mix. Maybe some product pins if you’re an e-commerce retailer.

Pinterest’s target for displaying certain types content will be changing all the time.

Publishing a wide variety of content will help ensure you have the right type of content available at the time Pinterest wants to display it.

Additional Notes

Pinterest’s home feed ranking team used to do manually what controllable distribution is designed to do algorithmically.

Yes, Pinterest’s home feed ranking team actually used to step in and adjust how often certain types of content appeared in users’ home feed.

Yaron Greif of Pinterest’s home feed ranking team describes the old process as “painful for both practical and theoretical reasons.”

“In practice, these hand-tuned boosts quickly became unmanageable and interfered with each other. And worse, they often stop working over time — especially when ranking models are updated. We regularly had to delay very promising new ranking models because they broke business constraints.

In theory, controlling content on a per-request basis is undesirable because it prevents personalization. If we show each user the same number of video Pins we can’t show more videos to people who really like to watch videos or vice versa.”

Pinterest says it’s committed to investing in the post-ranking stage of surfacing content. So it’s possible we may see this model applied elsewhere on the platform in the future.

 

[Source: This article was published in searchenginejournal.com By Matt Southern - Uploaded by the Association Member: Edna Thomas]

Categorized in Social

Google gives a detailed explanation of what causes algorithm to rewrite meta descriptions.

In a Webmaster Central hangout, a publisher asked Google’s John Mueller why their meta description was being rewritten. Mueller’s answer offered a peek into how Google’s algorithm chooses when to rewrite meta descriptions.

The question was specifically about a meta description on the home page being rewritten on the Google search results pages (SERPs) for branded search queries. The publisher used the example of using the modifier “UK” with the brand name.

There are no specifics mentioned in the question so there is no way to address the publisher’s issue directly.

But because Mueller’s answer is general, it provided an answer that gave some insight into why Google rewrites meta descriptions.

Here’s the question

“We have an issue with the meta description that is being displayed for hour home page.

So, even though we have a meta description that is being implemented on that particular page, somehow in Google when our website appears, the meta description is completely different.

And in some cases, if we search for our company name plus the word “UK,” the meta description makes no sense whatsoever. It’s just a bunch of words put together from various parts of the page.

I know sometimes Google goes hunting for various things on the page if it cannot find relevant content for that particular region.

So I guess my question is, because we have a lot of traffic that is coming up from branded searches… it is important for us to have the correct meta description showing up.”

What do we do to rectify the situation?”

 

 

John Mueller Explains What Causes Meta Description Rewrites

Before answering why Google rewrites meta descriptions, John Mueller stated he hasn’t seen the publishers specific search result and could not answer why it was specifically happening for a query he hasn’t seen.

“It’s hard to say without looking at the search results. So that’s kind of the one part.”

Then he offered explanations of what causes Google to rewrite meta descriptions.

First he states that you have to have a meta keyword:

“Usually what happens is we need to have the description meta tag on the page. That’s kind of the first step.

It sounds like you already have that set up.”

Reason 1: Poor Use of Meta Description

Now the explanation of what triggers Google’s algorithm to rewrite the meta description tag:

“The other thing there is that we need to be able to, I guess, trust the meta description on the page so that it looks kind of reasonable.

In particular, sometimes when we see a bunch of keywords that are just kind of collected in the meta description.

Then that’s something that our systems might look at and say well, this doesn’t look that useful for users.

So they’ll try to rewrite something else.”

Mueller is saying that one of the reasons why the meta description may be rewritten is because it’s focused more on keywords and less on what the page is about.

But more importantly, what makes that meta description a target for rewriting is that he said that it “doesn’t look that useful for users.”

Reason 2: Content and Query Matching Can Trigger Meta Description Rewrite

That “less useful” part, in the context of the above publisher, is relative to the search query. T

he publisher said that branded queries with the “UK modifier were being rewritten.

That “UK” search query modifier may be what’s causing Google to rewrite the meta description.

If the web page itself isn’t specifically sending UK related content signals then Google might choose to modify the meta description.

Adding modifiers to search queries can cause Google to rewrite the meta descriptions (and title tags too). This is especially going to happen when the keyword modifiers (like UK or Home Page) don’t exist in the written content of the page.

Suffering from fewer conversions for the same ad spend on your Google Ads?
Your ads might be getting click fraud. Check if you need to protect your ads from competitors & bots. Simple setup. Start your free checkup today.

 

Example of Query and Content Matching Causing Meta Description Rewrite

Compare the search query “Walmart” to the query, “Walmart Home Page” and you will see that the search query “Walmart Home Page” has a rewritten meta description.

example-google-rewriting-me-5ecd365921828.gif

Google is trying to show a relevant meta description for the term Walmart Home Page. But the words “Home Page” do not exist on the Walmart home page.

But those words do exist on the yellow star icon that has this alt tag: “Icon for spark” and if you hover over the Walmart and “spark icon” logo, the words, “Walmart Homepage” show up in an alt tag tooltip.

So what’s happening is that Google’s algorithm is trying to make the meta description relevant for the search query, “Walmart Home Page.”

The algorithm is trying to do that by rewriting the meta description. But as you can see above in the case of the search query Walmart Home Page, Google isn’t doing that so well.

John Mueller confirms what I wrote above:

“And most of the time when it tries to rewrite something, it’s based on the content on the page itself.”

What happened in the Walmart Home Page search query that caused Google to rewrite the meta description is that the words Home Page or Homepage do not exist anywhere on the Walmart home page except for in the alt tag for the home page button.

So Google grabbed some alt tag text associated with the Walmart Home button, selected the wrong alt tag, and displayed the phrase, “Icon for spark” in the rewritten meta description.

Reason 3: Search Query Influences Meta Description Rewrite

As I illustrated above, and John Mueller will say below, the meta description rewriting depends on the search query. And I would expand that to say that it depends on the search query and the content on the web page.

 

Here’s what Mueller said:

“And the other thing… you noticed, is the description can vary depending on the query that is used.

So the first thing that I would do is just take the normal branded query that you use and double check that the description that you provide in the meta description is actually pretty useful and not too… spammy or overdone.

And then go from there, essentially, to figure out… is this something where Google always gets it wrong?

Or is it something where sometimes Google’s algorithms pick up something else on the page and get it wrong?”

Google Meta Description Rewriting Explained

John Mueller gave a great explanation of the reason why Google rewrites search queries.

I know some people are going to react and say that Google’s rewriting is arbitrary. But it’s not arbitrary.

This article has described specific situations that cause Google’s algorithm to rewrite meta description tags.

Google’s algorithm rewrites meta descriptions based on the relationship between the search query and the web page content.

So if you have an issue with Google rewriting the meta tags, take a closer look at how the search query relates to the on-page content.

Watch Google’s Webmaster Central here:

 

[Source: This article was published in searchenginejournal.com By Roger Montti - Uploaded by the Association Member: Robert Hensonw]

Categorized in Search Engine

Google has started rolling out a new core search algorithm update that it calls the May 2020 Core Update. The new update comes months after the search giant released the last core algorithm update back in January. The goal behind updates like these are to try and improve the quality of results that users get when they enter a search query into the site. While this is good for an end user, many sites might see their performance fluctuate as a result of the core update. To avoid its results being manipulated, Google doesn't give out the details of its routine updates, merely advising content creators to focus on quality content.

 

A tweet posted by Google earlier today announced that the May 2020 Core Update has started rolling out for all users. The update would, however, take about one to two weeks to fully roll out.

Screenshot 1
Broader update to enhance Google search experience
The new update is a broad core algorithm update that would bring a list of changes to Google's search algorithms and systems. This is unlike the regular changes that the company releases incrementally to improve search results.
“Several times a year, we make significant, broad changes to our search algorithms and systems. We refer to these as ‘core updates.' They're designed to ensure that overall, we're delivering on our mission to present relevant and authoritative content to searchers,” Google noted in a blog post defining the core algorithm updates.

Apart from other changes, core updates are likely to affect Google Discover. Some sites are also expected to note drops or gains during such updates.

“We know those with sites that experience drops will be looking for a fix, and we want to ensure they don't try to fix the wrong things. Moreover, there might not be anything to fix at all,” the search giant said.

Having said that, webmasters and search engine optimisation (SEO) teams are advised to stay focussed on bringing quality content through their sites. The content should provide original information, reporting, research, or analysis along with a comprehensive description of the topic. It is also recommended to have a descriptive but not exaggerating or sensationalist headline. Furthermore, there are a list of content, quality, expertise, and comparative questions that webmasters and SEO folks should ask themselves about their content.

Drops, gains from search algorithm updates are common
It is natural that since Google makes certain changes at the algorithm level, some websites face drops, while the other ones get some gains in their traffic. There isn't any hard and fast rule to fix such impacts post an update starts rolling out. Nevertheless, it's better to consider analytics to understand ranking changes of your website.

The last core algorithm update that Google brought to its search engine took place in January. The company had also introduced a new design for desktop searches around the same time that faced some backlash from users initially.

 

[Source: This article was published in gadgets.ndtv.com By Jagmeet Singh - Uploaded by the Association Member: Deborah Tannen]

Categorized in Search Engine

Internet marketing and advertising is becoming increasingly difficult and expensive, especially for small to mid-level businesses. Veteran SEO expert Tony Rockliff urges business owners to utilize the power of YouTube as a promising alternative to the otherwise slow, painful and expensive build of a Google SEO campaign.

CLEARWATER, Fla.Feb. 24, 2020 /PRNewswire-PRWeb/ -- In 2020, according to the World Advertising and Research Center, spending on internet advertising will reach more than 50% of total global ad spend, an all-time record.(1) "A subset of internet advertising—search engine optimization (SEO) is now the major battleground in marketing today. SEO expert Tony Rockliff, founder and CEO of Tony Rockliff Productions, states as SEO "gets bigger, it gets tougher." For an increasing number of companies, especially SMBs, the smart move is to consider YouTube as an additional SEO powerhouse instead of the conventional reliance on Google. By using YouTube, business owners can combat the stiff competition for consumer attention and variating Google search algorithms.

 

Text Versus Video Content

According to a research study from Common Sense Media, more than twice as many young people watch videos every day as did four years ago, while the average time spent watching videos—primarily on YouTube—has roughly doubled, to an hour a day.(3) Video's popularity has exploded, while text takes a back seat. It is increasingly obvious in the industry that text-based content is saturated, and that if a company isn't willing to give it at least one year and invest considerable amounts, they shouldn't spend much time on traditional Google based SEO. (2)

The combination of the video-centricity of today's consumers coupled with the increasing expense and difficulty of attracting attention via text-based Google listings, Rockliff suggests, is what identifies YouTube an increasingly robust platform for video-savvy marketers.

How Businesses Can Adapt to YouTube

To capitalize on this opportunity, Rockliff urges marketers to research YouTube to qualify exactly what video content is needed, and which of this content will get the most responses from its viewers, or potential clients. He organizes the four major stages of YouTube optimization:

  • 1. Find out what is being searched for on YouTube in your area or niche that you can compete for.
  • 2. Create video content that answers what is being searched for, and also provides what YouTube is searching for, i.e. views per video, average time spent watching, engagement per video, and number of subscribers gained per video.
  • 3. Publish your videos properly and in an optimized manner.
  • 4. Promote your videos according to how and when YouTube wants to see them promoted.

Rockliff has been in search engine optimization since 1998 and online marketing since 1995. His online community membership site has grown to 1.3 million members and was receiving 1.5 billion hits per year before he sold it in 2002. Over the years, Rockliff has seen profound changes in both opportunity and approach of YouTube as a marketing strategy, and right now, YouTube represents a great prospect to get noticed and build a brand loyal following. This is especially useful for organizations that do not have an extensive marketing budget. "The key is to understand what you're selling and optimize all four major stages," Rockliff states.

Tony Rockliff will be speaking at the Podfest 2020 Multimedia Expo, March 6th-8th, at the Orlando World Center Marriott in Orlando, Florida. For more information, please see http://podfestexpo.com/speakers/

About Tony Rockliff Productions:

Tony Rockliff Productions was founded in 1995 by digital pioneer and trailblazer, Tony Rockliff. His video marketing company is based out of Clearwater, Florida, and brings over fifty years of audio/video marketing experience to the business. Remaining to be a top disruptor of the video marketing and media industry throughout his career, his world-renowned success is a product of his passion for storytelling through the art of video. Tony Rockliff Productions specializes in video and audio creation, producing music and videos, YouTube optimization, and building out-of-the-ordinary websites. Currently, Tony Rockliff Productions focuses on organic YouTube video marketing, a profitable niche of the industry that is host to 1.9 billion logged-in users per month. You can visit him here https://tonyrockliff.com/

 

  • 1. Handley, Lucy, "Global ad spend has slowed but 2020 looks set to be a bumper year," CNBC, October 24, 2019, cnbc.com/2019/10/24/global-ad-spend-has-slowed-but-2020-looks-set-to-be-a-bumper-year.html.
  • 2. Patel, Neil, "Everything I Taught You About SEO Was Wrong," neilpatel.com/blog.
  • 3. Siegel, Rachel, "Tweens, teens, and screens: The average time kids spend watching online videos has doubled in 4 years," Washington Post, October 29, 2019, washingtonpost.com/technology/2019/10/29/survey-average-time-young-people-spend-watching-videos-mostly-youtube-has-doubled-since/.

[Source: This article was published in finance.yahoo.com - Uploaded by the Association Member: Jasper Solander]

Categorized in Search Engine

Now that the Google January 2020 core update is mostly rolled out, we have asked several data providers to send us what they found with this Google search update. All of the data providers agree that this core update was a big one and impacted a large number of web sites.

The facts. What we know from Google, as we previously reported, is that the January 2020 core update started to roll out around 12:00 PM ET on Monday, January 13th. That rollout was “mostly done” by Thursday morning, on January 16th. We also know that this was a global update, and was not specific to any region, language or category of web sites. It is a classic “broad core update.”

What the tools are seeing. We have gone to third-party data companies asking them what their data shows about this update.

 

RankRanger. Mordy Oberstein from RankRanger said, “the YMYL (your money, your life) niches got hit very hard.” “This a huge update,” he added. “There is massive movement at the top of the SERP for the Health and Finance niches and incredible increases for all niches when looking at the top 10 results overall.”

Here is a chart showing the rank volatility broken down by industry and the position of those rankings:

 all-niche-data-jan2020-core-update-800x550.png

“Excluding the Retail niche, which according to what I am seeing was perhaps a focus of the December 6th update, the January 2020 core update was a far larger update across the board and at every ranking position,” Mordy Oberstein added. “However, when looking at the top 10 results overall during the core update, the Retail niche started to separate itself from the levels of volatility seen in December as well.”

SEMRush. Yulia Ibragimova from SEMRush said “We can see that the latest Google Update was quite big and was noticed almost in every category.” The most volatile categories according to SEMRush, outside of Sports and News, were Online communities, Games, Arts & Entertainments, and Finance. But Yulia Ibragimova added that all categories saw major changes and “we can assume that this update wasn’t aimed to any particular topics,” she told us.

SEMRush offers a lot of data available on its web site over here. But they sent us this additional data around this update for us.

Here is the volatility by category by mobile vs desktop search results:

semrush-catts-642x600.png

The top ten winners according to SEMRush were Dictionary.com, Hadith of the Day, Discogs, ABSFairings, X-Rates, TechCrunch, ShutterStock, 247Patience, GettyImages and LiveScores.com. The top ten losers were mp3-youtube.download, TotalJerkFace.com, GenVideos.io, Tuffy, TripSavvy, Honolulu.gov, NaughtyFind, Local.com, RuthChris and Local-First.org.

 

Sistrix. Johannes Beus from Sistrix posted their analysis of this core update. He said “Domains that relate to YMYL (Your Money, Your Life) topics have been re-evaluated by the search algorithm and gain or lose visibility as a whole. Domains that have previously been affected by such updates are more likely to be affected again. The absolute fluctuations appear to be decreasing with each update – Google is now becoming more certain of its assessment and does not deviate as much from the previous assessment.”

Here is the Sistrix chart showing the change:

 uk.sistrix.com_onhealth.com_seo_visibility-1-800x361.png

According to Sistrix, the big winners were goal.com, onhealth.com, CarGurus, verywellhealth.com, Fandango, Times Of Israel, Royal.uk, and WestField. The big losers were CarMagazine.co.uk, Box Office Mojo, SkySports, ArnoldClark.com, CarBuyer.co.uk, History Extra, Evan Shalshaw, and NHS Inform.

SearchMetrics. Marcus Tober, the founder of SearchMetrics, told us “the January Core Update seems to revert some changes for the better or worse depending on who you are. It’s another core update where thin content got penalized and where Google put an emphasis in YMYL. The update doesn’t seem to affect as many pages as with the March or September update in 2019. But has similar characteristics.”

Here are some specific examples SearchMetrics shared. First was that Onhealth.com has won at March 2019 Core update and lost at September 2019 and won again big time at January 2020 Core update

 onhealth-800x320.png

While Verywellhealth.com was loser during multiple core updates:

 verywell-800x316.png

Draxe.com, which has been up and down during core updates, with this update seems to be a big winner with +83%. but in previous core updates, it got hit hard:

 draxe-800x318.png

The big winners according to SearchMetrics were esty.com, cargurus.com, verywellhealth.com, overstock.com, addictinggames.com, onhealth.com, bigfishgames,com and health.com. The big losers were tmz.com, academy.com, kbhgames.com, orbitz.com, silvergames.com, autolist.com, etonline.com, trovit.com and pampers.com.

 

What to do if you are hit. Google has given advice on what to consider if you are negatively impacted by a core update in the past. There aren’t specific actions to take to recover, and in fact, a negative rankings impact may not signal anything is wrong with your pages. However, Google has offered a list of questions to consider if you’re site is hit by a core update.

Why we care. It is often hard to isolate what you need to do to reverse any algorithmic hit your site may have seen. When it comes to Google core updates, it is even harder to do so. If this data and previous experience and advice has shown us is that these core updates are broad, wide and cover a lot of overall quality issues. The data above has reinforced this to be true. So if your site was hit by a core update, it is often recommended to step back from it all, take a wider view of your overall web site and see what you can do to improve the site overall.

[Source: This article was published in searchengineland.com By Barry Schwartz - Uploaded by the Association Member: Edna Thomas]

Categorized in Search Engine

Michael struggles to find the search results he’s looking for, and would like some tips for better Googling

 Want to search like a pro? These tips will help you up you Googling game using the advanced tools to narrow down your results. Photograph: Alastair Pike/AFP via Getty Images
Last week’s column mentioned search skills. I’m sometimes on the third page of results before I get to what I was really looking for. I’m sure a few simple tips would find these results on page 1. All advice welcome. Michael

Google achieved its amazing popularity by de-skilling search. Suddenly, people who were not very good at searching – which is almost everyone – could get good results without entering long, complex searches. Partly this was because Google knew which pages were most important, based on its PageRank algorithm, and it knew which pages were most effective, because users quickly bounced back from websites that didn’t deliver what they wanted.

Later, Google added personalisation based on factors such as your location, your previous searches, your visits to other websites, and other things it knew about you. This created a backlash from people with privacy concerns, because your searches into physical and mental health issues, legal and social problems, relationships and so on can reveal more about you than you want anyone else – or even a machine – to know.

When talking about avoiding “the creepy line”, former Google boss Eric Schmidt said: “We don’t need you to type at all. We know where you are. We know where you’ve been. We can more or less know what you’re thinking about.”

Google hasn’t got to that point, yet, but it does want to save you from typing. Today, Google does this through a combination of auto-complete search suggestions, Answer Boxes, and “People also ask” boxes, which show related questions along with their “feature snippets”. As a result, Google is much less likely to achieve its stated aim of sending you to another website. According to Jumpshot research, about half of browser-based searches no longer result in a click, and about 6% go to Google-owned properties such as YouTube and Maps.

You could get upset about Google scraping websites such as Wikipedia for information and then keeping their traffic, but this is the way the world is going. Typing queries into a browser is becoming redundant as more people use voice recognition on smartphones or ask the virtual assistant on their smart speakers. Voice queries need direct answers, not pages of links.

So, I can give you some search tips, but they may not be as useful as they were when I wrote about them in January 2004 – or perhaps not for as long.

Advanced Search for everyone
Advanced Search for everyone.jpg
 Google’s advanced search page is the tool to properly drill down into the results. Photograph: Samuel Gibbs/The Guardian

The easiest way to create advanced search queries in Google is to use the form on the Advanced Search page, though I suspect very few people do. You can type different words, phrases or numbers that you want to include or exclude into the various boxes. When you run the search, it converts your input into a single string using search shortcuts such as quotation marks (to find an exact word or phrase) and minus signs (to exclude words).

 

You can also use the form to narrow your search to a particular language, region, website or domain, or to a type of file, how recently it was published and so on. Of course, nobody wants to fill in forms. However, using the forms will teach you most of the commands mentioned below, and it’s a fallback if you forget any.

Happily, many commands work on other search engines too, so skills are transferable.

Use quotation marks
4759.jpg
 Quotation marks can be a powerful tool to specify exact search terms. Photograph: IKEA

If you are looking for something specific, quotation marks are invaluable. Putting quotation marks around single words tells the search engine that you definitely want them to appear on every page it finds, rather than using close matches or synonyms. Google will, of course, ignore this, but at least the results page will tell you which word it has ignored. You can click on that word to insist, but you will get fewer or perhaps no results.

Putting a whole phrase in inverted commas has the same effect, and is useful for finding quotations, people’s names, book and film titles, or particular phrases.

You can also use an asterisk as a wildcard to find matching phrases. For example, The Simpsons episode, Deep Space Homer, popularised the phrase: “I for one welcome our new insect overlords”. Searching for “I for one welcome our new * overlords” finds other overlords such as aliens, cephalopods, computers, robots and squirrels.

Nowadays, Google’s RankBrain is pretty good at recognising titles and common phrases without quote marks, even if they include “stop words” such as a, at, that, the and this. You don’t need quotation marks to search for the Force, The Who or The Smiths.

However, it also uses synonyms rather than strictly following your keywords. It can be quicker to use minus signs to exclude words you don’t want than add terms that are already implied. One example is jaguar -car.

Use site commands

2618.jpg
 Using the ‘site:’ command can be a powerful tool for quickly searching a particular website. Photograph: Samuel Gibbs/The Guardian

Google also has a site: command that lets you limit your search to a particular website or, with a minus sign (-site:), exclude it. This command uses the site’s uniform resource locator or URL.

 

For example, if you wanted to find something on the Guardian’s website, you would type site:theguardian.com (no space after the colon) alongside your search words.

You may not need to search the whole site. For example, site:theguardian.com/technology/askjack will search the Ask Jack posts that are online, though it doesn’t search all the ancient texts (continued on p94).

There are several similar commands. For example, inurl: will search for or exclude words that appear in URLs. This is handy because many sites now pack their URLs with keywords as part of their SEO (search-engine optimisation). You can also search for intitle: to find words in titles.

Web pages can include incidental references to all sorts of things, including plugs for unrelated stories. All of these will duly turn up in text searches. But if your search word is part of the URL or the title, it should be one of the page’s main topics.

You can also use site: and inurl: commands to limit searches to include, or exclude, whole groups of websites. For example, either site:co.uk or inurl:co.uk will search matching UK websites, though many UK sites now have .com addresses. Similarly, site:ac.uk and inurl:ac.uk will find pages from British educational institutions, while inurl:edu and site:edu will find American ones. Using inurl:ac.uk OR inurl:edu (the Boolean command must be in caps) will find pages from both. Using site:gov.uk will find British government websites, and inurl:https will search secure websites. There are lots of options for inventive searchers.

Google Search can also find different types of file, using either filetype: or ext: (for file extension). These include office documents (docx, pptx, xlxs, rtf, odt, odp, odx etc) and pdf files. Results depend heavily on the topic. For example, a search for picasso filetype:pdf is more productive than one for stormzy.

Make it a date

1700.jpg
 Narrowing your search by date can find older pieces. Photograph: Samuel Gibbs/The Guardian

We often want up-to-date results, particularly in technology where things that used to be true are not true any more. After you have run a search, you can use Google’s time settings to filter the results, or use new search terms. To do this, click Tools, click the down arrow next to “Any time”, and use the dropdown menu to pick a time period between “Past hour” and “Past year”.

 

Last week, I was complaining that Google’s “freshness algorithm” could serve up lots of blog-spam, burying far more useful hits. Depending on the topic, you can use a custom time range to get less fresh but perhaps more useful results.

Custom time settings are even more useful for finding contemporary coverage of events, which might be a company’s public launch, a sporting event, or something else. Human memories are good at rewriting history, but contemporaneous reports can provide a more accurate picture.

However, custom date ranges have disappeared from mobile, the daterange: command no longer seems to work in search boxes, and “sort by date” has gone except in news searches. Instead, this year, Google introduced before: and after: commands to do the same job. For example, you could search for “Apple iPod” before:2002-05-31 after:2001-10-15 for a bit of nostalgia. The date formats are very forgiving, so one day we may all prefer it.

 [Source: This article was published in theguardian.com - Uploaded by the Association Member: Carol R. Venuti] 

Categorized in Search Engine

The new language model can think in both directions, fingers crossed

Google has updated its search algorithms to tap into an AI language model that is better at understanding netizens' queries than previous systems.

Pandu Nayak, a Google fellow and vice president of search, announced this month that the Chocolate Factory has rolled out BERT, short for Bidirectional Encoder Representations from Transformers, for its most fundamental product: Google Search.

 

To pull all of this off, researchers at Google AI built a neural network known as a transformer. The architecture is suited to deal with sequences in data, making them ideal for dealing with language. To understand a sentence, you must look at all the words in it in a specific order. Unlike previous transformer models that only consider words in one direction – left to right – BERT is able to look back to consider the overall context of a sentence.

“BERT models can, therefore, consider the full context of a word by looking at the words that come before and after it—particularly useful for understanding the intent behind search queries,” Nayak said.

For example, below's what the previous Google Search and new BERT-powered search looks like when you query: “2019 brazil traveler to usa need a visa.”

2019 brazil

Left: The result returned for the old Google Search that incorrectly understands the query as a US traveler heading to Brazil. Right: The result returned for the new Google Search using BERT, which correctly identifies the search is for a Brazilian traveler going to the US. Image credit: Google.

 

BERT has a better grasp of the significance behind the word "to" in the new search. The old model returns results that show information for US citizens traveling to Brazil, instead of the other way around. It looks like BERT is a bit patchy, however, as a Google Search today still appears to give results as if it's American travelers looking to go to Brazil:

current google search

Current search result for the query: 2019 brazil traveler to USA need a visa. It still thinks the sentence means a US traveler going to Brazil

The Register asked Google about this, and a spokesperson told us... the screenshots were just a demo. Your mileage may vary.

"In terms of not seeing those exact examples, the side-by-sides we showed were from our evaluation process, and might not 100 percent mirror what you see live in Search," the PR team told us. "These were side-by-side examples from our evaluation process where we identified particular types of language understanding challenges where BERT was able to figure out the query better - they were largely illustrative.

 

"Search is dynamic, content on the web changes. So it's not necessarily going to have a predictable set of results for any query at any point in time. The web is constantly changing and we make a lot of updates to our algorithms throughout the year as well."

Nayak claimed BERT would improve 10 percent of all its searches. The biggest changes will be for longer queries, apparently, where sentences are peppered with prepositions like “for” or “to.”

“BERT will help Search better understand one in 10 searches in the US in English, and we’ll bring this to more languages and locales over time,” he said.

Google will run BERT on its custom Cloud TPU chips; it declined to disclose how many would be needed to power the model. The most powerful Cloud TPU option currently is the Cloud TPU v3 Pods, which contain 64 ASICs, each carrying performance of 420 teraflops and 128GB of high-bandwidth memory.

At the moment, BERT will work best for queries made in English. Google said it also works in two dozen countries for other languages, too, such as Korean, Hindi, and Portuguese for “featured snippets” of text. ®

[Source: This article was published in theregister.co.uk By Katyanna Quach - Uploaded by the Association Member: Anthony Frank]

Categorized in Search Engine
Page 1 of 7

airs logo

Association of Internet Research Specialists is the world's leading community for the Internet Research Specialist and provide a Unified Platform that delivers, Education, Training and Certification for Online Research.

Get Exclusive Research Tips in Your Inbox

Receive Great tips via email, enter your email to Subscribe.

Follow Us on Social Media

Finance your Training & Certification with us - Find out how?      Learn more