What is Google RankBrain?

RankBrain is an algorithm for machine learning (EN: Machine Learning), which Google uses to sort search results.

RankBrain is therefore a machine: Artificial Intelligence.

It also helps Google process search queries and understand the intent of a user. Website to understand.

So: What makes RankBrain different?

Before RankBrain, 100% of the Google algorithm was coded by hand.

Of course, human engineers are still working on the algorithm.

But today, RankBrain also works behind the scenes.

In short: RankBrain optimizes the algorithm independently..

Depending on the keyword, RankBrain increases or decreases the importance of backlinks, content recency, content length, domain authority, etc.

The next step involves examining how Google searchers interact with the new search results.

If users prefer the algorithm, it will remain.

If not, RankBrain resets to the old algorithm.

And Google's RankBrain measures whether the page is optimized for the User Intent is optimized.

Now that you have an overview of RankBrain, let's delve deeper into how it works.

Google RankBrain Chapter 1: How it works

RankBrain has two main tasks:

• Search queries (keywords) understand
• Measuring how people interact with the results (user satisfaction)

Let's dismantle each of them.

How RankBrain understands every keyword you search for

A few years ago, Google had a problem:

15% of the keywords that users entered into Google had never been seen before.

15% doesn't seem like much.

However, if you perform billions of searches per day, that's 450 million keywords overwhelming Google every single day.

Before RankBrain, Google scanned pages to see if they contained the exact keyword someone was searching for.

However, since these keywords were brand new, Google had no idea what the searcher actually wanted.

So they made a guess.

Let's assume you were searching for "the grey console developed by Sony§".

Google searched for pages containing the terms "grey", "console", "developed" and "Sony".

RankBrain now understands exactly what you are asking.

And it delivers 100% accurate results:

Not bad.

What has changed?

Previously, Google tried to match the words in your search query to the words on a page.

RankBrain is actually trying to figure out what you mean today.

How?

By matching previously unseen keywords to keywords that Google has previously seen.

For example, Google RankBrain may have detected that many people are searching for a grey console developed by Nintendo.

And they have learned that people searching for a grey console developed by Nintendo want to see a range of results related to game consoles.

When someone searches for "the grey console developed by Sony", RankBrain delivers similar results to the already known keyword ("grey console from Nintendo").

So it shows results across consoles.

In this case, the PlayStation.

How RankBrain measures user satisfaction

Sure, RankBrain can make an attempt to understand new keywords.

The algorithm can even be customized.

But the big question is: If RankBrain displays a series of results, how does it know if they are actually good?

In other words, RankBrain shows you a range of search results that you think you might like.

If many people like a particular page in the results, they will give that page a ranking boost.

And what if you hated it?

You discard this page and replace it with another page.

And the next time someone searches for this keyword (or a similar term), they will see how it works.

What exactly does RankBrain observe?

Very close attention is paid to how you interact with the search results.

Specifically, it concerns:

• average organic click-through rate
• Length of stay
• Bounce Rate
• Pogo-Sticking

These are referred to as User Experience Signals (UX Signals).

Let's look at an example:

They developed sore muscles in their back while playing tennis.

So you search for "retracted muscle" on Google.

Like most people, you click on the first result.

Unfortunately, the intro is full of fluff and filler ("Your back is an important muscle group...").

So press the back button and check out the second result.

This one isn't much better.

It is full of general advice such as "rest and ice on the back".

So you press the "Back" button in your browser again and give result #3 a shot.

Bingo!

This result is exactly what you are looking for.

Instead of going "back", you need to spend 5 minutes reading through the physiotherapy routine on the page.

And because you got what you wanted, you didn't revisit the search results.

This back-and-forth movement is called "pogo-sticking".

And RankBrain pays close attention to that.

When Google detects that users are quickly leaving a page to click on another search result, a strong message is sent to Google: "This page stinks!".

And if Google notices that many people are stopping to follow a particular result, they will give that page a boost to make it easier to find.

Google RankBrain Chapter 2: Keyword Research

As you have seen, Google can understand the intent behind a keyword.

Does this mean that traditional keyword research is dead?

No!

This means you may need to optimize the keyword research process to make it more RankBrain-friendly.

Ignore long-tail keywords (they are outdated)

In the past, it made sense to create hundreds of different pages.

Each one was optimized for a different keyword.

For example, you create a page that has been optimized for the "best keyword research tool".

And another one that has been optimized for "the best tool for keyword research".

And the old Google would classify each of them according to their respective long-tail keywords.

RankBrain now knows that these terms are essentially the same.

They therefore show almost identical search results.

In short, the optimization It no longer makes sense for long-tail keywords.

What should you do instead? Read on…

Optimize keywords

Instead of long tails, we recommend optimizing content around medium tail keywords.

Keywords with a medium level of meaning are located in the middle section of the package.

You will receive more search volume than your average long tail.

But they're not terribly competitive either.

Below you will find a number of keywords related to the topic of "fitness diet".

The terms in the middle are keywords for the middle end.

If you optimize your page for a medium-sized tail keyword (and make that page fantastic), RankBrain will automatically rank you by that term… and thousands of similar keywords.

In short, we recommend that you optimize your page around a single keyword.

(Just make sure it's a medium-reach keyword.)

Then let RankBrain sort your individual pages by a range of different related keywords.

Google RankBrain Chapter 3: Optimizing for CTR

As you saw in Chapter 1, the organic click-through rate is an important RankBrain ranking signal.

The question is: How can you get people to click on your result?

Well, that's exactly what we'll be covering in this chapter.

Fill your title tags with emotion

There is no doubt: Emotional titles get more clicks.

The copywriters have known this for years.

In recent years, this idea has been supported by data.

In fact, CoSchedule found a clear correlation between highly emotional headlines and social media shares.

Here, for example, is a generic SEO-optimized title tag:

Productivity tips: How to achieve more

Not bad.

But it lacks the "oomph" that makes people click.

Here's how you can transform the title tag into an emotional powerhouse (while remaining SEO-friendly):

Manage your to-do list with these 17 productivity tips

It's not always a good idea to create highly emotional titles.

But whenever you can, you should.

Add brackets to your titles

In fact, this little tactic works so well that we frequently use parentheses in most of our titles.

Here are some examples of brackets you can use:

• (2022)
• [Report]
• [Updated]
• (Case study)
• (Tips & Tricks)

Use numbers (and not just for list entries)

The data from multiple sources (including BuzzSumo) are unambiguous: Numbers improve the click-through rate.

The best part?

You can use numbers in your title even if your content does not contain any list items.

For example, I published this case study last year:

As you can see, I've included not one, but TWO numbers in the title.

Create power words in your title tags

Power words are terms that have a serious emotional impact.

You guessed it: Using Power Words will make your title stand out and get you more clicks.

Here is a list of some of our most popular power words of all time:

• Effective
• Insane
• Case study
• examples
• Proven
• Study
• New
• Powerful

Don't forget to optimize your description tag for CTR.

Yes, description tags no longer directly help with SEO.

However, we have found that an optimized description can significantly increase your click-through rate.

Here's how to create a description tag that delivers results:

1. First, make it emotional.

Just like with a title tag, you want your description tag to convey emotions.

2. Next, show why someone should click on your result.

Is your content comprehensive?

Based on research?

Funny?

Sell ​​that in your description.

3. Finally, specify your target keyword.

Google makes it bold, which can make your result stand out.

Google RankBrain Chapter 4: Optimizing Content for Bounce Rate and Time on Site

You need to show Google that your result makes users happy.

And the best way to do that?

Improve your dwell time.

Does Google really quantify dwell time? Yes!

Dwell time is the time a Google searcher spends on your website after clicking on your result.

The longer someone spends on your site, the better.

This is what Google says: “Man, people love this result. Let’s add a few more places.”

And if someone leaves your website after 2 seconds, Google will be notified:

"This result stinks! Let's drop a few positions."

Therefore, it is logical that RankBrain measures dwell time – and swaps the results based on this signal.

A Google employee said recentlyGoogle relied 100% on off-page signals (especially backlinks).

Although Google still uses backlinks, this employee points out:

And the data supports this.

In analyzing a large number of Google search results, we found a correlation between high rankings and low bounce rate.

Here's how to reduce bounce rate and boost dwell time.

Now it's time to explain a few simple strategies you can use to increase the time visitors spend on your website.

1. Slide your content over the folder.

When someone clicks on your website from Google, they want their question answered NOW.

In other words, they don't want to scroll down to read your content.

For this reason, we strongly recommend removing anything that pushes your contents under the fold.

Instead, you want the first sentence of your content to be centered at the front and back.

This way you immediately hook your reader.

2. Use short intros (5-10 sentences MAX)

Believe it or not, we spend MORE time on our intros than on our headlines.

This is because 90% of your readers decide whether to stay or leave based on your intro.

And after many tests, we found that short intros work best.

Why?

If someone searches for something on Google, they already know about that topic.

A massive intro is therefore not necessary.

3. Publish detailed content

Longer content = better dwell time.

Obviously, reading a 2,000-word guide takes longer than reading a 400-word blog post.

But that's only part of the equation.

The other reason why long-form content improves dwell time is the fact that longer content can fully answer a search query.

Let's say, for example, you are looking for a marathon.

The first result you click on is a 300-word article.

It kind of answers your question… but you want more.

So you click the back button to find something better (as you may recall, Google calls this "pogo-sticking").

And this time you hit the jackpot.

You will find a comprehensive guide that includes EVERYTHING you need to know about running a marathon.

So you take a cup of coffee and read the instructions from beginning to end.

All this reading requires serious time and attention.

Long-form content performs so well that we ONLY publish content with at least 2,000 words.