Mon – Fri: 08:00am – 4:00pm, Sat-Sun: Closed

PC MADNESS
PERTH IT SPECIALIST


Cloud Services, Business IT Support, Web & Design Services,
Computer Store - Sales & Service for PERTH WA.

How search engines work and what to do about it (do SEO)

Written by Benjamin Denis

Search engines including Google, but also including search engines on sites such as Amazon, Netflix or YouTube and search on your computer or email software – work to the same principles:

  • The content you want to search is indexed first
  • A user enters a search query as words
  • The search engine returns results from the index that best match the query
  • The results are ranked with the intention of putting the most relevant results at the top of the results

If you are a WordPress site editor in almost any country of the World, your main interest in search engines – and for reading this article – is ranking on Google. Always remember, though, that other search engines exist, and that Google may not always be the most popular search engine.

By learning how search engine work, you can optimize your site for better ranking and more traffic!

Unique marketing opportunity

Visibility in a search engine can generate traffic, but it is also a unique marketing opportunity: being visible in the very moment that a person is interested in finding you, your product or your content. If you have an Indian restaurant in Seattle, for example, then being visible in Google when people in Seattle search “Indian restaurant” is an amazingly effective way of earning a new client. With most alternative marketing tools – such as handing out flyers on a street-corner – you are pushing messages out to people who have not asked for this information and there is much less chance of converting them to clients. There is even the possibility you may annoy one of them and lose a potential client!

Google does not search the web

From early search solutions used to find files on a computer, it became obvious that opening and reading each file on the computer, one after the other, was not an efficient way to find matching results. As the storage capacity of computers increased searching directly through all the files on a disk was too slow and it also increased the risk of damaging fragile disk-drives.

Did you know? Because search was not an option, early computer users were encouraged to organize files in directories and sub-directories. Applied to the web, this led to the creation of Internet directories such as Yahoo or DMOZ whose purpose was to organize the World’s websites in a directory structure. In the 1990s these sites were much more popular tools than search engines. DMOZ was even integrated into Google as Google Directory from 2000 to 2011.

Search engine software uses data stored in indexes to help it find results more efficiently. An index is part of a database that stores information in a format that makes it quicker to search. For example, rather than searching in multiple files for a word, an index will organize data by word: Listing all the files that contain a specific word in one index.

Google created and maintains its index by crawling the web. When it discovers a new page, it indexes the content of that page. It then regularly goes back to check to see if that page has changed – and if so, it indexes it again. When you search on Google you are searching Google’s index and not on the web itself. It is therefore possible that Google proposes a link to content that is no longer online, or that it omits content that is online but has not yet being indexed.

Pause for a moment and think of how big Google’s index is and how big a job it has keeping it up to date. Google says it crawls and indexes hundreds of billions of web pages (and as reported by Kinsta, WordPress powers 37% of all the websites on the Internet which is a cool thought). The index is well over 100,000,000 gigabytes in size and requires large industrial buildings to house the hardware that store the data. Crawlers, which are computer programs, are regularly re-visiting and re-indexing millions of web pages every day.

One of Google’s data centers (Dublin, Ireland)
One of Google’s data centers (Dublin, Ireland)

Keyword queries

Searches are performed using queries that are expressed in words. You can type these words into the Google search bar or use voice recognition instead of typing.

Different people have different methods for searching Google. For example, a person in Seattle looking for a restaurant may just type “restaurant” into Google (knowing that it will provide only local, relevant results), some may type “Seattle restaurant”, others “restaurants in Seattle” and so on. Someone, somewhere, may type – or say –  something like “please can you suggest a nice indian restaurant for me in downtown Seattle please, thank you very much” and they all get quick, relevant results (which proves that it pays to be polite!).

Google Search

Google is quick at returning the results. We can therefore deduce that it is not searching through all the pages it found on the web to find pages that match the keywords you typed. It is possible that it has indexes for popular searches (this means maintaining a list of URLs that correspond to exact multi-word queries). This is certainly the case for “Restaurants in Seattle”. When a user types this query, Google already has a list of relevant pages that answer that question.

However the query ““please can you suggest a nice indian restaurant for me in downtown Seattle please, thank you very much” has probably never been searched before and it is unlikely that Google has a pre-prepared a list of relevant pages for this query. This is not a rare occurrence, Google indicates that every day, 15% of searches are for queries that have never been searched before.

In this case Google breaks down the query to the most relevant words and expressions to match it to its different indexes. Over the years, Google has got better at interpreting these more complex queries and currently does it better for other search engines in most languages.

Algorithms to rank results

It is believed that for any query, Google first creates a list of only 1000 pages and then ranks these from 1 to 1000, eliminating some results along the way. There is an algorithm for finding the best 1000 pages and another for ranking these 1000 results from 1 to 1000.

SEO expert Jason Barnard has a theory that Google has many algorithms that calculate different ranks for each page based on different signals (topicality, quality, page speed, etc). It then multiplies them together to calculate a bid. The page with the highest bid is ranked highest. He also points out that because scores are multiplied, a very low score for any factor can significantly reduce a bid and the page’s rank.

Illustration from Jason Bernard’s article “How Google Search Ranking Works” in the Search Engine Journal
Illustration from Jason Bernard’s article “How Google Search Ranking Works” in the Search Engine Journal

This is only educated guess-work though. The details of how Google works exactly are a closely guarded secret and Google remind us that they change and improve the algorithm every day.

The main factors, though, are certainly the use of words in individual pages on your site and the links that those pages receive from other pages on the web (whether they may be links from your own site or links from pages on other sites). Minor factors, but potentially with a big impact for some keywords, may be linked to user experience, page speed, social networks, and online reputation.

After ranking the pages, Google may also decide to filter out some results. This could simply be to reduce the number of results from one site, but it could also be filters against spam or adult content.

Usually less than a second after sending your query, you get results back from Google.

Also bear in mind that search results are also personalized to your geographic location and your search history. These factors can make a big difference on search from one computer to another.

SERPs and snippets

For each query, Google returns results as a SERP (Search Engine Results Page). It is very rare that Google finds no results for a query.

As well as results from the web search, Google may add advertising from Google Ads and results from other Google search engines such as Google News, Google My Business, Google Images, …

The results from web search are typically listed as 10 results per page. Each result has a link to as web page and a short description. This result is called a Snippet. The Snippet may also be enriched with the websites’ logo, reviews, images, prices and other useful information.

Simple snippet from desktop SERP
Simple snippet from desktop SERP
Enriched snippet from mobile SERP
Enriched snippet from mobile SERP

The information for the snippet can be provided by the website itself using specific tags such as KEYWORD, META DESCRIPTION or Schema.org structured data. The quality of this information may influence Google in ranking your page, but it is also important in influencing the users who are looking at the results to click on your snippet and generate traffic. You may only rank third but have a very eye-catching snippet that means that most searchers will click on your link rather than the first two results.

In 2013, Dr Peter J. Meyers did a mock-up of 24 different SERP features. Although it is a bit old now, it is still a good introduction to the variety of results possible in Google.

Serp Features
Serp Features

What is SEO?

SEO is an acronym for Search Engine Optimization. This is the work you do to optimize your website for search engines – so that you can get listed (for free) in Google when there is a relevant search for your site.

From better understanding how search engines work, you can see that the main tasks in SEO are:

  • Make sure that Google can index your site correctly, can find all your content and links between pages. Providing a sitemap is a good idea here,
  • Understand what queries your target audience use on search engines and make sure that you have pages or posts that use the keywords of those searches and answer the question the search engine was asking. Most SEO plugins give you tools to analyze pages for keywords,
  • Ensure that tags important for SEO are correctly completed. Most importantly TITLE and META Description tags. You will need a SEO plugin to do this,
  • Where relevant, provide Schema.org structured data to describe your site or the content of your pages to obtain rich snippets. Again, using a SEO plugin that allows you to add relevant Schema to your posts and pages to gain rich snippets,
  • Develop external links from other web sites,
  • Analyze your ranking in search engines and the traffic you receive from them.

SEO is also used in another definition to mean Search Engine Optimizer, i.e. the person who is going to do the work above.  If this person is you, then continue reading our articles to learn more about getting better visibility and more traffic for your WordPress site!

Written by Benjamin Denis

Latest Google Updates Q3 2020

Written by Benjamin Denis

Summer passed and Google did not release a major update of their algorithm as we have learned to expect. Then, in the last week of the Quarter (officially in Fall though) SEOs reported major fluctuations in search results and claimed to have spotted a major update! This turned out to be a bug in Google’s index and is still being corrected as we write this post.

In fact, Google had 3 indexing problems in the last week of September that may have affected your site’s visibility and traffic. We will have a look at these first and then go through the other news from Google from July to September 2020.

Canonical Indexing Bug

Reports of major changes in ranking started around September 23rd with, in parallel, reports of users seeing odd values for canonical URLs in Google Search Console. Canonical URLs are Google’s way of treating duplicate content: when many pages have the same content, Google choses one page as the canon (i.e. original version) and will show that in search results rather than its copies. You can also specify your own canonical URL in the HTML code of pages. See Google’s guide on consolidating duplicate URLs.

The index of canonical URLs got mixed up around September 22nd and it appears that completely random URLs were being used as the canonical versions for pages in the same site for a part of the index. The bug was being slowly corrected by October 6th. It affected ranking for thousands of sites.

It seems rare to have this sort of massive error in Google, but something similar also happened in April 2019.

Google News Indexing Bug

From 6:45pm to 8:45pm ET on September 28th Google News did not index any news stories – although obviously many news sites did publish new stories during these hours. The bug was revealed on the Google Search Liaison account on Twitter @searchliaison.

Google SearchLiaison Tweet
Google SearchLiaison Tweet

This is an interesting account to follow for information on Google updates and, now, bug reports.

Mobile-Indexing Bug

In parallel to the canonical indexing bug, Google also reported problems with indexing mobile versions of pages. Through the Google Search Liaison twitter account again, they communicated on October 2nd: “The mobile-indexing issue impacted roughly about 0.2% of our index, beginning in early September but really spiking from around the middle of this week through late yesterday. We’ve since restored about 1/4 of those URLs & keep reprocessing more”.

In March, Google announced that it would be moving to mobile-first indexing in September 2020, but this deadline was pushed forward to March 21st.

Ranking on Google Discover: New Guidelines from Google

In July Google published new guidelines on how to feature in their popular Google Discover service used mainly on smartphones through the Google Search application. Any website can be listed on Google Discover, there is no need to be approved for Google News, but it will mainly index fresh posts with a clear publication date, and information about the publisher and author (similar to Google EAT recommendations) and will filter-out click-bait titles. Having high-quality images, at least 1200 px wide is very important, and you should enable the max-image-preview:large setting (this is enabled by default in SEOPress).

Read the full Guidelines from Google

New Guidelines From Google
New Guidelines From Google

Virtual Webmaster Unconference WordPress session

Still unable to organize physical events Google’s Aurora Morales & Martin Splitt (pictured below) organized an online unconference for webmasters on August 26th. Unconference, not because it was online, but because it was organized around interactive sessions with webmasters rather than presentations by Google staff as seen in previous Webmaster Conferences. One of the 17 sessions concerned Google’s plugin Site Kit for WordPress, the roundup from Google simply stated that this “showed that users were confused about data discrepancies they see between Analytics and Search console in the plugin”.

The succinct roundup of all 17 sessions can be found here.

Aurora Morales & Martin Splitt working hard to organize
Aurora Morales & Martin Splitt working hard to organize the Unconference

Googlebot will work over HTTP/2

In September Google announced that its crawling robot, Googlebot, will use HTTP/2 to connect to websites where this is possible and when it deems it useful to do so. HTTP/2 is useful for Google because it increases the speed at which pages are downloaded. Whereas this may help some sites get indexed quicker and more often, having HTTP/2 is not considered as a ranking factor.

Although HTTP/2 was adopted by most browsers as early as 2015, a lot of webservers still do not use it. You can test if your site “speaks” HTTP/2 by using KeyCNDs HTTP/2 Test.

It is generally advised that you should upgrade to HTTP/2 for SEO because it should improve page speed as experienced by users on Chrome – information that Google does use as a ranking factor. Changing from HTTP/1.1 to HTTP/2 is a change to the server software, done by your hosting provider. No changes are needed to your website.

The news here is more the surprise that Googlebot did not use HTTP/2 before 2020. John Mueller, Google Webmaster Trends Analyst, had already stated in a Webmasters Hangout in November 2015 that he expected it to be ready by the end of that year, or early 2016.

Local Ranking Factors 2020

Whitespark have taken over the mantle from David Mihm and Moz in producing the Local Ranking Factors survey. By asking a series of questions to several Local SEO experts, the survey gives interesting insights into how Google’s local ranking algorithms work – specifically for two types of results: the Local Pack (supplied by Google Maps) and the organic results.

Results published in september show an increased importance for Google My Business for ranking in the Local Pack, but less importance for ranking for localized organic search.

BrightLocal published a long review of the 2020 results: Local Search Ranking Factors 2020: What Affects Local Rankings?

Extract from the Local Ranking Factors report
Extract from the Local Ranking Factors report

Unfortunately, no question in the survey deals specifically with the question of Schema.org structured data as a ranking factor. But On-page factors are important. In a separate article on How to Optimize for Local Search, Brightlocal do cite “Use Local Business, Organization, Product and Service Schema” as one of the 4 things you need to do on-page to optimize for local search.

If you haven’t added structured data to your WordPress site yet, read our articles on How to add schema markup to WordPress using the SEOPress plugin.

Google launch Web Creators and a new WordPress plugin

In September Google launched an interesting initiative that they say is “A community for web creators to grow and get inspired”. A web creator is anyone who creates content for the web and that includes bloggers.

Google has launched Web Creators as a blog and also dedicated accounts on TwitterInstagram and YouTube.

One of the first announcements on this new platform was the official release of the Google Web Stories WordPress Plugin. This has been shortly followed by the news that Google Discover now features a Web Stories carousel.

One of the first announcements on this new platform was the official release of the Google Web Stories WordPress Plugin. This has been shortly followed by the news that Google Discover now features a Web Stories carousel.Video Player

PC Madness Trading Hours​

Monday: 8:00am – 4:00pm

Tuesday: 8:00am – 4:00pm

Wednesday: 8:00am – 4:00pm

Thursday: 8:00am – 4:00pm

Friday: 8:00am – 4:00pm

Saturday: CLOSED

Sunday: CLOSED

Public Holidays: CLOSED