Facebook Ads’ long slide into ineffectiveness for authors continues, with word that placement controls are being taken off us.

Placements does what it says on the tin: it’s where your ad shows up. Feed, Reels, Stories, Marketplace, the right-hand column, Audience Network, Messenger, and a dozen others. For a lot of authors the old-fashioned newsfeed is the only placement that reliably converts, largely because our readers are older and that’s where they live. So we tick Manual Placements, untick everything except Facebook Feed, and sleep a little easier at night.

That’s the bit that’s going.

What’s actually changing

But before we throw our laptops across the room, some facts, or at least my attempt to extract facts from the actual announcements.

What Meta is removing is the ability to exclude placements at ad set level. Along with that go platform exclusions (Facebook vs Instagram), device exclusions (mobile vs desktop) and operating system exclusions (iOS vs Android).

But here’s the part I don’t see many others reporting properly: it only applies to campaigns running the Sales and Leads objectives. Every other objective keeps full placement control, at least for now.

Most of us are on Traffic. Not because Traffic is better, but because we’re sending readers to an Amazon product page and we cannot put a pixel on an Amazon product page. No pixel, no conversion event, nothing for a Sales campaign to optimise against. Traffic has been our primary choice for some time and if it’s yours, your placement controls are staying put.

Also…

Account-level exclusions survive regardless, but they only cover three placements: Audience Network, Facebook Marketplace and the right-hand column. Feed-only isn’t on the menu and won’t be.

On Sales and Leads there’s a checkbox, switched on by default, that lets Meta spend up to 5% of your budget on each placement you’ve excluded, whenever it reckons that’ll improve performance. So even the exclusions that still exist leak a bit. Jon Loomer has been all over this one, and he’s right that the default-on part is annoying. Read him here.

Oh, and there’s no firm date for any of it. Advertisers are getting notices, the change has been running as a test, and Meta hasn’t committed to a switchover.

The irony, which is not lost on me

Work out who this affects the most, and it’s the big growth area in indie publishing.

It’s the authors selling direct.

Your own store, your own checkout, your own pixel firing on a real purchase, which means a real Sales campaign, which means you’re in the group losing placement control. The people who took the advice, built the shop, wrangled the tax and did the hard version of this business are the ones getting the feature taken away.

The rest of us still funnelling everything to Amazon are untouched.

For now.

I’d like to tell you Traffic is a safe harbour. I don’t believe it is. Meta has spent five years pulling levers out of our hands one at a time, and they always start with the objectives where their automation has the most data. Sales and Leads first because that’s where the conversion signal lives. Traffic is not immune, it’s just further down the list, I suspect.

What do we actually do?

If we can’t tell Meta where to show the ad, and we can’t really tell it who to show it to either (interest targeting has been quietly dying for years) then the creative and the offer are the targeting. That’s how the system works now, and it’s how it’ll work from here.

Meta’s algorithm has been switching to learn from our ad creatives and the product details at the end of our target link.

So, put a Vulcan bomber on the ad with a headline written for a man who remembers the Cold War, and the AI will try to find that man, wherever he happens to be scrolling. Feed, Reels, Stories, doesn’t matter.

Run a vague ad with a stock photo of a paperback and it’ll find nobody in particular, at considerable expense. So from that point of view, I agree with Pierre Alex Jaunty writing in the 20Books group, that the platform can still be profitable, we just need to be better at our own marketing.

And we need to adapt, as always. I was personally cynical about video ads on Facebook. Historically I saw them gain clicks but not sales. But Meta has created a user expectation now and that means some form of animations/movement is fast becoming the minimum for a successful campaign.

So the job for the rest of 2026 looks like this. Get better at creatives and design them for the way the platform works today.

Make vertical versions of everything, and make the first second earn the second second. Also, test offers as hard as you test images, 99p first-in-series/ box set / reader magnet.

And judge campaigns on cost per acquisition/sale (using benchmarking) rather than sitting up reading placement breakdowns like tea leaves! (Something I am guilty of).

I’m still annoyed. Goddammit, Meta. But of everything they could have taken off us, a checkbox isn’t the worst of it.

They can have the checkbox. I still control my ads.

A number of people sent me the above note that was sent to writers on the RARE email list. RARE is a huge series of reader events that romance writers can pay to attend with a stall.

Now, of course they are absolutely within their rights to declare they don’t want AI-written books sold at the event, but the above notice, to me at least, is sinister. It has overtones of a dictatorship where you are encouraged to spy on your neighbour and report them.

It also seems, from the wording, to cover ANY and all AI use in an author business.

I just attended the Romance Writers of New Zealand conference, and I am sure it was typical of the current state of AI use with authors. While I am certain there were authors attending who want nothing to do with any AI, the vast majority are employing it throughout their author business, in places where it moves them forward, treating the new technology as just another tool to help us run our businesses.

If RARE are seeking to ban any author who uses AI in any way, I wonder who will be left to attend?

The reality is authors will want to take part, and so they will go along and pray that their neighbour doesn’t report them to the authorities.

I think RARE need to clarify and take a look at the language they use.

Maybe I’m alone, do let me know what you think below. I’m genuinely curious to know if I have become out of step with the self-publishing community and that this is now normal and acceptable?

BTW you can hear Cissy Mecca and me discuss this, along with other industry developments in this week’s Self Publishing Show.

James

In December 2025 I added my voice to those calling for an end to the wave of AI cancellations that seemed to be sweeping platforms like TikTok.

To be honest I wasn’t hopeful. The indie writing community has always had a subset of self-appointed police who deem it legitimate to force their take on others. My romance writing friends are particularly aware of this phenomenon. Romance books are policed more than any other genre. For instance, writers have to tread an impossible to define line between diverse representation and cultural appropriation. I even found out recently that spicy romance authors are expected to write their fictional characters using fictional condoms, lest they be accused to promoting unsafe sexual practices.

I don’t make light of the issues; diverse representation in novels is an area that deserves respectful consideration and promoting safe sex is laudable. I do, however object to groups of authors shouting ‘J’ACCUSE!’ at anyone who doesn’t comply with their own version of how to interpret these unwritten rules.

So it was no surprise that ‘AI anything’ was added to the reasons for demanding a boycott of an author. And ‘anything’ might simply be an AI character on a cover.

But now we have this same imposition of a notional set of rules being wielded by major publishers and agents.

This platform (Substack) is encouraging such a hunt by introducing an AI detector. It claims accuracy but its score is just an estimate. Isn’t it wrong to allow an estimate to throw career damaging accusations at someone? We can turn it off, but that presumably will only increase the suspicion of an anti-AI reader. I’ll be curious to see how this entirely human-written article fares…

In book publishing, we’ve had a few high-profile cases (I suspect there are plenty more we haven’t heard about), where major companies and agents are dropping authors after suspicion was raised that they used AI to assist their writing. In some cases the firms relied on AI to deduce this. Oh, the irony.

The speed at which suspicion can become professionally devastating should concern us all.

And it’s got me wondering, where the settled point is?

By settled point, I mean the place of acceptance.

It may be hard to imagine this, but I suspect that at some point in the future most writers will use AI, not just for plotting but in the writing process as well. It will be assumed as the norm by publishers, not hunted down as the tell-tale signs of fraud.

At some point we’ll realise that creating great stories and characters is what great novelists do. That’s the essence of the book and it’s less relevant how the words got down on the page.

I’m not arguing for this acceptance now, I just happen to believe it’s where we will end up, whether you, I or anyone else likes it.

In fact, I am probably like you in that as it stands now, I don’t want to read a book ‘written’ by AI. But that’s because not enough time living with AI has passed for me to believe that A Novel by A Novelist is amazing because they created it, regardless of whether they used AI at every stage.

As Freida McFadden said to her AI accusers, if you think writing a book with AI is cheating and easy, why don’t you do it?

Being forced to prove a negative (that she didn’t use AI which seems to be the position publishers put authors in these days) allowed her to understand the limitations of AI, that you can’t just press a button and get a great book.

We are just starting to understand that only in the hands of a great storyteller and character creator will it produce anything great.

Call Me, I'll Hide the Body, by Jerry Falade is the title most recently in the headlines after a $2M bidding war was followed by accusations of AI use and the book was dropped.

Falade denies using AI, but the publisher says his story changed and they have enough suspicion to drop the deal. We don’t know if AI was used, but if it was (there is at least a chance it was), IT IS A GREAT BOOK! That’s the point people seem to be missing and that’s the basis for me believing where we will end up.

So, whether an author used AI to create the book will, I think, become irrelevant. We’ll get to a point when enough great novels we’ve enjoyed are revealed to have been co-written with AI that it will stop being a thing.

I think this, because we’ve seen it so many times before.

My first career was in tech in the aerospace industry just as computer aided design wiped out the need for an entire generation of technical drawers. These men and women worked on drawing boards creating the technical diagrams for every single nut, bolt and switch that made up the the aircraft. Overnight CAD appeared. Some people quickly adapted, some felt they were too old to learn a new technology, and many of them fiercely opposed the introduction of the software citing critical safety concerns.

In time, computer aided design became the norm.

I was also one of the many in the industry vowing never to fly on a Boeing 777! The revolutionary aircraft was designed from the start with computer assistance, it was the first to enter service with permission to fly long routes across oceans with only two engines. In addition some of the pre-service test work that would have been carried out in real-world flights in the past, was carried out as computer simulations instead.

I, along with most others outside of that specific programme, was aghast! I didn’t want to put my life in the hands of ones and zeroes from a computer.

But I was wrong, and the technical experts were right. The 777 has been in service for three decades, to date 1,700 examples have been built.

The point is humans designed and built it, they just used computers to assist at every stage, including the flying.

Just like an author coming up with all the ingredients of a novel before employing AI to assist the writing.

Very few of us are at the point of acceptance yet and I am sure the vast majority of you never want to be.

It takes a mind-shift to accept that an author who used AI to generate some of the words, is still the author, and we are currently before the mind-shift. The process we will go through will probably involve disclosure as a norm, until it’s rare to find an author who didn’t use AI at all at any stage, at which point the disclosure becomes ignored. This might be a few years away, but I believe it’s coming because mind-shifts happen.

The way the B777 was designed and prepared for service is now the norm and not questioned by anyone.

At some point in the future, I’ll read a book by A N Author, who I love. And as I read I’ll think how wild their imagination is and how clever they are, and I won’t think about them using AI to organise the plot or even generate words, because that will be the norm and I will know from experience it takes talent to create a book that good regardless of how it’s written.

And if the B777 and Aircraft Drawing Office comparison are a stretch, using a technical industry to discuss the creative arts, what about digital animation?

Every frame of Disney’s Snow White and the Seven Dwarfs was hand drawn in 1937. That was the definition of animation, a human drew a picture. Today’s animated films are produced through digital workflows that ‘draw’ every pixel.

But the films are still man-made. The stories and images conceived by humans, written by humans whether or not they used AI for any actual word or image generation will, I think, become irrelevant.

Did you like the film? Was it art that humans wished to consume? These are the only questions that will be relevant once we’re used to the new way of working.

I’ve been thinking about how it used to be. I wouldn’t say easy as such, but we probably didn’t realise how much harder it would get.

The novel writing and the marketing have always been hard, but only now do I realise how easily the market forgave us. For about a decade, if you understood a genre, wrote competently within its conventions, put a cover on it that signalled the right promise, and turned up again four months later with the next one, there was a very good chance you would earn money, and for many, enough to make it a career. The algorithm rewarded consistency more than it rewarded brilliance. Read-through carried you. The ‘next book’ method of marketing worked.

That was a genuinely good deal. It was also, if I’m honest about it, permission to be a bit lazy.

I know, because I took it.

My own confession

I poured everything I had into my first novel. Years of it. Research trips, redrafts, readers, the lot. It found an audience, and that audience wanted more, and so I gave them more. But this time the books took months, not years. They were competent and they delivered what the cover promised. I’m not ashamed of them, but I know something else to be true. They weren’t my best.

I could feel the difference between a story I’d wrestled with and a story I’d assembled. The thing is, the numbers didn’t punish me for it, so I didn’t change course. That’s the honest version. I wasn’t cynical about it, I was just responding to what the market was telling me, and the market was telling me that good enough, on schedule, was a viable business.

Last year, the market stopped saying that.

What actually changed?

The obvious answer is volume. Anyone can generate a manuscript now, and a great many people are. But volume alone was never the problem, Amazon has been a firehose since 2012.

What’s changed is what volume does to discovery. When there were a lot of books, a good book could still be found by the machinery: ads, categories, also-boughts, a well-tended mailing list. When there are an enormous number of books, that machinery gets more expensive and less reliable for everyone. Ad costs go up. Attention gets sliced thinner. The signal-to-noise ratio of a category page collapses.

So the thing that gets scarcer isn’t books. It’s reasons to pick one.

And the most durable reason to pick a book — the one that has survived every platform change, every algorithm update, every format shift — is another human being saying you have to read this.

The good news, with an expiry date

Right now, most AI-generated fiction is mediocre. Structurally fine, sentence-level fine, and hollow. Nobody did the unglamorous work of revision. Readers notice. They may not be able to put their finger on why a book feels thin, but they close it and they don’t tell anyone about it.

I want to be careful here, because “AI can’t write a good book” is the kind of claim that ages badly and I’m not going to bet my career on it holding for ever. Assume it’s temporary. And also know that in the right hands AI can write a decent book. But the right hands are expert-level. An accomplished author with a knack for technology. And by the time they’ve poured their expertise into teasing the manuscript out of AI, the line between AI-written and human-written has blurred anyway.

But the general point sticks, most AI-written fiction is pedestrian. A bit like my second and third books :(

What doesn’t change is the top. Nobody raves about a middling book. Enthusiasm — real, unpaid, unprompted enthusiasm — only ever comes from work that moved someone. The robots aren’t competing in that category, and it’s the one bit of the playing field where more competitors don’t actually hurt you, because the competition isn’t for shelf space. It’s for the thing that happens in a reader’s chest at 1am when they should have gone to sleep two chapters ago.

Fewer than a thousand

Damon Courtney of Bookfunnel fame made this point from the stage at SPS Live in June, reaching back to Kevin Kelly’s old essay about a thousand true fans. The idea that you don’t need a mass audience, just a small number of people who will buy anything you make.

Damon’s twist was that the number might be smaller than a thousand now. I’d go further: for most of us the number that matters is the handful of readers who talk. Who put your book in someone’s hands. Who post about it unprompted. Ten of those are worth more than a thousand passive readers, because they compound and passive readers don’t.

You do not get those readers by being competent.

What this doesn’t mean

It doesn’t mean marketing is dead and craft is all. That would be a lovely thing to believe and it’s wrong. You still need the mailing list, more than ever, because it’s the one audience nobody can take away from you. You still need paid ads to prime the pump, because word of mouth needs mouths to start in. You still need covers that signal correctly and blurbs that convert.

All of that is the launch platform. What’s changed is what has to sit on top of it. The boost, the bit that takes a book from breaking even to actually profitable, is increasingly going to come from a small number of readers having their minds blown and telling people about it.

Marketing gets you the first reader. Only the story gets you the second.

Which means we have to get better

That’s the uncomfortable part, and I’m including myself squarely in it. And I don’t mean produce more, or produce faster. I mean get better: at premise, at hooks that dig in and refuse to let go, at characters people argue about, at the deeply unfashionable business of revising a draft until it stops being merely fine.

I’ve spent a decade telling indie authors that the business skills were the differentiator, because for a decade they were. I now think the differentiator is going back to the thing we all started out wanting to do, and doing it much harder than we did when the market let us off.

Nobody wants this answer. It’s slower, it’s less controllable, and there’s no dashboard for it. But it’s the one part of this business that can’t be automated, out-spent or out-published, and if we do it well enough, the readers do the marketing we can no longer afford.

If you want some direction to improving your craft, standby for a shameless plug:

Two of the courses in our Summer Sale go directly at this: Suzy K Quinn on writing a book people actually can’t put down, and Jennie Nash on the revision work that turns a finished draft into something worth raving about. They’re in the sale until it ends — code SUMMER30.

https://www.learnselfpublishing.com/bestseller

https://www.learnselfpublishing.com/revise

This is not coming from an anti-AI place. Regular readers will know I am what pro-AI authors call ‘AI forward’, rather it’s part of the nuanced approach to AI I am personally developing and something I would encourage you to do too.

‘Can I do this with AI’ has been the first question I ask myself before every task for more than a year. If it involves spreadsheets or a form of project management, the answer is an insta-yes. And thank god for that.

But I am also aware it is essential to keep me, the human, in the loop at all times.

I know, I know, if you’ve been banging the anti-AI drum you are shouting at me ‘urm yeah asshole, we’ve been telling you this for two years’. But this, as I say, is not about anti-AI, it’s about how it’s used to platform my humanity, not replace it.

This week I am sending 52 thank you emails to SPS Live speakers and sponsors. Of course the emails themselves will be written by me, albeit I am sure the recipients will understand the top and tail will be very similar for each person and even before AI they would have understood some copy and pasting had taken place.

But using AI to create the draft, based on our spreadsheet of names and roles, with the correct email addresses and ‘cc’ in place, has saved me days. Literally days because with my aversion to repetitive tasks, I can only do five or six at a time.

And that tedious process of copying the email addresses, pasting them into a new email, copying the previous email’s words, pasting them, amending them, was so mind numbing that I probably got lazy in the past and the emails were very similar.

Having all 52 emails, sitting in my drafts, ready for me to open, amend and send, is transformative. It has saved me ALL the boring part and made saying sincere thank yous easy and enjoyable. I now spend MORE time on the personal message. AI has allowed the humanity to take up more of the result, not less.

But I am also finding myself saying ‘no’ to AI a little more. If the end result is completely AI, I have to question whether that is helping me achieve my business aims, which revolve largely around me being a human!

So when it comes to email banner creation, newsletter composition, this sort of blog first draft for instance, AI is used with caution. (Only for proofing for this article fyi).

Just because you can use AI for a task doesn’t mean you should.

I have now amended that pre-task question into the world’s shortest checklist:

1. Can I use AI for this?

2. Should I?

I remain steadfastly AI forward, but those of us enthusiastic about the new technology, need to create our own temperance or we risk losing what makes us worth listening to.

I think this is a micro version of the overall implementation of AI across society. Eventually the complete nay-sayers will dwindle, and we’ll all end up in this hinterland of whether AI is right for a particular task.

Partly this is about protecting us from error-correction.

Humans are not error-free and ultimately that might be the only reliable sign that something is AI: the lack of mistakes.

I’m not just talking about typos - although that’s fine I’d like to eliminate as many as possible - but the od typo is at least a sign of humanity! (See what I did there?) It’s more about the bigger mistakes.

For us creators we need to understand that people follow us and our journeys as much as they consume our content. In my case I have made plenty of public mistakes. My first novel process was tortuous to the point of embarrassing and it all played out via podcasts and live shows.

I was inundated with well-meaning advice on what I was doing wrong and those interactions were very often the first contact I had with fellow authors I now consider close friends. As a minimum, mistakes we make are conversation starters, but actually I believe it’s far more profound than that.

Mistakes define us. So by definition, if we allow AI to erase our mistakes, we risk erasing ourselves. Had I used AI to plan, implement and publish my first novel, it may have been a significantly smoother process but I would have lost something incredibly important along the way. I would have lost a two-year story of failure and success that allowed people to empathise with me and understand me. I honestly think my inability to hide my failures is the reason some people actually like and follow me. It’s the recognition of frailty and the knowledge that you’re not alone in making mistakes that allows people to understand themselves and feel more confident on their own journey.

We want to live in a human world. A world where we feed off human frailties and mistakes as much as the successes. As novelists more than anyone else, we understand it is the flaws that make characters complete.

So caution is needed in our race to automate our lives.

AI will continue to take hold of much of my daily work, and as mentioned above: THANK GOD, but I’m not too entrenched to borrow from the angry anti-AI mob and their underlying truth (or fear) that AI is not human and it risks eliminating the essence of humanity. As with my general philosophy in life, all things in moderation.

For our creative industry the advent of generative AI comes with a shadow. Many of us (most of us, I believe) have quickly adopted AI tools in our day-to-day workflows, and it has been transformative. But the threat to creatives is not just about taking our jobs. Among some creatives there is a visceral anger that AI could even be considered capable of producing art. And that even if it did, even if the viewer liked the painting, or enjoyed the novel, that it doesn’t count as art.

The problem is, that discourse is useless. It’s like someone in Florida firing a gun into a hurricane. It might feel good to vent, but it will make no difference.

I’m sure ‘human-only’ creative spaces will emerge. A few publishing platforms are already setting out markers for this, although serious questions remain about how this can possibly be governed.

But it’s time to move on from our emotional reactions to AI to attempting to navigate an uncertain future. While of course I have respect for an author who refuses to touch any element of AI, I genuinely fear they are disadvantaging themselves for no reason.

None of us know what 2027 will look like, let alone 2030. And what’s happening out there, in the rest of the AI universe is where our attention should be, not arguing among ourselves about an AI figure on a cover.

In the past week I noticed three significant moments that indicate the pace of change we are facing.

Firstly, Anthropic, the owner of Claude AI, warned that AI will soon be capable of building itself without a human in the loop. If you’re a sci-fi reader, this is a terrifying moment. At what point does the AI believe itself to be superior to humans? Safety protocols seem barely mapped out, and yet here we go unleashing what’s called ‘recursive self-improvement’.

Secondly, and more positively, in China, instead of relying on the world’s leading battery scientists to refine the technology and eek out extra miles for cars, they asked AI to look afresh at the problem. The result was a new chemistry for batteries that is already providing a ‘five-minute’ charge for 250 miles of range in a car. That’s quicker than it used to take to fill up.

Thirdly, AI was tasked with producing a new vaccine, which it did, and it is light years ahead of where humans had got to in over 230 years. The AI vaccine works in a fundamentally different way and works on ALL variants as well as working in animals to prevent the next pandemic. If that is the case, think of saved lives. Seven million people died in the last pandemic, and tens of millions from Spanish Flu.

Three stories in a single week.

So what does it mean for us?

I’m not frightened for storytellers.

Everything I’ve described is AI doing what AI is is good at: searching vast spaces, spotting patterns, iterating faster than any human team alive. None of it is AI deciding what is worth saying.

A model can generate a competent chapter. But it hasn’t experienced fear in a cockpit when an engine flames out, it hasn’t fallen in love with the wrong person, or carried a grudge for thirty years.

That stake, the human behind the words, is what readers invest in.

So my argument is simple. The flood is coming whether we like it or not. We can spend our time regurgitating arguments about whether AI art “counts,” or we can use these tools to do everything around the story faster and cheaper. The research, the admin, the marketing, the translations, the covers, so we can pour more of ourselves into the one part no model can replace: the story, and the human who lived enough to tell it.

For some, including those with disabilities, that will include getting the words down that tell their story.

That isn’t a retreat into “human-only” purity. It’s the opposite. It’s putting the human back at the centre on purpose, and using every tool available to protect the time and the energy to do it well.

If you’re ready for that conversation, you’re setting yourself up to navigate what’s coming. Being worried is legitimate, of course, but being curious may save you.

The authors who thrive in 2030 won’t be the ones who shouted loudest into the hurricane. They’ll be the ones who learned to read the weather.

Warning, long post.

This week on The Self Publishing Show, Cissy and I are talking about metadata in the new era. During my note making I came to the conclusion that this is even more profound than we first thought and it very likely requires remedial action on our books’ listings.

For years, we’ve told authors that metadata matters. Usually, what we meant was fairly practical: choose the right categories, fill in your seven keyword boxes, write a decent book description that ‘sells’ your novel, make sure your subtitle is doing some useful work, and don’t accidentally bury your book in the wrong part of the Amazon store.

Amazon now has Rufus, its own AI shopping assistant. Readers are increasingly using conversational search. Recommendation engines are getting more sophisticated. Ad platforms are using more automated targeting. Meta ads, Amazon Ads, and AI-driven discovery tools are all trying to work out what your book is, who it is for, and when it should be shown to someone.

That means your metadata is doing more than helping with search.

It is helping AI understand your book.

In this article I’ll set out the new approach I believe we need to take.

Metadata is your book’s machine-readable identity

Let’s narrow it down to two principles:

We are replacing advertising with truth.

We are replacing maximum audience appeal with niche.

In the old world, metadata helped answer one main question:

Can a reader find my book?

In the new world, it has to answer a slightly different question:

Can a machine understand exactly what kind of book this is, and who might want to read it?

A vague blurb may have hurt you before but now it means certain invisibility. This is especially important for fiction, where readers are often not just buying ‘a book’. They are buying a specific reading experience.

They might want:

• a fast-paced British military thriller

• a cosy village mystery with humour and warmth

• a dark romantasy with dragons and enemies-to-lovers tension

• a psychological thriller about a marriage falling apart

• a Cold War spy novel with authentic period detail

• a police procedural led by a damaged detective

The more clearly your metadata communicates that experience, the easier it becomes for both readers and AI systems to understand where your book belongs.

AI discovery is more conversational

Readers do not always search in neat keyword phrases.

They may not type:

“British military thriller ex Royal Marine”

They may ask something more conversational:

“What’s a fast-paced action thriller with a former soldier as the hero?”

Or:

“Find me a book like Jack Reacher but with a British feel.”

Or:

“I want a cosy mystery series set in a small town, with humour, friendship and not too much gore.”

The search isn’t just a single sentence, it will become a conversation in which the reader will narrow down their optimal book. AI systems are built to deal with this kind of language. But they still need information to work with. Your book page, reviews, categories, keywords, subtitle, series information and author data all become part of the cloud of information around your book.

So the question becomes:

If an AI had only your book page to work from, would it understand your book correctly?

If the answer is ‘unsure’, your metadata needs work.

The blurb has two audiences now

Yes, readers will still use your blurb etc to make their human choice, that has not changed.

But the metadata also has to explain the book clearly enough for machines to interpret it.

That does not mean writing robotic copy. It does not mean stuffing in keywords. It does not mean flattening your blurb into a dull list of tropes.

It means being clear.

A blurb for the new era must communicate:

• genre

• subgenre(s)

• protagonist

• setting

• central conflict

• stakes

• tone

• reader promise

• relevant tropes or hooks

For example, this kind of line gives both the reader and the machine useful information:

A fast-paced British action thriller about a former Royal Marine pulled into a deadly conspiracy.

That is doing a lot of work. It tells us the pace, nationality/flavour, genre, protagonist type and broad story engine.

Compare that with:

When the past returns, everything changes.

That might sound dramatic, but it says almost nothing. It could be a romance, a thriller, a family saga, a ghost story or literary fiction. It might create mood, but it does not create clarity.

And in the AI age, clarity matters.

Stop thinking of keywords as hacks

One mistake authors make is treating Amazon keywords as a little box of tricks.

The better approach is to treat them as reader-intent labels.

You are not trying to trick Amazon. You are trying to describe the book accurately enough that the right reader can find it. Think truth not selling.

Weak keyword thinking looks like this:

thriller, thriller book, best thriller, crime thriller, action book, suspense novel

Better keyword thinking is more specific:

British military thriller
former Royal Marine action thriller
fast-paced revenge thriller
special forces conspiracy thriller
short action thriller novella
ex-soldier hero thriller

Specificity beats generic noise.

The best keywords are not necessarily the most obvious ones. They are the phrases that connect the right reader to the right book.

Categories are classification signals, not just chart opportunities

Categories have often been treated as a tactical game.

Which category is easiest to rank in? Where can I get a bestseller tag? Which niche has the least competition? Those of us who are fans of the data guys like Alex Newton and Dave Chesson have long used their analytics to predict the above.

But there is now a greater danger than ever before in chasing categories that do not truly fit the book.

In the AI-driven discovery world, categories are not just chart positions. They are classification signals.

If your book is a cosy mystery but you place it somewhere that suggests crime mystery, you may get visibility in the wrong place. More likely the system will note the disconnect between data and consumer and your book will effectively be flagged as a poor performer in search results.

The test should be simple:

Would a reader browsing this category be pleased to find my book here?

If the answer is no, you are probably creating confusion.

The whole book page needs to tell the same story

This is where many authors go wrong.

The metadata does not operate in isolation. Your title, subtitle, cover, categories, keywords, blurb, reviews, A+ content and ads all work together.

Or at least, they should.

Imagine this:

• The cover looks like literary fiction.

• The subtitle says psychological thriller.

• The blurb reads like a police procedural.

• The keywords target domestic suspense.

• The Facebook ad says “for fans of Jack Reacher”.

That is a confused signal.

The reader is confused.
Amazon is confused.
The ad platform is confused.

And confused systems do not produce efficient sales.

The goal is alignment. I’ve been teaching this as the Swiss Cheese model for a while. Make sure your holes are aligned so the maximum number of people flow through each stage.

Your cover, title, subtitle, blurb, categories, keywords and ads should all be pointing in the same direction. They do not need to repeat each other word for word, but they should describe the same reading experience.

Reviews are metadata too

Annoyingly, this is also true.

Your official metadata is what you tell Amazon and readers about the book. But reviews are also part of the information ecosystem.

Reader language matters and now is the time to consider whether you’ve inadvertently mis-sold your book in the past.

If your book description says:

A laugh-out-loud cosy mystery

but the reviews repeatedly say:

Darker than expected
More violent than I thought
Not really cosy
Reads more like a police procedural

then that is important information.

You are attracting the wrong readers. In the new era that will mean attracting no readers.

AI systems can learn from reader language. Even if you cannot control reviews, you can learn from them.

Look for repeated phrases in your reviews:

• fast-paced

• emotional

• gritty

• funny

• atmospheric

• twisty

• romantic

• violent

• cosy

• cinematic

• slow burn

• great worldbuilding

• like Reacher

• like Agatha Christie

• like romantasy with teeth

Those phrases can tell you what readers think your book actually is.

Sometimes your readers understand your positioning better than you do.

It didn’t occur to me for one second that I wrote mysteries. I thought I was writing traditional military thrillers.

But my readers told me otherwise.

My favourite review remains ‘Most boring book ever’! At the time I shrugged, as we’re supposed to, and just assumed the reader was not in the mood.

But not long after, I got a much more helpful 2-star:

I rewrote the blurb to lean heavily into the mystery aspect of the story using keywords like ‘secret’ and ‘uncover’ and the book did better (it currently has 2,200 ratings at 4.5). But in all honesty I need to go back and start again. In the past a small tweak to the description may have helped, but in the new era I need to be crystal clear, this is a mystery-suspense not an action-thriller or face AI invisibility.

A simple metadata audit

So what can you and I do right now?

Here is a practical exercise.

Take your most important book — usually book one in a series — and write one clear positioning sentence:

This is a [genre/subgenre] about [protagonist] who must [central conflict] before [stakes], ideal for readers who enjoy [tone, trope or comparable authors].

For example, I would write this for Kerry J. Donovan’s Ryan Kaine books:

This is a fast-paced British military thriller about a former Royal Marine who must uncover a deadly conspiracy before innocent lives are lost, ideal for readers who enjoy action-heavy thrillers with tough heroes and military detail.

Now compare every part of your book page against that sentence.

Does the cover support it?
Does the title support it?
Does the subtitle support it?
Does the blurb support it?
Do the categories support it?
Do the keywords support it?
Do your ads support it?
Do the reviews broadly confirm it?

If the answer is no, you have found the problem.

What to fix first

If you are going to audit your metadata, start with these areas.

1. The first paragraph of your blurb

This is prime real estate. Do not waste it being coy.

Make sure the first paragraph gives a clear sense of genre, protagonist, situation and tone.

2. Your subtitle

The subtitle can be a useful classification tool, especially for genre fiction.

For example:

An Acid Vanilla Assassination Thriller

or:

A Ryan Kaine Action Thriller Novella

That sort of subtitle helps readers and systems place the book quickly.

3. Your seven keywords

Replace generic terms with specific reader-intent phrases.

Think about genre, protagonist, setting, tone, trope and comparable reader appetite.

4. Your categories

Make sure you are not chasing visibility in a category that misrepresents the book.

5. Your ad copy

Check whether your ads are promising the same experience as your Amazon page.

If the ad attracts one kind of reader and the product page describes another kind of book, conversion will suffer.

What not to do

The answer is not to stuff your book page with every possible keyword, clarity is more important.

Do not call your book romantasy if it is really epic fantasy with one romantic subplot.

Do not call it cosy if there is violence.

Do not call it a psychological thriller if it is really a detective mystery.

Do not chase fashionable labels unless the book genuinely delivers on them.

This is not just an ethical issue, although that matters. It is a commercial issue. Misleading metadata attracts the wrong readers, produces poor conversion, damages reviews and confuses recommendation systems.

And while that has always been true, in the new era everything is on steroids.

The aim is not to game the algorithm.

The aim is to remove confusion.

Metadata is positioning

The old way of thinking about metadata was administrative. The better way to think about metadata is positioning.

You are telling readers, retailers, algorithms and AI systems what your book is and why someone might want it.

That means metadata deserves proper attention.

It should be reviewed.
It should be tested.
It should be aligned with the cover and ads.
It should evolve as you learn more about how readers respond to the book.

That makes it one of the most important parts of modern book discovery.

And, frankly, one of the easiest things for authors to improve.

A quick homework task

Open the Amazon page for your most important book and ask yourself:

Would a stranger know exactly what kind of book this is within five seconds?

Then ask:

Would an AI that reads every letter and looks at every image on the page, have a crystal clear idea of who to recommend it to?

You do not need to trick the system.

You need to tell the truth about your book more clearly.

That is the new rule of book discovery.

James

In the early hours of yesterday morning (early in the UK) Draft2Digital, the leading aggregator of indie books to the book-selling websites, announced a change to their fee structure.

If you aren’t aware, this is the gist. In the past setting up a new account and uploading your books came at no cost to the author. However D2D takes a 10% skim of the royalty and hands the author the rest. In contrast, D2D competitor PublishDrive, allows authors to keep 100% of their royalty but it charges a monthly subscription, which starts at $9.99.

Yesterday, in response to a significant increase in substandard content books uploaded in the age of AI, D2D introduced an annual fee and a one-off charge to open a new account. The one-off fee for new accounts will be $20. The annual fee will be $12/year, unless the account earns $100 royalty share or more (about 32 book sales in the year).

It’s clear that D2D is struggling with the influx of low-value books. Anecdotally, a D2D employee told me earlier this year that on the same day Charlie Kirk was murdered, they had a clutch of ‘Charlie Kirk’ biographies submitted. These books were no doubt generated by AIs from a Wiki page or other links. They almost certainly contained mistruths and inaccuracies and in the time it took to produce them, they will not have undergone any serious fact-checking or editing. Simply put, the world could do without these books. They rip-off innocent readers.

TikToker Sophia Smith Galer reported finding a ‘biography’ of her for sale…again full of hallucinations and lies parading as fact.

Content farms that generate a 1,000 books a month are a real thing. There is no regard for quality, just the appearance of a normal book that will make a couple of sales. That’s all they need for a decent profit. They take advantage of the ‘free-to-enter’ indie publishing landscape.

For us indies, launching our books into this maelstrom of nonsense has made marketing significantly harder. Plus the retailers are fed up and it’s too easy for them to see the whole indie community as the problem.

And, there’s evidence that Amazon is favouring big publishers over indies with their algorithm change. That could be Amazon’s way of dealing with the content farms, and indies are being disadvantaged. In contrast, D2D want to do something that will stave off the majority of those mass submissions and leave the field clear for us indies and our few books a year.

So you would think the indie author community would welcome these changes. Especially as the annual fee is an entirely inconsequential amount.

You would be wrong.

‘Angry’, ‘Hurt’, ‘Frightened’ were just three of the adjectives I’ve read in the last twenty-four hours. Many fellow authors have announced their intention to leave D2D forever.

I’m sorry, what?

$12/YEAR!!!

Can we just focus on the charge that has precipitated this backlash.

$12/year. $1/month. $0.03/day.

If this fee were significant, it would be entirely reasonable to attack me for being bourgeois and wealthy enough to be oblivious to what life is like for poorer families. But I’m sorry but it’s hard to imagine the person living in the West, for whom $1/month is the cause of such angst. Even someone living on the streets in London would in all honesty not notice $1 (£0.73) across a month.

Do those same authors loudly protesting and threatening to leave D2D ever buy a coffee? Do you know how long it would take to pump $12 worth of petrol in the UK? Seven seconds.

It would take seven seconds to pump the entire D2D annual charge. An annual charge that allows you to distribute your books to every major bookseller in the world and gives you access to a host of useful tools such as a book formatter.

One of the objections seems to be that authors who don’t make money, ie those that sell fewer than 32 books a year through D2D are unfairly penalised. But the problem D2D is trying to solve only works if it focuses on the low-earners because that’s the strategy of those badly written AI books; for them it’s a quantity over quality game.

To get around that, D2D could simply charge everyone $12/year regardless of sales and maybe they should have, but only for optics. But how does someone else paying $12 solve the problem of your own $12?

I do wonder if the AI onslaught has created a level of paranoia in the indie community that stops us acting rationally. It’s never enough, apparently to look at a policy and weigh its pros and cons, the community seems to reach for the pitchforks immediately.

How about we think of it this way. Largely hidden from view, the age of AI is turning our industry upside down. Navigating this transformation will be painful and challenging. D2D is doing what it believes is the right thing to fend off the onslaught of poorly written, mainly non-fiction, AI books. That leaves more space for discoverability for all-human authors and indeed those who use AI but use it as a tool to produce a decent book.

It also preserves D2D’s relationships with the retailers and it starts the process of separating the indie community of writers from the content farms. Surely this is a good thing?

And all for the cost of a banana a month.

I’ll start with an apology for causing alarm and upset with my previous post.

The idea of a $300 charge per book to upload to KDP sent many authors into a panic, immediately foreseeing the end of their author careers. I absolutely did not want that and it was my fault for not ‘work-shopping’ the idea more thoroughly before airing it.

(There was also an unnecessary pile-on from some, using insults and insinuations that I choose to ignore, while noting the names…).

Clearly the figure of $300 was the main issue. Even if it covers all derivatives, including translations, it would adversely affect more authors than I considered.

But I will stand by the central theme of the post. And a more modest charge per title is one way of addressing it. I discuss this for purely discussions sake. And we should be talking about this, not shutting people down who are trying to address an industry wide issue.

So whether it’s $50 or even less (maybe with three free books a month), there is an argument for a charge.

We’re all aware of the radical change that has happened to Amazon KDP in the past three years, following the rise of generative AI.

You will know from this blog, I love the new technology and I’m interested in the folks who use AI to write and sell books, some of them are very impressive. It’s a new era of publishing.

But...when KDP slapped a three-a-day limit of the publication of new titles, I realised, as you did, that there were likely thousands of accounts uploading more than 100 books a month. (There still are, of course, having multiple KDP accounts is hard to police).

Set that fact alongside last year’s reported algorithm change, which seems to hold indie titles back from the chart positions we used to enjoy and I suspected KDP was solving a reader-experience issue by punishing indies.

Amazon, understandably, wants to protect the reader experience. They are up against AI generated novels, thousands a day, and AI driven marketing carefully balanced to produce a small profit per title. It’s a new business model that relies on quantity over quality. 25 sales of a free-to-produce AI book with an AI cover and blurb, multiplied by 100 a day, is a very profitable business.

So I thought...does it matter to Amazon that 1,000,000 new books sell 25 copies each, rather than 100,000 new books selling 250? Yes and no. No, ultimately the profit is the same for Amazon, albeit with higher storage and processing overheads. But yes, if the reader experience is affected and readers drift away from the platform.

So Amazon’s answer seems to have been firstly the three-a-day limit (not really enforceable) and then to make the algorithm very conservative, so the books you see in the charts come from accounts with big sales history.

A friend noticed on Monday this week, of the top 100 titles in the .com KDP chart, only THREE were from indie publishers. That’s a crisis. That’s a fundamental change to the nature of KDP, a platform built on the back of indie writers.

A fee to upload a title undermines the quantity business model. If done right, it favours indie writers, makes no difference to trad and renders the 100-books-a-day model obsolete.

Many of those who fiercely objected to my idea are prominent AI writers. As I said, I’m not anti-AI, even writing. But using AI to craft a novel (I’ve seen it done, it’s hard work) is one thing. Setting up a workflow that produces 100 romance novels a day, auto-edited, with auto-generated AI covers, is another.

‘The market will sort it out’ was a common response. But the market DOES now sort it out - those books sell just a few copies each. However the knock on effect because of the sheer volume, is to make smaller indies’ discoverability extremely difficult, to say the least. Arguing that discoverability remains the same as it did four years ago is nonsensical and not evidence based.

The problem is that producing a book that sells 25 copies and then dies because it’s bland and generic, is profitable if you do it many times every day. Charge $50 (refundable after you earn $100 for instance) and you eliminate that loophole.

Use AI to produce fewer but better books, which is what many indie AI writers do, serves readers better and we can co-exist together competing for the same pool of sales. After all readers only spend so much to go around.

So, on reflection the $300 attention-grabbing headline did just that, but instead of igniting a debate, it set the dogs loose. On me! I hope this explanation helps to fuel a better discussion.

You can hear the genesis of the idea on this week’s Self Publishing Show episode, and a hastily recorded response we added last night...

Click here to watch or listen to the episode if you’d like to hear me talking about it.

Oh and I reserve the right to delete unnecessarily rude comments. I’m too tired for that shit and I have enough of my own anxiety, thanks.

James.

Three books a day, for a platform designed for individual authors!? Who, you might reasonably have wondered, is uploading more than three books a day?

Of course generative AI was the answer. It has enabled the hyper-fast production of books and almost overnight, KDP with its zero barrier to entry, felt the result.

Most of these fast-produced and published books are non-fiction; cookery, self-help, celebrity biogs, anything that’s easy to generate in ten minutes on Claude. Some of them are novels.

The result is a saturated marketplace. To be fair, KDP has long been saturated. There were always thousands of books a month appearing on Amazon, but the exponential rise in numbers is changing the landscape. Discoverability for authors who spend a year writing and publishing a single novel, is now harder than ever. Their handcrafted stories are lost in the flood of AI titles.

I think it’s time for a radical move by Amazon KDP. Charge $300 per title to publish.

You’ll get the $300 back after the title has made $500 in royalties ($1 added for every $1 earned in royalties over $500 until the $300 is paid back). Think of it as a reverse advance.

Overnight, the wash of low-content, bland, generic AI books will disappear.

The business model for that AI production line relies on a few sales to unsuspecting consumers before the bad reviews and lack of interest leave the book buried.

But 25 sales multiplied by 1,000 books a year is a business that pays regardless of the quality of content.

Not all AI-generated content is without merit. You will know from this page, I am very open-minded about the new technology’s abilities. Writers such as Coral Hart produce AI-generated books that sell, using their experience of writing dozens of novels in the past combined with deep market understanding and a talent for expert prompting. Her system produces upmarket books compared to the usual AI generated mulch.

But I doubt Coral is producing three books a day. In fact I’d be surprised if it was even three a week. So AI writers such a Coral will also benefit from this change. It’s a system that rewards those who put the most effort in. This change will force those writers to spend even more time on a single title to ensure it will earn back the advance.

I’m not naïve enough to imagine the Amazon board is sitting around wondering how they can make KDP better for indie authors. In fact, I am certain they sit around wondering how to make KDP better for Amazon shareholders; that is their job, after all. And there will be voices around that table that see this sudden glut of titles that sell little-and-often as the new model. Does it matter to Amazon if they sell 25 copies of a million new books, rather than 2,500 copies of 10,000 new books?

Well, yes, I think it does. The end result in $ might be the same, but Amazon also has to consider the reader experience. This is an oft-quoted reason for policies regarding content changes to existing titles. Amazon cares about the reader experience.

So I hope there is at least one voice around the Amazon boardroom table that wonders if the dilution of KDP is serving their loyal customers in the long run.

A $300 charge doesn’t stop the author who wants to write and publish, even if they use AI, but it does stop the ‘author’ who produces bland AI books with the sole intention of making a handful of sales as their business model. And that’s a good thing for the reader, which should, rightly, be Amazon KDP’s primary concern.

James.

You’ll know I am not on the fence about AI. I see our industry being transformed by the new technology and I feel I have an obligation to adapt or risk becoming irrelevant. (This is not unique to our industry, of course).

But even within that acceptance, I understand how actually writing books with AI is still a step too far for many. I also understand what I don’t understand. And my own experience experimenting with using AI to generate writing has been dismal at best. The writing it comes up with is generic and a long way from something I could consider publishing or offering to readers.

So I was intrigued when I saw the writer Coral Hart enthusiastically telling a New York Times reporter how she makes six figures a year, across a range of pen-names, generating books with AI.

I wanted to know more and I also wanted to make sure she was OK, after putting her head so far above the parapet. (A pity that I had to do that). So I dropped Coral a note and a couple of days later we had a call in which she demonstrated her system.

It was absolutely fascinating. And, as I suspected, simply ‘pressing a button’ is not going to give you a viable novel from AI, not now, not ever. In fact, I would say there’s at least a week’s work with AI to get a useable platform on which to start a first draft. A week that includes market research down to trope level, and reviewing the AI’s initial work on characters, plot beats, and so on. After all of that, you have a first draft that (just like our first drafts), requires development editing, followed by line and proof editing.

It’s definitely much faster than the twelve-plus months it takes me, but still not the ‘press of a button’ that so many assume.

More importantly, it is a wholly human led process and something else I will say for the record: watching Coral Hart write a book using AI? It is a creative process. She is able to make this work because she’s written 90 books, she understands story, tropes and the commercial fiction market.

As I have been saying for a long time, AI is the ultimate ‘Garbage in, Garbage Out’ platform. The more you work with it and input your creative talent, the better it will be.

This is good news for writers, not bad news. AI will streamline our process and make us more efficient, but it cannot make a non-writer, a writer. That creative flair you have for storytelling? That understanding you have of the reader market? That will see you thrive in the new era.

I know this subject makes many of you feel uncomfortable, but I promise you a thorough understanding of the reality of AI is essential for you and me to navigate the future. Shouting it down and aiming barbs at fellow writers is not useful and not cool.

(If you want to know which series I’m talking about, it’s in code at the end of the post).

It was gripping all the way through. The MC was SO HATEFUL I actually had to turn away from the screen when he was ‘being nice’ to his victims, duping them into believing he was a wholesome good guy with their best interests in his heart.

Huge kudos to the actor, who we know from comedy. He was chillingly good.

So it all boiled down to the finale, of course, because that’s how psych thrillers work.

But our MC didn’t get his comeuppance. Instead of the wonderful pay-off we had earned by watching his mounting deceit and evil, we got to see him sail off into the sunset, presumably ready for Season 2.

And this is not the first case recently.

Endings are hard, we all know that, but unless I’m missing something, it is traditional for the baddie to get caught and for the victim (and by extension) the reader to enjoy his/her ultimate failure.

I don’t understand this trend for the baddie getting away with it? If you want the series to have legs commercially just do what every previous writer has done from Agatha Christie to Freida Mcfadden; make the MC the good guy. OR, at least have him banged up, or thought dead. As The Night Manager proved, death is no barrier to a come back.

It’s possible there was some shallow ‘eat the rich’ message in here, as the victims were wealthy and maybe we were supposed to root for the killer? But the killer was so unhinged, NO-ONE apart from a fellow psycho is rooting for him. And the family, despite the usual levels of familial dissonance, loved their kids (and loved their dog). So if we were supposed to want them dead, that wasn’t written into the script.

Justice is satisfying. Injustice is annoying when it comes just before the end credits.

There’s a reason good authors stick to the same outlines and plot beats. Because a thousand years of storytelling has taught us some thing should not be played with.

____________________________________________________________________________

Obviously naming the series would be a spoiler, but if you want to know…

Using A=1, B=2 etc, here it is:

13, 1, 12, 9, 3, 5, 15, 14, 1, 13, 1, 26, 15, 14, 16, 18, 9, 13, 5

The comments were mostly supportive, but as per, the criticism was loudest.

I should have the intelligence and maturity to empathise with writers and artists who are scared for their careers, but in reality, my enthusiasm for the new technology means I’m blinded by the upsides and possibilities.

As a novel writer, I am happy to embrace the arrival of generative AI.

I’ve been fascinated with technology my whole life and this era is on a par with the digital revolution of the 1990s. That revolution has transformed the world. Not all for good, in fact the level of vitriol (in a few cases), that I received on the previous post is an example of the dark side of the digital world we live in, where it’s common for people to attack from behind a keyboard in a way they simply wouldn’t do in real life.

But the benefits of the digital world? Immeasurable.

The benefits of AI? The same. The advances in healthcare alone are astonishing. AI, with its ability to consume and process petabytes of data is finding links and cures that were beyond humans even with the most powerful pre-AI computers.

The legal world has benefitted from AI doing a lot of the grind, freeing up lawyer hours for the less straightforward casework. In fact in most industries, AI is improving and advancing existing practices.

But most of all, for me, it is the democratisation that comes with tech revolutions. The access to information has elevated everyone in society and levelled the playing field between those with access to an expensive education and those born in countries where school is a luxury.

In the author world, AI has closed the gap between those authors lucky enough to have wealth and those who live month-to-month. Translations, covers, ad design, copy editing, proof editing, developmental assistant are just some of the services available for a few dollars that were previously beyond reach for many writers.

And while of course I sympathise with the bank teller, supermarket worker and factory worker who have been replaced by machines, I find it hard to sympathise with a cover designer, translator or editor who instead of embracing AI and transforming their workflow and productivity, chooses to ignore it, deliberately setting themselves at a disadvantage and then complaining that others are stealing work from them.

But let’s address, head on, the more challenging aspects of the rise of AI in the creative industries.

In the author world, this can be summed up in two key areas:

• AI writing books for us.

• LLM companies ‘stealing’ copyrighted work to build their companies for profit.

Firstly, I have no issue with writers using AI in the writing process.

In fact the competitive author in me is all too happy if they rely on AI for their entire work, because even a mediocre novelist such as me can out perform any of the current crop of LLMs. The stories they produce without very specific guidance are perfunctory at best. They lack, by definition, humanity and soul and it’s easy to stand out against them in a crowded market.

But by working closely with AI, using your humanity and life experience to craft a compelling story while relying on the technology for writing help, can produce a good novel. In fact for writers with some form of neurodivergence, AI can be a saviour. Bridging the gap created by their cognitive limitations. And what a fantastic thing? Someone’s dream of being a writer is fulfilled, like an amputee embracing the latest technology in prosthetic limbs.

I don’t need to encourage writers to use AI, because as someone very publicly in favour of the new tech, I can tell you writers confess to me all the time that they use it in their workflow. I use the word ‘confess’ because that’s how it feels to them thanks to the regular witch-hunts led by the loud hold-out minority.

I’m drafting my fourth novel, Spider, at the moment. I’m obeying the orders from my friend who’s helping with the story development to press on drafting without looking back. The next pass will be to refine the story, but after that, when it comes to polishing the prose, you can bet your life I’ll be AI-ing the hell out of it. That final refinement, copy editing or whatever you want to call it, is a perfect match for the logical skills of the computer.

And I’ll tell my readers. I intend to add a note in the book explaining how I used AI, as I suspect some of the mob will come after me once it’s published.

Secondly the ethical issues surrounding the wholesale sweeping up of data to train the LLMs without regard for copyright or even payment to the creators of those works…

The payment for the works is the lesser issue. If after a long and expensive court case, OpenAI and Anthropic each pay me for the cost of my three published ebooks, I won’t be retiring on the $15. Even if you have 100 published books, it’s a relatively small amount next to the income you are likely making.

Unauthorised use of our work to create a commercial entity is the bigger issue.

But let me ask you this? When a myriad of thriller authors wrote books about a former military man, now a loner who wanders from town to town solving crimes and helping the less fortunate, should they have paid Lee Child?

They read his books, watched the Jack Reacher films and TV series and decided this was a lucrative market to be in. They wrote novels that followed the same basic premise, sold them for profit and some case built very health self-publishing empires.

Really, what is the difference?

If we say to Anthropic and OpenAi what you did was illegal, how will that be policed at individual creator level?

When Flashdance appeared a couple of years after Fame (I’m showing my age here) we all knew what had happened. Alan Parker had a huge hit with Fame and he created a market for that type of inspirational journey film. The similarities are stark and even Irene Cara, a star from fame and the singer of the iconic theme song, was asked to write and perform the title song for Flashdance.

Was that stealing? Again, really, what is the difference? Without Fame it’s hard to see there would have been a Flashdance. Without Top Gun there wouldn’t have been Iron Eagle (some might say that would have been a good thing) and without James Bond, there’s no Jason Bourne.

The LLMs needed training like humans. They operate like people.

What is the alternative? Do we say that a new piece of technology that’s going to assist musicians, it cannot listen to any previously recorded and copyrighted music to learn?

Would you be able to write a novel without having previously read copyrighted material? And you sell your books for profit, right?

AI’s here. It’s amazing at some things and awful at others. You are creative and talented, don’t ignore something that can make you better. Get it to do the grind you don’t want (copy editing!) and free your creativity to write even better and more amazing books.

But she couldn’t use it. Not for copyright issues, not because it would breach any retailer’s terms… She couldn’t post it because she was scared of her fellow authors.

The image was created with AI. Of course it was, how else in late 2025 are you going to get the absolutely perfect image for one particular story?

Unlike me, she writes in romance, a genre I love, full of people I admire, but one that’s also known for its online drama. A small but aggressive contingent has made pile-ons and blacklisting almost routine.

A small group of romance writers, possibly in conjunction with some readers, but I haven’t seen much evidence this comes from readers, have formed themselves into a secret police outfit. They scour Amazon for what they see as the tell-tale signs of AI in any aspect of a book and then they mobilise their army to destroy it, and the author.

Why? Good question.

Maybe they’re still bitter about the car’s devastating impact on the horse and cart industry? That is, after all what’s going on here. And it happens, predictably every time there is a transformational technological development. Most of us embrace it whole heartedly, adapt our working processes, even if that threatens our old way of life, and make the most of the change.

But for some, there is a false belief that new transformative technology can be suppressed. That the genie can be put back in the bottle and that it’s the right thing to do, so that we continue to work in the way we always used to. In this case, ‘protecting’ the jobs of cover designers.

Let’s put aside the fact that you cannot in fact put the genie back in the bottle and let’s ask why you would want to? Why now? Why didn’t we freeze the cover design process before the advent of Photoshop (1990)? I’m old enough to remember traditional designers criticising advertising imagery they saw as ‘obviously Photoshop’. I doubt there is a designer left on the planet who doesn’t use it today.

The practice of trying to stop technological progress is as old as time. The word ‘Luddite’ is often used to describe someone who is less than effective at using a piece of technology or who doesn’t understand it. But the Luddites fully understood the tech, in their case the automated weaving machines in 19th Century factories. They physically smashed them up to prevent them taking workers’ jobs.

It was of course futile and as with AI today, they needn’t have bothered. The automation that swept through English factories in the Napoleonic era saw an explosion of production and the hiring of many more workers.

It’s not uncommon for a tech breakthrough to lead to more jobs, not fewer. How many people does the IT industry employ? It’s an industry that literally didn’t exist in 1945. And of course the advent of the personal computer led to wild claims that we would all be replaced.

Embrace AI., I say. Let the courts decide on the rights and wrongs. Use it to copy and proof edit, to design your covers, to analyse your advertising data. And if you are an editor, cover designer or ads professional, increase your own productivity and offer new services as a direct result.

I am a techno-thriller author with specific image requirements of military aircraft from the 1960s. Until recently my second novel Dark Flight had a futuristic x-plane on the cover to represent the fictional x-plane in the book. Trouble is, it was clearly an actual XB-70 Valkyrie, a real aircraft from the 60s. It’s something a lot of my readers noticed.

Along comes AI and in seconds I can create an even more atmospheric cover with a wholly fictitious aircraft, much closer to the one I had in my imagination.

None of us are immune from disruptive changes to our industry. In fact the irony is those romance writers waving their pitchforks are part of a multi-billion dollar disruption of the publishing industry. Why don’t they campaign against indie publishing taking all those trad jobs?

In the end my friend, had to give her A.I. designed character to her regular (human) cover designer and ask her to create it, as near as she could, without using A.I. How ridiculous is that? All because she feared the pile-on.

So stop it please. In two years this will all look silly as there will not be a cover designer anywhere who does not use A.I., or an author who doesn’t use it throughout their production and marketing process.

James.

Targeting is shifting from the largely manual process we’ve used for a decade to an algorithmic AI system known as ADVANTAGE+.

A year ago, I would still have advised people to keep manual control over some key areas, such as placements (where your ad is shown). My books, for instance, appeal mainly to men aged 55+, and they still inhabit the Facebook News Feed. So what’s the point of wasting placements on Instagram Reels?

But we need to guard against becoming set in our ways. Just because that was true in January doesn’t mean it’s still true in December. Which is why we test.

It’s something we teach in the Ads for Authors course, and it remains our mantra.

In the next couple of posts, I’m going to lay out a useful methodology for testing, because, like anything worthwhile in life, it’s depressingly complicated!

SET A SPECIFIC AIM FOR YOUR TEST

First, you must decide what you’re testing. The list of possibilities is long, but here are the fundamental elements you should understand for your books:

• Creatives – what type of images work best?

• Copy – what type of language leads to sales, not just clicks?

• Demographics – men vs women vs undeclared; age ranges

• Geography – Eg: USA vs UK, and sometimes specific regions or states

• Placements – where your ad is shown

• Interest Targeting – manually choosing interests vs leaving it to ADVANTAGE+

• Campaign type – traffic vs engagement etc.

In my example, I want to test ADVANTAGE+ vs manual campaign settings. I’ve chosen a three-stage test:

1. Full ADVANTAGE+

2. ADVANTAGE+ targeting with manual placements

3. Manual targeting with manual placements

This is a wide-scope test designed to give an indication of where Meta’s Advantage+ currently stands for authors, or at least for me and my books.

Just one tweak for each stage—which brings me to the second consideration:

BUILD A CONSISTENT TESTING FRAMEWORK

Make no other changes.

No promos during the test period, no changes to blurbs, and no new releases. It’s also advisable to shut off other ads such as Amazon or BookBub Ads. If you don’t want to risk losing momentum on those campaigns, leave them running but make no changes during the test period, and ensure the last change was at least seven days before your Facebook test begins.

Ensure all other campaign elements remain consistent throughout. In my case, I’m not testing creatives, which means the images and copy must remain identical in all three campaigns.

However tempting it is to tweak your ads—and we all love to tweak—it will render your test results invalid. Resist!

Now, we’re still not quite ready to run the test. There’s one more important step.

DEFINE YOUR RESULTS BEFORE YOU START

It’s essential to know before you begin what constitutes a result.

If you’re testing direct-to-Amazon product page campaigns, you’ll want to use Amazon Affiliate links and rely only on those results. These links tend to under-report, but as long as they under-report consistently, they are more reliable than benchmarking—where you simply compare sales during the test to the previous period. Benchmarking effectively doubles the test window, as you need an equal “clean” period with no changes, promos, or tweaks.

In my case, I want to track results more reliably, so I’ve chosen to run a lead generation campaign. This means I can literally count the email subscribers and divide that number by the spend to get a precise metric for comparison.

Lastly, consider whether those superficial results are the final results. In my case, not all email addresses are equal. So it’s important to tag subscribers as they enter my Email Service Provider so I know which campaign they came from. Over the following weeks and months, I can further refine my assessment by measuring subscriber quality.

For example:
Having 100 new email addresses at £0.30 each might look good against 100 at £0.90, but if 66% of the latter group buy a book or engage with my emails—versus only 10% of the cheaper subs—that flips the superficial results on their head.

Right—800 words in—and we’re finally ready to set up the campaigns.

It will take an intense few hours, but after all that planning, I recommend you set up every campaign at the same time and schedule the start/end dates in advance.

I did this a month ago, in the next post we’ll look at the results and I’ll let you know whether ADVANTAGE+ should be the new standard for authors.

James.

Since I stopped communicating with them I’ve receive almost daily reminders to pay quickly or else lose my chance to be considered for that Netflix series my novel so richly deserves.

Instead, I reported their account to the bank, Frost who are based in Texas. While I have not heard back from Frost, this morning I received this strange email from the scammers:

It’s extremely unusual, to say the least that a scammer would urge you not to pay, but to wait until they are ready. My guess is Frost Bank has frozen their account. I have no other evidence that’s happened, but I can’t think of another reason for the above request.

I dare say it won’t hold them up too long, but I am pretty pleased that we got to disrupt these thieves who prey on my fellow writers.

The reason I do this is because these scams are becoming a daily nuisance for indie authors. A constant barrage of requests from ‘Book Clubs’ (they do not exist), offers of film options and so on.

Jan contacted me to say she’d received a fairly believable email purporting to be from a ‘Creative Strategy and Author Collaborator’, complete with US credentials such as a phone number and bio pic.

By the way, if someone wants to try the phone number in the email please let us know what happens, I assume it will be a message, if the line is connected it all. (I would call myself, but annoyingly my UK phone won’t let me, even though I’m still in the States.)

Jan was smart when she received this. Firstly, the book ‘Amelia’ chose is a short story and the language used in the email didn’t fit. Even though I’m certain Jan can write with ‘lyrical stillness’ and she has ‘storytelling that is both timely and deeply human’!

She gave the email to ChatGPT to analyse, which is an excellent idea. The AI soon picked up on the likely AI nature of the writing.

Talking of ‘Amelia’ another ‘Amelia’ has been in contact with another author who wrote to me. Jessica got a friend request from a well known indie author and after chatting, the author suggested she tried an amazing marketing guru who had done wonders for her.

I had the start of this scam recently, when I received a friend request from Melanie Harlow. I have friends who know Melanie and I was quickly able to ascertain the person on X was an imposter, but I suspect the recommendation of a marketing professional was on its way.

Happily Jessica got suspicious pretty quickly about ‘Amelia’ the marketer and she referred the contact to https://writerbeware.blog/. This is a site run by Victoria Strauss and a small team. They do far more digging than I can and the blog is a great repository of scam evidence.

I will keep flagging these scams as well. The more we talk about it, the more authors are forewarned.

I am hoping AI, the cause of many of the scams, will come to our rescue soon. As shown above, it will be pretty good at detecting scams and could act as a filter in our inboxes, or at least add a line to each message warning us to be cautious.

For now, please spread the word to your writer friends.

James.

It’s obviously a scam, but I wanted to know exactly how it works and I very much want to raise awareness in the author community.

So we got to the business end of the process in the last post. They require a set of strictly formatted documents, which they claim you can DIY but they quickly reject your versions and insist you use professional services from a sister company.

I went along and sent my ‘substandard’ documents to the service company.

Did you guess how much they would ask for?

$955 is the lucky winner.

I was getting the point that I would have to pull out, obviously not wishing to pay these people. I was also concerned about sending them a PDF of my novel.

But AI came to my rescue again…I asked Claude to create a version of my manuscript, with the original opening two chapters, but some made up nonsense for the other 100,000 words.

Claude was very much up for it.

We went with option 2 - circular, repetitive. It was quite trippy reading it:

So now I feel confident I can submit my ‘novel’. I thought it was worth doing this, just so I can find out exactly how I was supposed to pay them.

Desiree Brown, commenting on my last post, suggested it would be Wallmart Gift cards - a usual untraceable option favoured by money launderers. The actual request surprised me:

This is a genuine US bank account at a Texas based bank.

A surprise to me, but some research shows it’s common for overseas scammers to use a US based account, often in Texas, using a local, either someone who is part of the scam and on a cut, or stolen ID details.

The great news is, I was now able to report them to Frost Bank via their Fraud Hotline (I’ll let you know how they respond).

So that’s how this particular scam works. And it does work. Not for me, but they wouldn’t be doing this if it didn’t. As I mentioned before, I suspect those who fall for it are too embarrassed to talk openly about it. They shouldn’t be. AI makes these approaches appear credible.

Our only weapon is awareness. If you’re an author, please tell your friends.

And remember, if an offer comes in that looks too good to be true…is is.

James.

Most recently I received a DM on X purporting to be from a producer at Archewell Productions - a company owned by Harry and Meghan.

It isn’t, it’s a scam.

How could I tell? Firstly, any professional production company pursuing an author for an option would do the minimal research and contact me using the information on my website, rather than sending a DM with formatting errors, on social media.

Secondly, after I responded, the first actual email came from ‘Ben Browning’, a producer who left the organisation some time ago (found via a simple web search).

Thirdly, if it’s too good to be true, it probably isn’t.

But how does this scam work?

Here’s how it is playing out. My initial response played along:

The next email, carefully worded to raise excitement levels, contained a request for four specifically formatted and well crafted documents.

You’ll note that a deadline is created, of ‘November’ and, crucially, an offer of professional help. This is obviously the scam.

I thought I’d call their bluff, at least for a while, by producing the documents as requested. Using AI myself, I delivered the following:

I was very happy with my work, using Claude, particularly as I wove in some surreal ridiculousness to my authentic story set in the Cold War. This included a sequence in which the lead test pilot needs to fight a unicorn with lasers for horns in Narnia. And the appearance of the ghost of John Quincy Adams.

Imagine my disappointment when ‘Ben’ replied explaining that after ‘careful review’, my proposals didn’t meet industry standards and that he ‘strongly advises’ me to use the professional services they offer.

No mention of the unicorns or ghosts, despite the ‘careful review’.

More time pressure was added, telling me that another author already has a contract. No time to waste!

I think it’s time to give in, and find out how much they intend to extract from me.

How much do you think they’ll ask for ?

Stay tuned…

Subscribe now

CAN YOU SPOT THE SCAM?

It’s always the underbelly of society that leverages new technology first.

I worked in broadcasting and video production for a while and I can recall being told to follow the porn industry from a tech point of view, because they always lead. (VHS, DVD, Blu-Ray, Streaming etc etc).

Now, in our writing corner of the world, AI has been seized on by the nefarious who wish to relieve authors of their hard earned cash.

Each approach will be personalised to you. It will be full of praise and at least some detail that seems like it could only come from someone familiar with your book.

Of course it’s flattering, and of course authors fall for it.

Here’s three of the most common.

The initial contact will come as a DM on a social media platform, typically Instagram or X or maybe an email.

1. The Book Club Scam

This seems to be the single most common scam at the moment. You’ll get an email or DM waxing lyrical about your book and announcing that it’s been selected for a book club in Somewhere, North Carolina (etc). After a bit of back and forth, you’ll discover there is a fee for your book to move forward to the actual club. Usually in the $100-$200 range.

At this point, most authors will back out, but I am sure plenty do not. Flattered by the attention and the deep insight the requester has into your book, they’ll press ahead.

So for the scammer, using mailing lists, automations and a decent AI subscription, it becomes a numbers game. Send out 1,000/day. A 5% success rate nets $10,000 / day.

2. Befriended by a Famous Author Scam

For me, this has been harder to spot. I have hosted nearly 500 podcast episodes and I’ve met pretty much everyone in the indie community as well as a bunch of famous authors from trad. That means it’s not out of the question to receive a friendly email from a well known name.

Most recently it was Melanie Harlow. We haven’t met in person but I know of her excellent following and how well regarded she is. I didn’t think it was out of the question that she would message me. Maybe she wants to come to The Self Publishing Show Live?

It wasn’t her, of course. I checked with a friend who does know her. Melanie asked me to report the post, which by the way, was impossible on X as only the person being impersonated can report an impersonator…

I’m not sure how this would have worked. No doubt ‘she’ would have offered some form of promotion to her audience. I didn’t follow it up, I simply reported the post as a scam and told them I knew they were lying.

They blocked me. Which got me thinking. It might be time to take a different tack…

3. The Film Rights Scam

This scam, like the others, works on flattery. Your book is SENSATIONAL (which it is btw) and we want the rights!

Most recently this happened via another DM on X, purporting to be from Archewell Productions. A real production company you might recognise as it belongs to Harry and Meghan…

It was at this point, I became more curious to find out exactly how these scams work…

Fighting AI with AI

I can imagine writers who end up scammed feel embarrassed. It might be why we don’t hear about the details of what exactly happens when you engaged. If this is you DO NOT BE EMBARRASSED. There but for the grace of god go all of us.

Because of that lack of common insight into how these scams work, I’ve decided to start following a few up. I’ll post the results here.

Today, I’m starting with ‘Simon Dan’. Let’s see if he promises a Zoom call with H & M.

To whet your appetite - here’s how the scam started.

I’m now into an email conversation with real life Hollywood producer Ben Browning, or more likely, some bloke in in his underpants in a Florida bedsit. Or possibly a Karachi warehouse…

The first hint at how it works comes from a request for a complex sequence of documents about my book, which they kindly offer to help me compile, should I wish to use their services.

For now, Claude and I are working on the papers ourselves and just like the scoundrel at the other end, I am AI-ing the hell out of it, and having some fun... Here’s a peak at what me and Claude are going to weave into my biography:

My next post will layout all the interactions and we’ll see how far I get refusing their professional help before they lose patience.

James.

Subscribe now