How do search engine algorithms work?
The mystical weapons of ranking.
This is an opaque topic that we hope to illuminate and make accessible to non-technical readers.
There's a point that's important to keep in mind, and it is as simple as this: there is a hierarchy to page ranking.
From the top down it looks like this:
Your domain authority is determined by the sum score of the different pages rankings on your website.
The ranking of each individual page is determined by the sum score of different factors.
In order to score the individual sections of a page on your website, Google uses smaller, more specific algorithms.
In order to establish the overall domain authority of a site, these different scores (keep in mind we're simplifying this) are plugged into a larger algorithm to determine overall trust and authority.
When you're working on a page for your website, you probably think about things like:
How can I get this paragraph on a featured snippet?
What kind of markup/structured data can I use to rank this page number one?
While it's important to ask these questions, it's even more paramount to ask this question:
What structure/ranking factor am I operating in?
It's impossible to win the game if you don't understand the rules.
So let's discover the rules.
What's an algorithm?
Who better to ask than Google?
"A process or set of rules to be followed in calculations or other problem-solving operations, especially by a computer."
So what isn't an algorithm?
An algorithm is not a formula.
While this seems picky, understanding this difference is critical to your success in driving organic traffic.
The easiest delineation is this:
A formula is a specific step in a larger set of instructions (the algorithm).
Let's use a sports analogy.
You coach a football team and you have several pass plays.
These plays are clearly marked in the playbook. Each player knows where they should go, when they should go there, who they're blocking, etc.
These are formulas. Specific steps that result in specific outputs.
You have two minutes left on the clock in the 4th quarter with no timeouts left. You must find the order and cadence of these pass plays that will allow you to get in field goal position with minimal time left on the clock.
You're actively creating an algorithm (a specific arrangement and implementation of the plays) to solve this problem.
The specific plays don't change throughout a game (disregarding audibles and such), but the order and cadence in how you use them do.
This is the algorithm: assembling the specific formulaic steps to achieve a particular outcome.
Algorithm Inception
So what does football have to do with search engine algorithms?
Take a look at some of the fundamental components of a website:
Internal and external linking
URL structure
Content
Images and video
Page load speed
As discussed earlier, each of these areas can be further subdivided and scored using formulas and algorithms.
Each of these subdivided systems is then fed into a higher-level algorithm that processes the sum of its parts.
That higher-level algorithm is what ultimately determines where you land in the search engine results pages.
Always remember this hierarchical relationship. Each smaller factor is fed into an increasingly broader-scope system until the final rank for a webpage is determined.
When people say "The Google algorithm", what they really mean is the enormous collection of algorithms and formulas that work together in producing a page’s SERPs rank.
So what kinds of algorithms are there?
There's Panda (not to be confused with the data manipulating python package).
It helps Google judge content as good or penalizable based on specific factors.
Keep in mind that this "one algorithm" probably contains many subdivided formulas and algorithms themselves.
Then there's Penguin.
This one is used to determine link quality.
As with Panda, there is likely one "core" algorithm that receives information from many specific sub-algorithms to determine a final "good" or "bad" determination for a link.
But we don't have to get too deep into the weeds here.
We can classify algorithms as part of a more general category, like:
Organizing/clustering algorithms
Task-specific algorithms
Entity-linking algorithms
Now we move on to an important aspect of the inner workings of algorithms.
Why do search engine algorithms use entities?
This is a newer topic of discussion that's criminally under-utilized by SEOs.
Google defines an entity as, "A thing or concept that is singular, unique, well-defined and distinguishable."
So bowling is an entity. Sports is an entity. And so on.
To continue our analogy from earlier, each player on a football team is an entity, but the team as a whole is also an entity.
Just as with algorithms, there is a hierarchy at play.
The way that Google understands the world wide web is as a collection of entities.
Using its knowledge graph, Google links entities based on shared characteristics.
These linkages allow Google to understand what topics are mentioned alongside each other and the context that related topics add to each other.
Entities are crucial to implementing semantic SEO (guide coming soon), which is a newer concept addressing how you should build content for Google and your users.
How do these search algorithms use entities?
We've discussed algorithms and entities, but it may not be apparent how they relate.
While you may understand a little bit more about how Google sees and ranks content, you may not understand the relevance between a football team and a football player.
Let's pretend we have a webpage about local SEO. There's a page that links to it that discusses technical SEO tactics.
Since Google understands that these two entities are related (both are SEO subdivisions), it will rank this collection of pages higher than a page about SEO linked to a page about football.
This was a huge leap in Google's ability to find and promote great content because it became much easier to detect spammy and black-hat elements.
It was much easier to manipulate search results through keyword spamming than it is to manipulate Google's understanding of entity relations.
How do these search algorithms work?
We've covered a few topics today so let’s put it all together.
Context is crucial
When you see that there's been a Google algorithm update, you know that that single algorithm is a smaller piece in a giant puzzle.
Understanding this hierarchy lets you understand the up-wind and down-wind effects of a new update, which gives you an advantage in adjusting your website for that update.
Entities are your friend
Now that you know how Google understands and relates content, you're far more equipped to write high-ranking content than most professional SEOs.
Make sure you're writing about related entities, and make sure you're also describing how they're related. Google loves this kind of content and you're likely to see that favor in your rankings.
User intent above all
If you take away only one thing, let it be this: all of these algorithms and formulas work together to produce results that answer a user's query.
That's what optimizing your website is for. To present information in a way that Google says, "This is the most well-structured and complete answer for this particular query".