Turning a web serial into a book

You have finished, or nearly finished, and people keep asking where they can buy it. So you start reading about publishing, and within an hour you have hit the question every serial writer hits: what happens to the 200 chapters sitting on Royal Road?

The answer everybody gives is the same. Stub it, leave ten percent up, you are fine. It is repeated in forum threads, in author newsletters, and on the sites that exist to advise self-publishers.

I went and read the agreement. The ten percent is not in it.

This is not legal advice, and I am not a lawyer. What follows is what the document says, quoted, with a version date. The agreement governs, it changes, and yours is the one that counts. Read it yourself before you act on any of this.

The archive question

One piece of terminology first, because it does all the work here and it usually gets skipped.

This is not a publishing problem. It is a KDP Select problem.

Publishing on Amazon does not require exclusivity. The clause everyone is anxious about lives in the terms for KDP Select, an optional program you opt into, and what it buys you is Kindle Unlimited. You can publish on Amazon tomorrow, leave all 200 chapters up, and never enrol.

If that sounds like the article could end here, it cannot, and the reason is that Kindle Unlimited is where the readers are.

The mechanism authors point to is not the per-page rate, it is the subscription. A Kindle Unlimited reader has already paid this month. Borrowing your book costs them nothing further, so the question in their head is not "is this unknown serial author worth ten dollars" but "do I feel like reading this". In progression fantasy and LitRPG, where readers get through work at genuinely high volume, removing that decision is most of the game. The consistent report from authors and from the people who advise them is that KU substantially outearns wide distribution in these genres. I have not found a rigorous source for the size of the gap and would not trust one that claimed precision, but the direction is not in dispute.

Royal Road treats it as the default too, and institutionally rather than anecdotally. Stub is one of five official fiction statuses on the site, and the knowledge base defines it as content removed "because of exclusivity contracts with, for example, Amazon's Kindle Unlimited". The platform built a setting for authors leaving to KU, and named KU as the example.

So "just do not enrol" is a real option that costs real money, which is a different thing from a way out. That is exactly why the clause deserves reading properly rather than being taken on trust.

And the archive is not a liability you are shedding. It is the readership: years of comments, a following, a review score, and the algorithmic position that produced all three. Deleting it is not housekeeping. It is switching off the thing that worked.

So the question is not "how do I get rid of it." It is "what am I required to do, what am I choosing to do, and what does each cost."

What the agreement actually says

Here is the entire exclusivity clause of the KDP Select terms, in full, from the version last updated 27 September 2024:

1 Exclusivity. When you include a Digital Book in KDP Select, you give us the exclusive right to sell and distribute your Digital Book in digital format while your book is in KDP Select. During this period of exclusivity, you cannot sell or distribute, or give anyone else the right to sell or distribute, your Digital Book (or a book that is substantially similar), in digital format in any territory where you have rights.

That is all of it. One paragraph.

Three things worth reading slowly.

There is no percentage. Not ten, not any. There is no sample exception, no promotional excerpt provision, no carve-out of any kind anywhere in the KDP Select terms. If you are operating on the ten percent rule, you are operating on something that is not in the contract you signed.

The verbs are "sell or distribute." Not "publish", not "make available". Whether a free web serial constitutes distribution is a question I am not qualified to answer and neither is a forum thread. But note that the clause is not quite the flat "nothing anywhere" that Amazon's own marketing page implies, either. It is narrower and stranger than both the folklore and the summary.

"(or a book that is substantially similar)." This is the sentence serial writers should read twice, and it is the one nobody quotes. It means the obligation does not attach only to the exact file you upload. A version you cut, restructured and edited is still plausibly substantially similar to what is on Royal Road. The common instinct, "I revised it heavily, so it is a different work now," is not obviously a defence, and you should not assume it is one.

Where the ten percent rule came from

I cannot prove where a piece of folklore started, so treat this as a reconstruction rather than a fact.

The general KDP terms, section 5.2.1, grant Amazon the right to market your book by "making chapters or portions of your Books available to prospective customers without charge." That is Amazon's sampling right, the Look Inside feature, and it runs to roughly ten percent of a book. Separately, Amazon's help documentation uses a ten percent figure for bonus content, meaning how much supplementary material should sit inside your book.

Neither of those is a permission granted to you to publish ten percent of your book elsewhere. But you can see how a right that Amazon holds, sized at about ten percent, becomes a right you hold, sized at about ten percent, after enough retellings.

Now the honest part, because it cuts the other way.

Plenty of authors stub to ten percent and nothing happens to them. That is true, it is widely observed, and I am not going to pretend otherwise. Amazon does not appear to hunt for stubbed serials.

But non-enforcement is not permission, and the asymmetry here is brutal. The upside of being right is that a few chapters stay up as a funnel. The downside of being wrong is not a warning or a takedown. Section 5.4.8 of the general terms sets out that if Amazon determines you breached the agreement, you forfeit unpaid royalties, and it is account-level. People do lose KDP accounts.

That is a bet you can absolutely choose to take. Take it knowingly.

Two traps in the timing

These are in the agreement, they are specific, and almost nothing written for serial authors mentions them.

Unpublishing does not end your exclusivity. From section 5 of the KDP Select terms:

If you unpublish your Digital Book, we will remove it from the Kindle Subscription Programs, but you must continue to comply with these commitments, including exclusivity, through the remainder of the Digital Book's then-current 90-day period of participation in KDP Select.

So the instinct that gets people into trouble, pull the book and put the chapters back up, does not work. Unpublishing removes it from Kindle Unlimited and leaves your obligation running to the end of the 90 days. The archive stays down until that clock finishes, not until you press unpublish.

It renews itself. From section 3:

Your Digital Book's participation will automatically renew for additional 90-day periods, unless you opt out through the KDP website before renewal.

Ninety days is not a term you serve and walk away from. It is a subscription you have to actively cancel, and if you were planning to restore your archive at the end of it, you have to remember to opt out first. Miss the window and you have committed to another ninety days without deciding to.

Between them, those two clauses mean the gap between "I want my archive back" and "my archive is back" can be six months if you time it badly.

What stubbing costs on Royal Road

Every discussion of this is about what Amazon requires. Almost none of it is about what the removal does to you on the platform you are removing from, and Royal Road documents a cost that belongs in the decision.

Set a fiction to Stub and, in the knowledge base's own words:

This tag excludes the fiction from certain ranking lists.

That is not a footnote. Royal Road is a discovery engine and the lists are the discovery. A stubbed fiction keeps its followers and the readers it already has, and stops competing for the ones it does not. So the cost of enrolling is not only that a new arrival cannot finish the story. It is that fewer new arrivals meet it at all.

Marking it is not optional, either. Stub is a required status rather than a courtesy, and the staff enforce it, precisely because of the effect on those lists.

Two things soften this.

Nothing is destroyed. You delete the chapters by hand, but they remain in Royal Road's database and can be restored later. If you leave Select at the end of a term, the archive goes back up. Worth knowing before you spend a weekend making backups in a panic.

Stub is a recognised state, not a black mark. Readers in these genres know exactly what it means. It reads as "this author is publishing", which to a lot of them is a quality signal rather than a loss.

Note what is missing, though. There is no percentage in Royal Road's documentation either. The ten percent is not their rule any more than it is Amazon's. Both halves of this decision are being made on a number that neither party wrote down.

A serial is not a book

Set the archive aside. Even with the rights question settled, the manuscript is not ready, and this is the part writers underestimate because the words are all already written.

Every chapter ends on a hook. That is correct for a serial, where the hook has to survive a week of waiting. Read forty of them in one sitting and it is exhausting, and worse, it stops working. A hook that fires every 2,500 words is not tension, it is a tic.

You have been recapping. A returning reader needed reminding who Ellis was. A reader who met Ellis nine minutes ago does not, and every recap now reads as the book not trusting them.

Arc breaks are not act breaks. A serial arc ends where the story reached a satisfying pause and you needed one. A book's acts turn where the protagonist's situation fundamentally changes. Sometimes those coincide. Often the arc break is in the wrong place by a few chapters, and the book sags exactly where the serial was strongest.

There are cadence artifacts everywhere. Author notes. "Sorry this one is late." A chapter that exists because you had a schedule and a deadline and nothing else. They are invisible to you because you wrote them under the pressure that made them necessary.

The cutting pass

The single most useful thing you can do is also the least technical: read the whole thing, start to finish, as a reader would.

Not as 200 documents. Not with the editing surface open. In one continuous surface, in as few sittings as you can manage, taking notes but not fixing anything as you go.

Almost nobody does this before publishing, and it is the only way to catch the things that only exist at book scale: the pacing sag in the middle third, the character who vanishes for thirty chapters, the reveal that lands flat because you seeded it twice, the running joke that was funny weekly and is wearying in one go.

What comes out, roughly in order of how much it improves the book:

1. Recaps. Nearly all of them. This alone can cut several thousand words and it makes the book feel confident.
2. Chapter-end hooks, thinned. Not all. Keep them where a genuine act break lands and cut them to a normal scene ending everywhere else.
3. Filler chapters. You know which ones. The week you were ill, the week you had nothing and posted anyway. In a serial they kept the schedule alive, which was a real job. In a book they are dead weight.
4. Author notes and everything around the words that was addressed to a reader waiting rather than a reader reading.
5. Chapter breaks, merged. Serial chapters are sized for a posting cadence. Book chapters are sized for the story. Two 2,000-word serial chapters are often one good 4,000-word book chapter.

While you cut, the continuity you tracked while drafting becomes load-bearing in a new way. Merging two chapters can put an event before the scene that set it up. Cutting a recap can remove the only surviving mention of something the ending depends on. There is more on the tracking method itself in keeping continuity straight past chapter 100.

What you are actually choosing

Strip out the folklore and the decision is small and clear.

Go wide, meaning publish on Amazon without KDP Select and on other stores too, and there is no exclusivity clause. Your archive stays up. Your existing readers keep reading and keep recommending. You give up Kindle Unlimited, which for a lot of serial writers is the majority of the revenue.

Enrol in KDP Select, and you get Kindle Unlimited page reads, and you take on an exclusivity obligation that runs in renewing ninety-day blocks, attaches to substantially similar work, and does not end when you unpublish.

Most advice presents the second as the default, and for these genres that is not a conspiracy, it is arithmetic. Kindle Unlimited is where the readers already are and have already paid. What is worth pushing back on is not the recommendation but the confidence around its details, which is how a number nobody wrote down became a rule everybody follows.

Neither option is the right answer. But the choice is yours, and it deserves to be made against the actual clause and the actual costs rather than against folklore.