It’s October 2015… Spectre is tearing up the box office, bringing Bond back to London. Jürgen Klopp has just arrived at Anfield, promising to turn doubters into believers. And over in California, Google is quietly admitting that for the first time in history, the machine isn’t just following rules—it’s starting to think.
For nearly two years, Google had been running a new system in the background. They called it RankBrain. It wasn’t just another update like Panda or Penguin, which were essentially digital wardens designed to punish bad behaviour. RankBrain was different. It was a student.
It was the first time Google deployed Artificial Intelligence (AI) to directly influence search results. It changed the game from “matching words” to “understanding intent,” and for British businesses, it marked the moment we had to stop optimising for robots and start optimising for humans.
This is the story of how it works, why it matters, and how it acts as the grandfather to the AI revolution we see today.
Please note: The content below may contain affiliate links. If you make a purchase through these links, we could earn a commission, at no additional cost to you.
1. The Problem: The 15% Mystery
To understand why RankBrain exists, you have to understand a specific headache Google had in the early 2010s.
Every single day, Google processes billions of searches. But here’s the kicker: back in 2015, 15% of those queries had never been seen before.
That might sound small, but 15% of billions is a massive number. These weren’t standard searches like “BBC News” or “trains to Manchester.” These were complex, long-winded, or weirdly specific questions.
The Old Way (Pre-RankBrain): If you searched for “that film where the guy gets left on Mars and grows spuds,” the old Google would panic. It would look for pages containing the words “film,” “guy,” “Mars,” and “spuds.” You might get a page about potato farming or a sci-fi blog, but not necessarily the right answer.
Google was essentially guessing. It had no “gut instinct.” It needed a way to interpret queries it had never seen by comparing them to things it had seen.
Enter RankBrain.
2. What Exactly Is RankBrain?
It’s a machine-learning artificial intelligence system.
In plain English? It’s a computer program that teaches itself how to do something, rather than being told exactly what to do by a human coder.
Before it, Google’s engineers wrote strict rules. If X, then Y. With RankBrain, the engineers said: “Here is a pile of data. Figure out the patterns.”
The “Pub Quiz” Analogy
Imagine you’re at a pub quiz. The question is: “What’s the name of the big spiky building near London Bridge?”
You might not know the answer is “The Shard.” But your brain makes connections:
- “London Bridge” = A location in London.
- “Spiky building” = A tall, glass structure.
- “Big” = Skyscraper.
Your brain connects these concepts and makes an educated guess. RankBrain does the exact same thing for Google. It looks at words not as isolated strings of letters, but as concepts that are related to one another.
RankBrain vs. Hummingbird
This is a common point of confusion. We previously discussed the Hummingbird update (2013), which was a complete overhaul of Google’s engine. So, where does RankBrain fit in?
Think of Google Search as a car.
- Hummingbird is the engine. It’s the main machinery that powers the whole thing.
- RankBrain is the sophisticated onboard computer system controlling the fuel injection.
RankBrain didn’t replace Hummingbird; it is part of Hummingbird. It’s a specific component designed to handle the queries that the rest of the engine finds too difficult or ambiguous.
3. How It Works: Vectors and “Word Maths”
This is the only technical bit, so stick with it. It’s fascinating.
RankBrain uses a process called Word Embeddings to turn words into mathematical objects called vectors. Imagine a giant 3D map where every word in the English language has a specific coordinate.
On this map, words that mean similar things are placed close together.
- “Chips” is close to “Fries” and “Potato.”
- “Lift” is close to “Elevator.”
- “Flat” is close to “Apartment.”
The Magic of “Word Maths”
Because words are now maths, RankBrain can do sums. A famous example of this vector logic is:
Paris – France + UK = London
RankBrain understands that the relationship between Paris and France is the same as the relationship between London and the UK.
So, if a user in Peterborough searches for “best boot for walking the dog”, RankBrain looks at the vector map. It sees that in the UK context, “boot” is close to “Wellingtons” or “hiking shoes,” and far away from “car boot” or “boot up a computer.” It understands the context without you having to explain it.
4. The British Context: Dealing with Our Dialects
For UK searchers, RankBrain was a revelation. Our language is notoriously tricky. We say things differently from our American cousins, and we have slang that changes every 50 miles.
Example 1: The “Chips” Conundrum
- User A (USA): Searches “chips.” Intent: Potato crisps (Lays/Walkers).
- User B (UK): Searches “chips.” Intent: Hot, fried potatoes from the chippy.
RankBrain analyses the user’s location and previous search history. If you’re in Norwich, it knows “chips” likely pairs with “fish” or “curry sauce.” If you’re in New York, it pairs it with “dip” or “salsa.”
Example 2: The “Football” Divide Before RankBrain, searching for “football results” was a gamble. You might get the NFL scores. Now, RankBrain knows that if a user in Liverpool types “football,” they definitely do not mean the sport played by the Miami Dolphins.
It understands that UK + Football = Premier League, while US + Football = NFL.
5. The Impact on SEO: The Death of Keywords (Again)
When RankBrain launched, the SEO industry had a collective meltdown. But once the dust settled, the rules became much simpler.
Keyword Stuffing is Dead
In the old days, if you wanted to rank for “Cheap Plumber Bristol,” you wrote “Cheap Plumber Bristol” five times in your first paragraph.
RankBrain killed this. It doesn’t care if you use the exact phrase. It cares if you answer the question. If you write a helpful article titled “Affordable Plumbing Services in the South West”, RankBrain knows that “Affordable” is a vector neighbour of “Cheap,” and “South West” covers “Bristol.” You rank anyway.
The Rise of User Satisfaction
This is the most critical shift. RankBrain isn’t just analysing what you type; it’s analysing how you interact with the results.
It looks at signals that suggest you found what you wanted. While Google denies using “Dwell Time” (how long you stay on a page) as a direct ranking factor, RankBrain essentially measures success.
- The Scenario: You search for “How to make a Victoria Sponge.”
- Result A: You click the top link. It’s a wall of text with no recipe. You hit “Back” immediately. (This is called Pogo-sticking).
- Result B: You click the second link. It has a clear ingredients list and a video. You stay for 5 minutes.
RankBrain notices this. It thinks: “Result A failed the user. Result B satisfied the user. Swap them.”
For content creators, this means your goal is no longer to trick a robot. It’s to satisfy a human.
6. Practical Tips: Writing for the RankBrain Era
If you own a website or write content in the UK, here’s your survival guide.
A. Focus on “Medium-Tail” Concepts
Don’t obsess over single words. Focus on complete concepts. Instead of a page just for “Red Shoes,” create a comprehensive guide on “Best Red Shoes for Winter Parties.” RankBrain loves context.
B. Answer the Question Immediately
British politeness often makes us waffle. We write long introductions. RankBrain hates waffle.
- Bad: “Since the dawn of time, man has wondered about the weather…”
- Good: “The best time to visit Cornwall is July, when the average temperature is 19°C.”
Get to the point. If you solve the user’s problem quickly, they won’t hit the “Back” button.
C. Use Natural Language
Write like you speak. RankBrain is trained on human speech. If your content sounds robotic, it won’t connect. Use UK spellings (Colour, Optimise) and local phrasing naturally—RankBrain understands it.
D. Emotional Engagement
Headlines matter more than ever. You need a high Click-Through Rate (CTR).
- Boring: “List of Top UK Beaches.”
- Engaging: “10 Secret Cornish Coves You Need to Visit This Summer.” RankBrain sees people clicking the second one and assumes it must be better.
7. The Legacy: From RankBrain to the Future
RankBrain was just the start. It was the toddler that learned to walk so that future AIs could run.
- BERT (2019): Built on RankBrain’s foundations to understand the nuance of words like “to” and “for” in complex sentences.
- MUM (2021): The Multitask Unified Model, which is 1,000 times more powerful than BERT and can understand images, video, and text simultaneously.
None of these would exist without that quiet update in 2015. RankBrain proved that machines could learn from us, understand our intent, and deliver what we need—even when we don’t quite know how to ask for it.
Conclusion
Google RankBrain was the moment the internet got a “gut feeling.” It moved us away from the robotic, keyword-heavy web of the early 2000s and gave us a search engine that feels like a knowledgeable friend.
For British businesses, the lesson is clear: Stop trying to outsmart the algorithm. The algorithm is now smarter than you. Instead, focus on being the absolute best answer to your customer’s question. If you do that, RankBrain will do the rest.
Further Reading and Resources
To deepen your understanding of the concepts discussed, we have curated a list of the most authoritative and practical resources available. These links cover everything from the official engineering documentation to advanced SEO strategies suitable for UK businesses.
Official Google Documentation
- A Guide to Google Search Ranking Systems The official word from Google. This documentation explains how systems like RankBrain, BERT, and others fit into the broader search ecosystem.
Industry Analysis & History
- Search Engine Land: FAQ: All About The New Google RankBrain Algorithm The original investigative piece from 2015 by Danny Sullivan. This is the “primary source” for the “third most important signal” quote and is essential reading for history buffs.
- Search Engine Journal: Google Algorithm History A comprehensive timeline tracking every major shift. Use this to see where RankBrain sits in the timeline between Hummingbird (2013) and BERT (2019).
Practical SEO Guides
- Backlinko: Google RankBrain – The Definitive Guide Brian Dean’s guide is widely considered the gold standard for actionable RankBrain strategies. It covers “Medium Tail” keywords and satisfaction signals in depth.
- Ahrefs: What is RankBrain? A concise, no-nonsense glossary entry that cuts through the hype. Perfect for a quick refresher on the technical definitions.