Full transcript
0:00Most authors don't actually know who
0:02they're writing for. They say things
0:03like, "My book is for everyone who loves
0:06a good story." That is unfortunately
0:07code for I have no idea who my reader is
0:10and I'm praying Amazon figures it out
0:11for me. And if that's you, it's not a
0:13moral failing. It's just a skill gap.
0:15And in publishing, that skill gap is
0:18expensive because if you don't know who
0:20you're writing for, you won't know how
0:22to title the book, how to design the
0:23cover, how to write the blurb, how to
0:25aim your ads, or even have to decide
0:27what tropes to lean into. Everything
0:29just becomes a guess. In this video, I
0:31want to close that gap. I'm going to
0:33show you how to use AI to build a reader
0:35avatar so specific, you can literally
0:37hand it to a large language model and
0:39say, "Write a book that this person
0:41can't put down." When you get that
0:42specific, titles get easier, covers get
0:44easier, blurbs get easier, and the
0:46writing itself gets easier because
0:48you're writing to one human instead of a
0:50vague crowd. And the ironic part is the
0:53tighter you write to one person, the
0:55more books you actually usually sell.
0:58For example, you could write a fantasy
0:59novel for people who like fantasy. Or
1:02you could write a morally gray romantic
1:05C with enemies to lovers and found
1:07family for 27-year-old women who binge
1:10BookTok at midnight and want to escape
1:12their boring job.
1:13Which one do you think is easier to sell
1:15to? So, here's where we're going in this
1:17video. First, I'll show you why
1:19publishing is actually one of the
1:20easiest industries in the world to find
1:22an ideal customer in.
1:24And then I'll walk you through
1:25step-by-step how to find the right
1:27comparison books, mine them for
1:28patterns, use AI to extract tropes and
1:31expectations, analyze covers and blurbs
1:34so you don't accidentally fight your
1:35genre, and then finally turn all of that
1:38into a detailed reader avatar that you
1:39can plug into whatever AI tools you use,
1:42whether that's ChatGPT or Claude,
1:45Autocrit's beta reader, ad generators,
1:47whatever. So, let's start with the good
1:49news. Most businesses have to fight
1:52tooth and nail to figure out who their
1:53ideal customer is. You're selling
1:55software or services or consulting, and
1:58the market is honestly pretty messy.
2:00Different industries, job titles, income
2:02levels, use cases. It's honestly quite
2:04chaotic. But books are way simpler. In
2:07publishing, the who is mostly baked into
2:10one decision, genre.
2:12Genre is just a fancy word for shared
2:15expectations. It's a bundle of tropes,
2:17emotions, pacing, and vibes that a
2:20specific group of readers reliably pay
2:22for. That means your ideal reader is not
2:24some mystical being you have to meditate
2:27to discover. They are the person who
2:29already loves that exact genre and
2:32subgenre. Your first job is simply to
2:34pick that lane clearly and specifically
2:36enough so that you can say, "I write for
2:38this type of reader." This is where a
2:40lot of authors stall. They'll say, "I
2:43write fantasy" or "I write romance."
2:45Fantasy is vague. Cozy fantasy with low
2:48stakes, cottagecore vibes, and a
2:50slow-burn queer romance, that's more
2:53specific.
2:54Romance is vague. Spicy small-town
2:57single dad cowboy romance with a secret
2:59baby and a redemption arc, that's
3:00specific. Once you claim a specific
3:03lane, your reader avatar almost
3:04pre-exists. You're not inventing them,
3:07you're discovering them. They're already
3:09on Amazon, already reading, already
3:11leaving reviews. You just haven't gone
3:14to look at it yet. So, your first real
3:16decision is what exact shelf or exact
3:19little corner of the bookstore do you
3:21want to own? If you're between ideas,
3:23pick one and commit. Confusion is more
3:25expensive than being slightly off. As
3:27long as you're in a real lane with real
3:29readers, you can refine over time,
3:31thankfully.
3:32Uh but I write for everyone is just a
3:34more sophisticated way to say I'm okay
3:36with selling nothing. Now, once you've
3:39picked that lane, we start to do the
3:41work. Step two is where you're going to
3:43go competitor hunting. This is where we
3:45stop theorizing and start looking at
3:47what real readers are already buying.
3:49You want five to 10 books that are as
3:51close as possible to what you're writing
3:54or want to write. Not just in the broad
3:56genre, but in the overall vibes. So,
3:58that could be heat level, tone, tropes,
4:01length of the book. And if your book is
4:03a fast-paced, snarky urban fantasy with
4:06a sarcastic heroine and a slow-burn
4:08romance, don't grab a lyrical, slow,
4:11literary fantasy about grief because the
4:13reader for that is different. So, the
4:14way I do this is stupidly simple. I open
4:16Amazon and just start where my book
4:18would live. That could be category
4:20pages, it could be keyword searches, it
4:23could be the customers also bought
4:25section. And I look at what's ranking in
4:27my exact subgenre. I click through, pay
4:30attention to covers, titles, and blurbs
4:32that feel like, "Yeah, this is close to
4:34what I'm doing." And then I find five to
4:3610 of those and I buy them. Now, I don't
4:39treat this like a casual browse. This is
4:42market research. I want enough samples
4:44that I can see patterns. So, five to 10
4:46competitors is usually enough. More is
4:48also fine. Ideally, you read them all,
4:50and I do mean all of them. Now, in some
4:53cases, you are already a fan of this
4:55genre, so you may have already read a
4:57bunch of books in the genre, which is
4:58great.
4:59Uh if you haven't, this may sound like
5:01work, and it is. Welcome to the part
5:03where you separate yourself from the
5:04authors who are just hoping for sales
5:07while you're actually building the thing
5:08that deserves the sales. You want to
5:10absorb how these books feel, what
5:13readers are clearly rewarding. Quick
5:15pause. If you're tired of the one-line
5:17magic prompts, I've got you. Join the
5:19Story Hacker AI waitlist at the link
5:21below and I'll send you my real prompt
5:23pack, the multi-hundred-word prompts
5:25that I use for full books. Click the
5:27link, drop your email, get the pack.
5:29Now, if you don't have the time to read
5:31each of these fully, at least get the
5:33skeleton of the book. And you might want
5:35this anyway for steps that we're doing
5:37later.
5:37Grab summaries from Wikipedia if you can
5:40find those or from blogs or literally
5:42just ask AI for a detailed plot
5:44breakdown of each book. And you just
5:46need enough of the story to know what
5:48promises they make, what they deliver,
5:50how they structure the emotional payoff,
5:52etc. So, now you've got a pile of books.
5:54You've got some knowledge of their
5:55plots. You've got all of their Amazon
5:57pages sitting in your tabs. That's raw
5:59material. Next, we're going to feed it
6:01to AI in a way that gives you leverage
6:03instead of hallucinations. Step three is
6:05where we use AI to extract the common
6:07tropes and promises across those books.
6:10Most people misuse AI here. They open
6:12ChatGPT and say, "Who is the ideal
6:14reader for my fantasy book?" That's
6:15honestly useless. It doesn't know you,
6:17it doesn't know your story or your
6:20your chosen genre. And it isn't magic,
6:22it's just very good at predicting text
6:24based on what you feed it. So, instead,
6:25we feed it with real evidence. So, I
6:29would take the summaries of those five
6:31to 10 books, and you can paste them in
6:33one by one or combine them into a single
6:34document with clear dividers. Then, in
6:36your AI of choice, you say something
6:38like, "Here are summaries of several
6:40best-selling books in this subgenre."
6:43And then you just insert your subgenre.
6:45"For each book, identify the key tropes,
6:47emotional beats, character archetypes,
6:48and recurring themes. Then analyze all
6:51of them together and tell me what they
6:53have in common. What do readers of this
6:55subgenre clearly expect and enjoy based
6:57on these patterns." What you're doing is
6:59asking AI to be your pattern detector.
7:01It's scanning the stories and surfacing
7:03what repeats. So, that could be things
7:05like enemies to lovers, grumpy sunshine,
7:07found family, slow burn, high angst, low
7:09angst, cozy vibes, brutal violence,
7:12whatever it is. And this is gold because
7:14tropes are just shorthand for the
7:16emotional experiences that your ideal
7:18reader is paying for. They are buying
7:21secret identities revealed or last
7:23chance second chances
7:24or found family that feels safer than
7:26blood, and not just a book. Once you see
7:29the same tropes and beats pop up across
7:31multiple successful books, you're
7:33looking at the skeleton of your reader's
7:34dream outcome in story form. You can go
7:38one layer deeper and even ask AI, "Based
7:41on these patterns, what does the typical
7:43reader of these books want to feel while
7:45reading? What are they trying to escape
7:47from? What kind of ending do they
7:49require to feel satisfied?" Now, AI
7:51starts giving you language like they
7:53want emotional intensity but a
7:55guaranteed happily ever after, or they
7:57want to feel seen as ambitious but
7:59misunderstood, or they want to escape a
8:01stressful life into a soft, cozy,
8:03low-stakes world. That's your reader's
8:05internal world. You're already closer to
8:07a real avatar just from this, but we're
8:09not actually done yet. Step four is to
8:11now cross-analyze the covers and blurbs
8:14because your reader doesn't meet your
8:16story first. They actually meet the
8:17packaging of your story first. So, open
8:20the Amazon pages for each of your
8:22competitor books, and don't skim. Study
8:24them like a detective. Start with the
8:26covers first. Ask if I squint, what are
8:28the common signals here?
8:31Things like color palettes, fonts,
8:32figures and symbols,
8:34heat level, darkness versus lightness,
8:36that sort of thing. You can even feed
8:38these to AI to get the similar analysis.
8:41Now, in most genres, the top sellers
8:42cluster around a very specific visual
8:44language. Thrillers will have a certain
8:47color and typography vocabulary. Cozy
8:49romance will have a different one.
8:51Readers are not consciously thinking,
8:53"Ah, yes, a dominant teal with a script
8:55font signals low heat contemporary
8:58romance." But their pattern-matching
9:00brain is actually making that
9:02comparison. They scan for familiarity.
9:04They want to know in like half a second,
9:07"Is this for people like me?" Then look
9:09at the blurbs. Copy and paste them into
9:11your AI and say, "Analyze these blurbs.
9:13What phrases, promises, and hooks are
9:15repeated? What do they emphasize about
9:18characters, stakes, and tropes? What's
9:19the typical tone?" You'll start to see
9:22the same thing show up over and over,
9:24whether that's slow burn, enemies to
9:26romance, grumpy boss, sunshine
9:27assistant, one small town, one big
9:30assistant, perfect for fans of X and Y,
9:33and you'll see whether they lean into
9:35heat or fade to black, or whether they
9:37emphasize plot or vibes, or just how
9:40explicit are these tropes. This part
9:41matters because your reader avatar
9:43doesn't just have internal desires. They
9:45have pattern expectations. If all the
9:48covers in your lane are dark, moody, and
9:50minimalist, and you come out with a
9:51bright cartoon cover because you like
9:53it, you're making your reader work
9:55harder to recognize your book as
9:56something that they would enjoy, and
9:58that's the friction that you don't want.
10:00Once again, your job is not to be
10:02original, especially at the level of
10:04packaging. Your job is to be legible, to
10:06fit the genre visually and tonally, so
10:08your right reader instantly says, "One
10:10of those. I love those." And then you
10:12can be original inside the pages, at
10:14least to a certain extent, because you
10:16still don't want to alienate alienate
10:18those readers. Now you've got your trope
10:20patterns from the story summaries, your
10:21emotional patterns from the analysis,
10:23your visual patterns from covers and
10:25language patterns from the blurbs.
10:26That's a lot of signal you've got. So
10:28now what we do is combine it. Step five
10:31is to turn all of that into a single
10:32detailed reader avatar and plug it into
10:34your AI tools. This is where the machine
10:37starts paying you back. Up till now you
10:38did the boring work finding competitors,
10:40buying them, reading or at least
10:42summarizing, and feeding all of that
10:44into the AI and asking smart questions.
10:46Now you create the human that you're
10:48going to write for. Take everything
10:50you've learned and give it to your
10:51chatbot of choice with one big prompt.
10:54It might sound like this in normal
10:55language, although you can absolutely
10:56tweak this for yourself. Something like,
10:59"I'm an author writing in your very
11:01specific genre, and I've researched 10
11:04books in this genre. Here are the common
11:06tropes, emotional beats, and themes that
11:08I've found." Then you just copy and
11:10paste everything.
11:11Here's what the covers and blurbs tend
11:12to emphasize. Then you copy and paste
11:15all of that. Using all of that, create a
11:16single detailed reader avatar who is the
11:18perfect target for this kind of book.
11:21Include their age range, gender if
11:22relevant, life situation, job, what
11:25their day looks like, what they're
11:26stressed about, what they're secretly
11:28craving, how they discover new books,
11:30what kind of tropes do they love, what
11:32they hate, what heat level they want,
11:35and how they feel when they finish a
11:37book that they adore. When AI spits that
11:39out, don't just skim it and move on.
11:41This is the person you're going to write
11:42to for the next 60,000 plus words. So
11:46you want them to feel like real people.
11:48Ask follow-ups like, "Write a
11:50day-in-the-life scene from this reader's
11:52perspective, focusing on how and when
11:53they read." Or have this reader write a
11:57one-star review of a book in this genre
11:59that disappointed them. What went wrong?
12:01And then have a reader write a five-star
12:03review of the perfect book in this
12:04genre. What did it deliver? And those
12:06answers give you a three-dimensional
12:08human. You know what they hate, you know
12:10what makes them gush. You know how they
12:12talk. From a business perspective, this
12:13is what we call a dream outcome, uh plus
12:16the fears and frictions around it. From
12:18an author perspective, this is your true
12:21north. Now, practically, what do you do
12:25with this avatar? Well, first you write
12:27to them, literally. When you're drafting
12:29and you're stuck, you ask, "Would your
12:31avatar keep reading this scene? Uh what
12:34would they want more of right now? More
12:35banter, more tension, more cozy feels?"
12:39And if you're not sure, you can even
12:40paste a scene into your AI and remind it
12:42of the avatar to say, "Act as this
12:44reader. Give me blunt feedback on this
12:47scene. Where would you skip? Where do
12:48you light up?" Second, you market to
12:51them.
12:51When you're writing your book
12:52description, you can say to the AI,
12:54"Using this reader avatar, write three
12:57versions of the book blurb that would
12:58hook them instantly using their own
13:00language and focusing on their favorite
13:02tropes." Now the copy is grounded in
13:04actual reader desires instead of
13:06whatever you think sounds cool. Third,
13:08you can plug this avatar into tools like
13:10Autocrit's beta reader feature.
13:12Uh give the avatar as context and have
13:14it review your manuscript specifically
13:16as that type of reader. You can ask,
13:18"Where did this book fail to deliver on
13:20the tropes you promised?" Or what scenes
13:23did you love most as this reader and
13:25why? And that's ridiculously powerful
13:27feedback early in the process before
13:29you've spent months in on revisions.
13:31Now, fourth, you can use this avatar
13:34everywhere else. We're talking ad
13:35headlines, social media posts, email
13:38subject lines, and whenever you're about
13:40to talk about your book publicly, you
13:41ask, "What does
13:43your avatar need to see or hear to
13:46recognize this as a book for them?"
13:48And that's how you stop writing generic,
13:50weak copy and start sounding like you're
13:52inside your reader's head. And the
13:53beauty is, once you've built this avatar
13:55once, you can reuse it. You can ask AI
13:57to make small variations like older
13:59versions of this reader or more
14:01budget-conscious versions of this
14:03reader, just like we'd segment leads in
14:05a business and nurture them
14:07specifically. You're building a tiny
14:09data-driven universe of your readers
14:10that you can keep writing for book after
14:13book. Now, let me zoom out a bit and
14:15connect this to why this matters beyond
14:16just one book. Most authors live in
14:19permanent uncertainty. They don't know
14:21if their idea is right, if their cover
14:22is right, if their blurb is right. They
14:25publish, pray, and then blame the
14:27algorithm when nothing happens. And
14:29that's like opening a gym not knowing
14:31who your customer is, slapping random
14:32offers on the wall, and then complaining
14:34that marketing doesn't work. You're
14:35doing something different. You're
14:37saying, "I'm going to choose a clear
14:39lane. I'm going to study what actually
14:41sells in that lane. I'm going to mine
14:43those books for patterns, and I'm going
14:45to use AI not as a magic eight ball, but
14:47as an amplifier of real reader data. And
14:50then I'm going to build one specific
14:52human in my mind to write to them so
14:55relentlessly that when they see my book,
14:57they will feel like it was built for
14:58them." That's how you build a career
15:00instead of a just a random lottery
15:02ticket. That's how you go from, "I hope
15:04someone reads this" to, "I know exactly
15:06who this is for and how to reach them."
15:08So your homework from this video is
15:09simple, but not easy, of course, as all
15:12good things are. Pick your exact
15:14subgenre, go find five to 10 competitors
15:17and buy them. Get summaries if you need
15:19to. Run those through AI and extract the
15:21tropes and the emotional beats and
15:22patterns. Analyze the covers and blurbs.
15:25Then feed all of that into an AI chatbot
15:27and have it build you one insanely
15:29detailed reader avatar. Give them a
15:31name, print them out, keep them where
15:33you write, and then next time you open
15:36your manuscript or your marketing
15:37dashboard, don't ask, "What should I
15:39write?" Ask, "What would make this
15:41person obsessed with my book or my
15:43series?" You do that consistently book
15:45after book, and it's going to be very
15:47hard for you to not win in the long run.