What Is a Knowledge Graph for a Book? A Plain-English Guide for Authors

A knowledge graph for a book is a simple map of who it's for, what it does and when it's used. Here's what it is, why Amazon cares, and how to build yours.

By Gundi Gabrielle

12 min read

Start Publishing With AI

🚀 AI-Powered Publishing Studio✍️ Full Writing App🎧 Audiobook & Translation Generators🤖 AI Publishing Team🔍 Viral Topic Gap Finder📢 Ads Generator🎯 AI Marketing Automation

Knowledge graph... sounds super techy, right?

It isn't. If you've ever said "this book is for tired working moms who feel guilty at night," you've already done the hard part. A knowledge graph is just that kind of answer, written down in a neat map that Amazon's computers can read.

I'll show you what one is, what goes in a book's graph and how to build yours. This post is part of my Amazon KDP algorithm guide.

What is a knowledge graph?

A knowledge graph is a map of things (called nodes) joined by named connections (called edges), like "this product is used by this kind of person."

Each fact has three parts: a thing, how it connects, and another thing. Amazon's own example: slip-resistant shoes, linked by a "used for audience" connection to pregnant women. Amazon Science says the graph can use the used_for_audience relationship to link slip-resistant shoes and pregnant women.

Stack up millions of those and you have a giant map. And the names on the connections are the whole point. They tell the computer why two things belong together, not just that they showed up in the same sentence.

👑 Want a map of your own book? Golden Bestseller Empire builds yours from your answers. → See how

Does Amazon really use knowledge graphs?

Yes. Amazon built a huge one called COSMO, which is its way of figuring out what a shopper actually wants when they type something in.

Here's why they bothered. The COSMO paper says other shopping graphs hold lots of product facts, but "they fail to discover user intentions". In plain words: knowing a shoe is black and size 8 doesn't tell you why somebody buys it. So here's what COSMO's graph looks like:

• ✦6.3 million nodes (the "things" in the web)
• ✦29 million edges (the connections between them)
• ✦15 relation types (the names on those connections)
• ✦Built from real behavior: what shoppers search for and buy

Amazon says the same thing in its own news post: it's "not just about matching keywords anymore", it's about "understanding what customers are actually looking to buy."

One thing to keep straight: COSMO does not rank your book. It figures out what the shopper means and brings the right books to the table. Ranking is still sales, conversion and relevance. I explain the split in the COSMO guide.

Why does a knowledge graph matter for a self-published author?

A knowledge graph matters because it makes you think clearly about your reader before you write a single word of the listing. Most authors start with the title and hope for the best. That's backwards.

When you build your graph first, you get answers to the questions every buyer is silently asking:

• ✦Is this book for someone like me?
• ✦Will it fix the exact thing that's bugging me?
• ✦When would I actually use it?

Your listing, your chapters and your ads then all give the same answers. A clear knowledge graph means every page of your book tells one story. And it costs you nothing but an afternoon of thinking.

What is a knowledge graph for a book?

A knowledge graph for a book is a map of one book: one big idea in the middle, joined to its reader, its result, its daily moments and its pairings.

The idea in the middle is your Core Node. Around it sit 18 connections in six groups. Here's the visual map from Golden Bestseller Empire:

That's a sample book, not a real one. Pretty, isn't it... :)

Amazon already has its own map of how ideas connect. Your job is to hand it a clean, complete picture of your book so it can file it in the right spot.

👑 Not sure what your book's center is? Golden Bestseller Empire helps you find it. → Start here

What is a Core Node in a book's knowledge graph?

A Core Node is the one short phrase your book is the authority on. Think of it as your book's address in Amazon's map.

I teach a simple formula:

Reader + Primary Gap → Transformation = Core Node

Here it is with a real example:

Working Mom + Guilt → Reset = "Working Mom Guilt Reset"

Three to five words, and a stranger can picture who it's for.

Two things trip people up:

• ✦It's not your title. "Working Mom Guilt Reset" is the territory your book owns. "How I Stopped Feeling Guilty at 6pm" is a title.
• ✦The last word must fit the gap. Guilt plus "calm" doesn't match, because calm fixes anxiety, not guilt. Here's what fits:

And only one Core Node per book. Try to own three and Amazon can't tell what the book is. That's why a series works so well: Book 1 owns "Working Mom Guilt Reset," Book 2 owns a neighboring idea, and together they cover much more ground.

Why does one precise idea beat five vague ones in a knowledge graph?

Because Amazon trusts the most common reason people buy far more than the rare ones. In the COSMO paper, about 35% of search-to-purchase connections were the "typical" kind, the default reason a reader searches in a niche. Only about 9% of co-purchase connections were typical.

Here's what that means for you:

• ✦Pick what most buyers in your niche really want.
• ✦Cut the edge cases, even real ones. They blur the picture.
• ✦Ask yourself: "Is this what the large majority of readers want, or what a few could want?"

A knowledge graph with one clear, typical idea in the middle is easy for Amazon to file. A knowledge graph that tries to say everything says nothing.

What are the six groups in a book's knowledge graph?

A book's knowledge graph has six groups: Audience, Classification, Functional, Context, Interest and Desire, and Book Specific. Together they hold 18 connections.

Each group answers a plain question:

• ✦Audience (3): Who is this for? Their job, their life situation, and what they call this kind of book.
• ✦Classification (2): What kind of book is it, and why do people buy it (gift, book club pick)?
• ✦Functional (3): What does it do, what goal does it help with, and what change can it create?
• ✦Context (4): When, where and in what moment is it used? Even the body side counts, like cortisol crashes.
• ✦Interest and Desire (3): What does the reader want, what else are they into, and what do they use it with?
• ✦Book Specific (3): The author, the series and the format.

Amazon's COSMO paper lists 15 of these connections. I add three for books: Author, Series and Format. That's the 18 I map in the course. I go through every one in the COSMO relations guide.

My favorite is the Context group. Naming the exact moment, like the 5pm arrival chaos or the school drop-off, gives Amazon one more door into your book.

What do concept node, LSI term and intent string mean?

They're three ways of saying the same idea. A concept node is the idea itself, an LSI term is how you write it, and an intent string is how your reader asks for it.

Say the idea is mom guilt.

• ✦Concept node: Mom Guilt, the idea Amazon already knows.
• ✦LSI term: "mom guilt," "guilt spiral," "evening guilt." (LSI just means the related phrases you use in your text.)
• ✦Intent string: "why do I feel so guilty after work?"

Same idea, three costumes. Your graph covers all three, so what you write matches what readers type or say.

Notice what's happening here. When your listing says "mom guilt" and "mental load," Amazon isn't counting keywords. It's recognizing ideas it already has on its map.

👑 Want all three filled in for you? Golden Bestseller Empire maps every node. → See the generator

What does a finished knowledge graph look like?

A finished graph has a plain answer in every group, and every answer points back at the same Core Node. Here's a sample:

No magic. It's clear thinking, organized. You can see it in the app as text, as the visual map or as a coverage table.

Does Amazon read the words inside my book?

Yes. KDP's own Read Sample page says: "When customers search for books on Amazon.com, we use the actual words that match inside the book".

So your graph isn't only for the listing. The introduction, the table of contents and every chapter subheader get read. The Look Inside window covers roughly the first 10% of the book. That's:

• ✦Your table of contents
• ✦Your introduction (the part that does the most work)
• ✦The start of chapter 1

Each subheader can confirm another node, so I teach 4 to 6 per chapter, written as the reader's real situation, not a dull label.

How do you build a knowledge graph for your book?

You build one by answering questions about your reader and result, naming your Core Node, then filling in each group in plain words.

1. 1Pick one reader. Who are they and what's their biggest struggle?
2. 2Name the change. What can they do after reading?
3. 3Write your Core Node. Reader + Gap → Transformation.
4. 4Add 5 to 7 side themes. These are the neighboring ideas your book also covers.
5. 5Fill in the six groups. Two or three plain phrases each.
6. 6Add your author signals. Your story, your series, your formats.
7. 7Cut the odd ones. Pick what most buyers in your niche really want, not every possible angle. One clear idea beats five fuzzy ones.

You can do it on paper in an afternoon (just draw the Core Node in the middle with six lines out to the groups). Easy to leave gaps, though. That's why I built a generator.

👑 Skip the blank page. Golden Bestseller Empire asks the questions and builds the map. → See the generator

What mistakes should you avoid when you build a knowledge graph?

The biggest mistake is making the graph about you instead of your reader. Here are the ones I see most:

• ✦Using your title as the Core Node. The Core Node is the territory, not the headline.
• ✦Using fancy words. Write the way your reader talks. "Mom guilt" beats "maternal affective distress."
• ✦Leaving whole groups empty. Even one plain phrase in each group helps.
• ✦Owning too many ideas. One Core Node per book, remember.
• ✦Never using the graph. It's not a school project. Use the same words in your description, your chapter subheaders and your ads.

Fix those five and your knowledge graph is already better than most.

What does a knowledge graph feed?

The graph feeds everything else on your listing and in your book. Everything uses the same words, so it all says the same thing.

Here's what comes from it:

• ✦Your outline and chapter details
• ✦Your title and subtitle
• ✦Your book description
• ✦Your A+ content
• ✦Your backend keywords
• ✦Your translations
• ✦Your Amazon ads keywords

Even the "Customers Say" tags come from your listing copy, so clear copy plants them. And since Rufus became Alexa for Shopping on May 13, 2026 (you'll find it in the Amazon app menu, not the search bar), the plain, human wording in your listing is what that assistant has to work with. More on that in the Alexa for Shopping guide.

If you wonder how this differs from keywords, I cover that in knowledge graph vs. keywords.

How does Golden Bestseller Empire build it for you?

Golden Bestseller Empire is the only app with an AI generator that builds your knowledge graph with one click, straight from your own Book Planner blueprint. I don't know of another app that does this.

Here's how it works:

1. 1You write the blueprint. You answer guided questions about your reader, their biggest gap, the change and your goals. Your answers, your words.
2. 2You click Generate. The app builds all 18 connections from your blueprint, each in your reader's own words.
3. 3You see it as a map. As text, as the visual map, or as a coverage table.
4. 4That one graph builds the rest: your outline, description, A+ content, backend keywords, translations and ads keywords.

It's AI-assisted, not AI-written. You write or dictate the book yourself. The planning and publishing work is what gets so much easier.

I used it to write and publish AI for Beginners. It was faster, easier, and it hit #1 within 3 days.

Related reading

Want to go deeper? These posts pair well with this one:

• ✦Knowledge graph vs. keywords: how the two differ.
• ✦COSMO relations explained for books: every connection, one by one.
• ✦Amazon bookpage ready for AI shopping: get your page set up.
• ✦KDP keywords: rules and limits: the seven keyword boxes.

What should you remember about a knowledge graph?

A knowledge graph is just your book explained clearly: who it's for, what it fixes and when they use it. Start with one reader, name one Core Node and fill in the six groups in plain words. Do that, and Amazon has a clean picture of your book, and so do your readers. That's the whole power of a knowledge graph.

Created By

Gundi Gabrielle

Top 100 Business Author · Carnegie Hall Conductor · Founder of Golden Bestseller Empire

Hey, I'm Gundi, and for over a decade, I've helped thousands of everyday people turn what they know into a bestselling book.

Along the way, I sold over 350,000 books, and was listed in Amazon's Top 100 Business Author Charts, ahead of authors like Simon Sinek, Brené Brown, Tim Ferriss, Robert Kiyosaki, or Brian Tracy.

No traditional publisher. No massive budget. Just understanding Amazon's algorithm and researching what readers actually want.

In my earlier years, I was a classical conductor & organist, touring the world, incl. Carnegie Hall & a performance for the Pope on St. Peter's Square.

These days, I love traveling as a Digital Nomad and orchestrating busy AI Agents 😎. Out of that came Golden Bestseller Empire — a complete AI-Powered Book Publishing Studio now helping authors around the world.

Can't wait to see you inside & help you publish your dream:

Start Publishing With AI →

Featured In: