Google Algorithms presentation

Submitted by:
Maitreyee Mathur 0801IT131048
Pratibha Anand 0801IT131061

Contents:
1.Introduction
2.How a search engine works?
3.PageRank Algorithm and working
4.Updates in the Google Algorithm:
i) Caffeine
ii) Panda
iii)Penguin
iv)Hummingbird
v)Pigeon
vi)Mobilegeddon

HOW DOES GOOGLE SEARCH WORKS?????
Google's search engine is a powerful tool. Without search engines like Google, it would
be practically impossible to find the information you need when you browse the Web.
•Google uses automated programs called spiders or crawlers, just like most search
engines.
•Also Google has a large index of keywords and where those words can be found.
•What sets Google apart is how it ranks search results, which in turn determines the
order Google displays results on its search engine results page (SERP).
•Google uses a trademarked algorithm called PageRank, which assigns each Web page
a relevancy score.
•So basically google does three things : 1. Crawl web comprehensively and deeply.
2. Index those pages found in web.
3. Rank those pages & return the most
relevant ones to the user.

CRAWLING:
•For crawling there exists a crawler i.e. a program that visits Web sites and reads
their pages and other information in order to create entries for a search
engine index. The major search engines on the Web all have such a program, which
is also known as a "spider" or a "bot."
•For crawling page rank is the primary determinant.
More the page rank means more people are linked to the page and the more
reputed they are, its more likely that your page is crawled earlier.
•In around like 2000, google used to do crawling for like months .
Later took 30days, then indexing for a week and then result.
Then after some time , after 1 month crawling , a recrawling was done on the high
page rankers to see if there is some change , then its updated and indexing was
done for 1 week and then result.

•Around 2003, significant chunks of web was crawled daily. And by the end of the
day each chunk was checked again for any changes. So a base main index was
maintained which got out of date till you loop back around & refresh it.
•And this process helped as we need not had to wait for the whole crawling to be
finished as we were getting incremental updates.
•This refreshing increased a lot in the coming years & so google was noticing
changes a lot faster than before.

INDEXING:
Earlier Google along with main base index had more things ; really fresh content,
layer of main index and then the supplementary index which were more like of
documents which were not changed quickly.
But now indexing has changed a lot.
Now indexing is basically taking things in word order.
e.g.:
If someone searches Donald trump in the search bar.

RANKING:
•It depends upon the page rank as well as other such 200 factors and google
balances all these factors to rank a page.
Take the same example; we will look the pages having the position of ‘Donald’
& ‘Trump’ next to each other, or the no of times it occurs or the no of links it
has to it and the content it has about it.
So any time someone searches something; that query is sent to the nearest
datacenter of google, there the search is sent to 100 different machines who
all at once look through their tiny fraction of web and then find the best
matched documents and then these results are merged so as to select &
provide the user the best document as per his/her query.
All this in just half of a second!!

Google Algorithms presentation

 The heart of Google’s searching software is PageRank, a
system for ranking web pages developed by Larry Page and
Sergey Brin, the founders of Google, at Stanford University.
 PageRank is a probability distribution used to represent the
likelihood that a person randomly clicking on links will arrive at
any particular page.
 It works by counting the number and quality of links to a page
to determine a rough estimate of how important the website
is. The underlying assumption is that more important websites
are likely to receive more links from other websites .
 Google interprets a link from page A to page B as a vote, by
page A, for page B.

How it works?
 N= no. of nodes / webpages
 Initialize PR(X)=100%/N

 Initially, PR of every node=9.1

 The PageRank theory holds that an imaginary surfer
who is randomly clicking on links will eventually
stop clicking. The probability, at any step, that the
person will not continue is a damping factor d( 0 to 1)
 It is an iterative algorithm and it continues to
calculate PR until the algorithm converges.
 PR is computed using this formula:

ABOUT CAFFEINE!!
•It was launched in 2009.
•Google Caffeine was the latest update coming to the Google search engine, but
unlike other updates, Google Caffeine was a reboot of the search engine.
•Rather than simply introducing new changes within the current system,
Google has opted to completely retool the search engine with a goal of
achieving greater speed, better indexing & a more relevant search results.
•Speed. This was the main goal of Google Caffeine. Search results were loading
up to twice as fast as previous results.
•Size. The more results that can be indexed, the better results that can be
achieved within the search result pages. Google Caffeine increased the size of
the index, with some search results pulling back 50% more items.
•Relevance. While speed and size are the easiest to test, the relevance of
Google Caffeine's search results made the biggest difference.

What Is the Panda update?
•Panda is a update in search algorithm named after the Google
Engineer, Navneet Panda .
•It was the first search filter that was rolled out In February 2011.
•It’s basically a content quality filter that was targeted at poor quality and
thin sites, in particular content farms, in order to prevent them from
working their way to Google’s top search engine results pages.
•Google could now determine more accurately which sites were “spammy”
and which sites would likely be deemed useful by visitors.
•Google said the update affected 12% of searches in the U.S.

Why Panda is called a update ?
• Panda is called an update because the filter runs periodically. And, every
time it runs, the algorithm takes a new shape.
•There has been a Panda update every 1-2 months since 2011, for a total
of 26 updates since February 14, 2011.
• Panda 1.0 update: The search filter was aimed at content farms — those
sites that employ many writers who create poor quality content around
specific keywords in a bid to rank in Google’s top ten results on a rank
tracker.
• Panda 2.0 update: This update, released in April 2011, was targeted at
international search queries, though it also impacted 2% of U.S. search
queries.

Panda 2.1 – 2.4: There were minor updates in May, June and July 2011 in which
Google incorporated more signals to help estimate the quality of a site. Low quality
content and pages were further penalized.
Panda 2.3: It was minor and focused on the user’s experience.
Panda 2.4: This update was rolled out in August 12th, 2011 and affected about 6-9% of
user queries. It was focused on improving site conversion rates and site engagement.
Panda 3.0: This SEO power search filter began to roll out on October 17, 2011 and
was officially announced on October 21, 2011. This update brought large sites higher
up in the SERPs – e.g., FoxNews, Android.
Panda 3.1 went live on November 18, 2011. It was minor and affected about 1% of all
search queries. Although this number seems low, it is still significant for any rank
tracker, considering the number of searches conducted every day.

Panda 4.0: On May 20, 2014, Matt Cutts tweeted that Google was rolling out
Panda 4.0. It was the next generation of Panda and it generated both winners and
losers. This update had a big impact.
Ebay lost a significant percentage of the top 10 rankings it had previously enjoyed.
Panda 4.0 was primarily targeted at larger sites that have been dominating the top
10 results for seed keywords.

What was the impact of panda ?
•The sites with large amount of advertising dropped in ranking.
•The sites with duplicate content, or those which copied content from other
sites were ranked lower while the ones with the original content were
ranked higher.
•The sites with poorly written content like those having spelling mistakes or
grammatical mistakes were ranked lower.
•The sites having a lot of links i.e. inlinks or outlinks were ranked low.
GOOGLE provided an advisory on its blog for the users for providing them
with some direction for self evaluation of their website’s quality in some 23
bullet points.

What to do and what not??
•Make the content of your site original if its not.
And for that we can use “copyscape” a tool for duplicate content detection
software.
•Make sure the content is not redundant.
•Don’t stuff your pages with a lot of keywords.
•Write the content for your site but not for the search engine.
•Try to get your content be written with the help of a proffessional having
good writing skills so as to avoid spelling and grammatical mistakes.
•Make your site easy to navigate.
•Keep a check on the time your site takes for loading. The one that loads
faster has a better rank.

Misconception with PANDA
I) It affects the site only.
But its later updates looked the page by page content and then ranked
the pages and in extreme case the rank of the site gets changed.
II) Panda is a penalty .
Its not since its whole aim is to increase the google quality and the
content we find there.

What is a penguin update ??
•On April 24, 2014, Google released the first Penguin update.
•While the Panda update was primarily targeted at thin and low-quality content,
Google Penguin was a set of algorithm updates that puts more focus on incoming
links.
•Google Penguin is a codename for a Google algorithm aimed at decreasing search
engine rankings of websites that violate Google's Webmaster Guidelines by using now
declared black-hat SEO techniques involved in increasing artificially the ranking of a
webpage by manipulating the number of links pointing to the page. Such tactics are
commonly described as link schemes.

Factors which are watched by the PENGUIN !!
There are four factors which will get your site penalized by Penguin. See the image
below:

1) Link schemes
•Links are still important, but high quality sites are using this as the best way to
improve search rankings.
•Link schemes are those types of activities geared at generating links that will
manipulate or induce search engines to rank your web pages. If you fall into the
trap of always building links to your site from every other site found in search
engines, you may be penalized by Penguin.
•A dedicated website that interprets lyrics and poetry, was penalized, because
Google found they were using link schemes to manipulate their rankings at the
time.

2) Keyword stuffing :
Matt Cutts already warned against stuffing your page with relevant keywords. No
matter how in-depth and easy-to-navigate the site is, Penguin will most likely find
and penalize it. In most cases, it’s easy to see why, especially if you’ve ever actually
seen a keyword-stuffed page.
Here’s an example:
Buying valentine’s gift for your spouse is a great step to take. This
year, valentine’s gift should be an avenue to express how much love you’ve for
him or her. Make sure the valentine’s gift is well-researched. But don’t stop
there. Make it a culture to always show love to your spouse, whether there is
valentine celebration, Christmas etc. When you show love, you get love. For
instance, when you show love today, you’ll live to be loved. Are you ready to
choose the best valentine’s gift?
That’s keyword stuffing and it’s contrary to the Google Webmaster Guidelines.

3) Over-optimization :
SEO is awesome, but too much SEO can cause over-optimization. If you over-optimize
your anchor texts, for example, this could get you penalized by Penguin. The best
approach is to incorporate social media marketing and gain natural or organic links to
your web pages.
One of the signs that you might be over-optimizing is having keyword-rich anchor texts
for internal links, i.e., anchor text that links to a page within your own web pages.
Here’s an example:
Learn more about Hp Pavilion 15 laptops and its features.
(Links to: example.com/hp-pavilion-15-laptops.htm)
Another example:
Do you know the best iPhone case that’s hand-crafted for you?
(Links to: example.com/best-iphone case-hand-crafted)

4)Unnatural links :
The funny thing about unnatural links is that they don’t look good to anyone – not
to your readers and not to Google. These links may appear on sites that are totally
off-topic. E.g. Cardstore lost their ranking through unnatural links that appeared in
article directories.
The moral of the story: Your links should be natural.

CONFIRMED PENGUIN UPDATES
Penguin 1 on April 24, 2012 (impacting around 3.1% of queries)
Penguin 2 on May 26, 2012 (impacting less than 0.1%)
Penguin 3 on October 5, 2012 (impacting around 0.3% of queries)
Penguin 4 (a.k.a. Penguin 2.0) on May 22, 2013 (impacting 2.3% of
queries)
Penguin 5 (a.k.a. Penguin 2.1) on October 4, 2013 (impacting around 1%
of queries)
Penguin 6 (a.k.a. Penguin 3.0) on October 17, 2014 (impacting less than
1% English queries). On December 1, 2014 Google confirmed that the
update was still rolling out with webmasters continuing to report
significant fluctuations.
Penguin 7 (a.k.a. Penguin 4.0) on September 23, 2016

•The Panda update is primarily concerned about quality content, while Penguin wants
to drown spammy or aggressive links that strive to manipulate any search engine rank
tracker.
•It’s important to keep an eye on both updates. When your site is penalized by Panda,
there’s a good chance that Penguin will affect your site, too.
•Even if your webpage contains unique, useful and in-depth content, the Panda
update likes you but you’ll still face a Penguin penalty if your links are low-quality.
PANDA
V/S
PENGUIN

 When did Google release it?
Google started using Hummingbird about August 30,
2013, and announced the change on September 26 on
the eve of the company's 15th anniversary.
 Why is it called Hummingbird?
This Google algorithm derives it’s name from the tiny
bird, which is known for being “precise and fast.”

 Why was it important?
 Panda, Penguin and other updates were changes to parts of the
old algorithm, but not an entire replacement of the whole.
 The Hummingbird was an entirely new algorithm.
 It approaches search engine queries in a brand new and
intelligent way utilizing new technology combined with older
features of the existing algorithms.
 Biggest change since 2010, when Caffeine Update was released.
Since then, Google only updated it’s algorithm but never
changed it.
 Hummingbird was important because Google Algorithm was
dramatically rewritten.

 How was it different ?
 Hummingbird is about synonyms but also about
context. Google always had synonyms, but with
Hummingbird it is also able to judge context—thereby
judging the intent of a person carrying out a
search, to determine what they are trying to find
out. This concept is called semantic search.
 Semantic search seeks to improve search accuracy by
understanding the searcher's intent and
the contextual meaning of terms to generate more
relevant results.

 Is Page Rank dead?
No.
Page Rank is one of over 200 major “ingredients” that
go into the Hummingbird recipe.
Hummingbird still depends on PageRank to decide
how important links to a page are deemed to be?
But this is along with other factors like whether Google
believes a page is of good quality, the words used on it
and many other things

Main Characteristics
 1. Knowledge Graph:
 a knowledge base used by Google to enhance its search
engine’s search results with semantic search information
gathered from a wide variety of sources.
 Google has been up and running for 15 years (at the time
of Hummingbird release), and since then, it has been
collecting what could very well be the largest database of
knowledge ever recorded.
 This knowledge Graph came into view after Hummingbird
release and was extensively used by Google to understand
the relevance and context of the searches.

 The Resurgence of Long Tailed Keywords
 The Hummingbird will take a search engine query using
long-tailed keywords and try to decipher the context of the
question rather than chase the specific keywords within
the question. The goal is to provide results that actually
answer the question
 Contextual and Conversational search
 In Contextual search, Google increasingly returns results
that match the intention behind the query. Results are no
longer limited to the words themselves, but include an
interpretation of intent behind the search.

 EXAMPLE:
 A search for “pay my bills through HDFC Bank”
used to bring up the home page for HDFC Bank but
now should return the specific page about paying bills
 Google interprets the query and returns an answer in
the search results as a Knowledge Graph integration

 Conversational search is a new kind of philosophy for
human/computer interaction. The principle
behind conversational search is that a user can speak a
sentence into a device, and that device can respond with a
full sentence.
 Example: “What’s the closest place to my home,to but an
iPhone?”
How Hummingbird will interpret it?
 It may better understand the actual location of your
home, if you’ve shared that with Google.
 It might understand that “place” means you want a brick-
and-mortar store.
 It might get that “iPhone ” is a particular type of
electronic device carried by certain stores.

Information card
 Hummingbird is supposed to answer the search
question.
 It does this by giving us an answer in the form of an
“Information Card” taken from its Knowledge Graph
and displaying it right at the top of the results page
either above or to the right
 Example:
 “Does Barrack Obama have children?”

 There are many
wonderful websites,
that want to give us
that information and
while they are at it,
show us what
wonderful surprises
they have to offer.
 Unfortunately,
Google has already
answered my
question in the
Information Card.
Why should I bother
to click on those
links when I have the
answer right here in
front of me?

 Google Pigeon is the code name given to one of Google's
local search algorithm updates.
 The update is aimed to increase the ranking of local listing
in a search.
 This update was released on July 24, 2014
 What was new??!!
 To improve the quality of local searches and provide more
relevant results to the user, Google relies on factors such as
location and distance.
 This update alters the local listings in the search results;
along with this, the local directory sites like Yelp are given
preference.

 Pigeon improves their distance and location ranking
parameters.” Increased specificity affected dense
neighborhoods, also known as “informal space.”
 Neighborhoods are known as “informal space,” meaning there
are no standard boundaries. So the definition of where a
neighborhood starts and stops can be different for everyone.
This makes it hard for services like Google to get the
neighborhood thing right.

 How does Pigeon help?
 The algorithm will return better results for queries that
use both the conventional term for a local
neighborhood and the colloquial term for the same
neighborhood.
 The algorithm will provide local results for areas that
are slightly outside of a searched-for local
neighborhood

When I just type ‘rest’,
 restaurants is the first suggestion that pops up
 Restaurants in Indore is also among the top
suggestions by Google.

What
happens
when I
simply
search for
restaurants?

 It refers to Google's algorithm update of April 21, 2015
 The main effect of this update is to give priority to web
sites that display well on smartphones and other
mobile devices.
 The change does not affect searches made from a
desktop computer or a laptop
 It was labeled as an update that changed the world of
web from the prospect of users as well as webmasters.

What makes a webpage mobile-friendly?
 Avoids software that is not common on mobile devices,
like Flash
 Uses text that is readable without zooming
 Sizes content to the screen so users don’t have to
scroll horizontally or zoom
 Places links far enough apart so that the correct one
can be easily tapped

What is known of Mobilegeddon update?
 It only applies to individual pages, not entire websites. So if you
only have one page that is mobile friendly, only that page will do well
in mobile search.
 It affects searches on mobile devices across all languages and
locations.
 One of the best ways to combat it is to switch to a responsive
website.
 A responsive website displays content differently depending on the
device used to view it, which produces more relevant layouts which
optimize the end user’s experience and also looks much better.

Age of Artificial Intelligence:
 Google uses artificial intelligence (AI) to better understand
search queries. A "very large fraction" of queries is being
interpreted by an AI system called RankBrain.
 Expect more accurate search results
 With machine learning, RankBrain better understands
what users are searching for. It has improved 54.6 % of
difficult to understand queries.
 Using AI, Google handles the increasing volume of search
traffic that comes as a byproduct of increasing Internet
access.
 RankBrain also contributes to page ranks.

Google Algorithms presentation

More Related Content

What's hot (20)

Similar to Google Algorithms presentation (20)

Recently uploaded (20)

Google Algorithms presentation

Editor's Notes