Google ranks pages by using an algorithm to search its massive database of everything it can find on the internet, and find pages that seem relevant to each individual search. For popular searches, it will have cached the results (or some table of top results at least) so that it can serve them even faster. For niche searches, it may work them out on the fly.
Even so, it has enough computing power at its disposal that any search result is returned in a fraction of a second.
It’s always less than a second. Try it.
How Google decides what order to show pages in is obviously more complicated. It’s not just about which web pages are most relevant to a search. It’s also about showing pages that most satisfy the user intent behind each search. It can tell if a user is satisfied by looking for long clicks.
(A long click is when someone searches, clicks on a link on the SERP, then doesn’t do that search again for a while. The implication is that the link satisfied what the user wanted.)
How do they guess your intent though? Well, they make an educated guess by seeing what is popular with people like you. By measuring the popularity of different answers to searches in different contexts they can do pretty well at knowing what will satisfy your searches.
This works well, mainly because you’re simply not as unique as your momma might have told you (not when it comes to clicking on things anyway).
Despite what I just said, we should probably start by discussing relevancy anyway. It’s one of the things that website owners are most in control of, and for many searches, it is still the key.
If someone searches for “Penguins sleeping patterns” then Google will try to show pages on that topic by searching its massive database of the internet. What is it searching for though? Here are the most important things:
- A page title that contains that term
- The term being used early on the page (in the first paragraph preferably)
- An image or video with the search term as its title or alt-tag
- The term being used in the meta description of the page (the snippet of test that appears in Google search under each link)
- A sub-heading containing that term (if there are any sub-headings on the page)
- The search term is used throughout the page (although not too much)
- Links to the page from other pages which are relevant to that search. (Links from irrelevant pages make no difference, or can actually hurt ranking)
Why you may ask, does it look for these things? It’s because pages that have these elements have, on the whole, proven themselves to be more relevant to searches than pages without them. Google has a small army of testers trying out different searches in order to poke holes in the algorithms it uses, and over time these factors have consistently proven themselves.
The reason why these factors have proven themselves is two-fold:
- If someone made the best page they could on a subject without knowing anything about SEO, then this is the behaviour they would probably exhibit.
- If you visit a page with all of these things, then you will probably think the page is about the search term you used.
These factors don’t just imply relevancy, they demonstrate it. If you want Google to think your page is relevant to a specific search, make sure it has all the above factors so that it actually is relevant.
“Keyword stuffing” refers to forcing keywords into a page in order to get that page to rank for that keyword. It also means you’re a jerk trying to game the system. Instead of keyword stuffing, you should just write naturally.
Google can understand synonyms so you don’t have to force anything into anywhere. Just make the best page you can on a topic – the above factors tell you how to do that.
If I search for “FIFA World Cup” there are millions (billions?) of pages that are exceptionally relevant to this search. However, if the World Cup is happening at the moment, it’s likely that when I search for “FIFA World Cup” I want to see the latest scores.
If the World Cup has been recently plagued by a scandal, then it’s pretty reasonable to assume I want to read about that scandal. So it’s not just that Google looks at the words we’re searching for. It also takes into account the context of what is happening in real life, right now.
It was billions.
Google does this by seeing what sort of web pages are very popular at the moment. It also checks what sort of related searches are popular at the moment. A spike in traffic for the search “FIFA scandal” will help Google’s algorithms to understand that right now people interested in searches containing “FIFA” might be interested in results for a FIFA scandal.
It will constantly test these results too. So if people are clicking on web pages about the FIFA scandal, then it will rank them higher. If not, their ranking will drop. Over time these links will drop in the rankings anyway as newer news is generated about FIFA.
One way Google knows whether something is recent is by the timestamp on web pages. This means when a page was published and when it was last updated.
To improve your rankings, make sure you publish and update pages on your site frequently (15 times a month at least ideally). If you’re not doing this as standard, then your site probably isn’t that currently relevant anyway.
Similarly to recent events, the context of a search changes significantly based on where you are. If an incredible event happens somewhere in the world, every world leader will potentially put out a statement about it. However, you’re probably not interested in them all equally. If you are in France, and the leader of Peru has made a statement – it probably doesn’t mean that much to you. At least not compared to the leader of France’s statement anyway.
Again, however, Google will test this. If people in France are flocking to the Peruvian leader’s speech (and linking to it lots), then Google will promote links to the Peruvian leader’s speech to French users.
As Google can also see your browser settings when you’re making a search, they’ll also know what language you have set as your own. Therefore it will only show you links in your own language. This naturally means that web pages based in countries that speak the same language as you will get a boost.
These two things together mean that your location is one of the most important factors in Google ranking search results.
If your website has a physical presence somewhere (ie a shop or local business), then register with Google Business Profile.
GBP is the tool that helps you manage your presence on Google Maps. Using it can boost your rankings when people are searching locally for you.
The time of day you search for something also demonstrates a huge difference in user intent. If you search for “croissants” first thing in the morning, you are probably looking for a cafe or bakery. If you search for “croissants” late at night, you are probably looking for a recipe or doing some online grocery shopping. Simple as that.
Too simple? Ok, let me complicate it for you… My guess is that Google, having developed an algorithm to understand user intent at different times of day, have discovered some truly odd things. I don’t mean that if you search for “cocktails” at 6am you’re an alcoholic, but if you search for “cocktails” at 6pm you’re a legend.
I mean some deep anthropological stuff, like if you search for “socks” between 3-4pm then you’re never going to earn above $50k.
Definitely 6AM cocktail searcher – via GIPHY
It’s likely that Google will never tell us what they’ve found though, so don’t sweat it. They might not even know what they’ve found themselves to be honest. This is the point. Google likely uses many algorithms to decide what is relevant based on extra factors, and while timing of searches is understandable to humans, they might not all be.
After all, Facebook shut down their AI experiments after two of them started talking, and made up their own language which only they understood. It’s not unthinkable that Google has an algorithm making AI that humans don’t fully understand. If this AI is improving their search results though, they might just let it run and run.
Was that too much? My apologies. Speaking of robots, however…
The first form of SEO was Technical SEO. It simply meant making it so that Google could index your website. Google has unsurprisingly become smarter since the 1990s however, so it can read and index almost any website now. This means that this form of SEO is less important than it used to be.
However, there are many ways in which Google judges websites based on their technical setup rather than on what is on a page or user intent. The main ones are:
- Page Speed – webpages taking ages to load makes people mad. Google doesn’t want to make people mad so it penalises slow pages in its rankings.
- Mobile Friendly – most people search via mobile phones these days. If a webpage works badly on a mobile phone then it’s going to go down in the rankings.
- Sitemap/Search Console – Google Search has a way for websites to gets stats on search, find out about problems, and feed info directly into Google. It’s called Search Console, it’s free, and you should register your site on it immediately. One of the most important features in Search Console is the ability to register a sitemap. Doing this helps Google to know about every new and updated page on your site (in real-time). Being registered on Google Search Console won’t give you a bump in rankings. It will however help your pages get noticed quicker and get penalised less.
- No funny business – if your site is set up in some nefarious manner Google will find out. If your site is stealing users’ data Google will know. Google can tell if you are putting ads on top of content so you can’t help but click on them. Google hates it when you use code to hide where links go. If you have a dodgy site – Google will find out, and they will penalize you. They have a very specific set of skills, and spotting bad websites is definitely one of them. Sending people to dodgy websites would hurt their brand infinitely so don’t think they won’t find out.
Google knows a lot about you. They know a lot about me too. For a long time, they were using that data as much as possible to give each individual the search results most relevant to them.
However, in 2018 they announced that they weren’t doing that much anymore. Apparently, it didn’t make that much difference, it was a lot of effort, and it was just… creepy. So Google pretty much stopped doing it. They’ll still use your recent previous searches, your location, and the time of day, to make decisions about what to show to you personally, but not much else.
They almost certainly still have all that data about you (which you can find out by downloading it all from here btw) but they don’t use it now. I personally expect they still use some of it sometimes, but as a whole, not so much.
Search Engines rank pages by:
- Indexing every webpage they can find so they can be searched faster
- Using what is on those pages to work out the relevancy to individual searches
- Using how those pages are made to decide if people will like them
- Seeing if other pages deemed relevant to a search have linked to them
- Seeing what is popular for that sort of search at the moment
- Checking what is popular for that sort of search in a users location
- Using other inscrutable algorithms to work out what people want to see
And hundreds of other less important factors. If you want your web pages to rank better in search engines, simply make the best web pages that you can. That is always the secret key to success in SEO.
A good page for people is a good page for search engines.