Back to the table

116 Advanced Sm Storycosm Series & Universe

Storycosm

A whole fictional universe holding many stories.

What it is

A storycosm is one invented world that holds several series, standalone novels and casts, any of which may turn out to be connected. The word is this site's own. It names a story world built to expand: the setting, the people and the backstory sitting under a single novel or under a whole series. Publishing has its own word for the same thing, and that word is shared universe. Agents and editors know it, so use it when you talk to them.

Why would a working novelist want one? The headline reasons are simple. A reader who finishes one of your series has a next book waiting. The research you did for one book pays for itself twice. Each new title lifts the older ones back into view. The headline cost is just as simple. Every book you publish narrows every book you have not written yet, and those narrowings multiply rather than add up.

Almost nobody designs one in advance. The usual path is two books, then a moment when you notice they are kin, then a universe recognized after the fact. So here is the thing to carry away: a universe is a bet that you will still want this world ten years from now, and the stake is the freedom of every book you have not written.

"Storycosm" is not industry vocabulary, and a query letter or a pitch needs a standard term the person across the desk already knows. Six terms circle this ground, and they mean different things. Keep them apart.

Nothing on this page is proprietary. All of it is ordinary working-novelist practice, worked out by crime, romance and science fiction series novelists long before this table existed. The practices are theirs. If you want the long view, Mark J.P. Wolf's Building Imaginary Worlds, listed below, studies an invented universe as a subject in its own right rather than as the backdrop to one story, and Karen S. Wiesner's Writing the Fiction Series, also listed below, gives the one practical account in print of how a universe forms, including growth from a standalone.

A series is a set of books read in sequence and held together by a person, a place or an ongoing story. A universe is the larger container. It can hold several series with different leads, plus standalones, novellas, and books joined by nothing but shared ground. Here is the test. If the same protagonist or the same ongoing plot runs through every book, you have a series. If a reader can finish one whole sequence and then start a second one with a completely fresh cast in the same world, you have a universe. A crossover is an event inside a universe: two casts meet, a character makes a cameo, a spin-off launches, a reader of both sequences gets a payoff. The Crossover page owns that event and the extra consistency work it brings. This page is about the universe as a built object: what binds it, how far apart its parts should sit, what it earns, and what it must not do.

When the casts change from book to book, something else has to hold the books together. Six things usually do the job. Each does different work, and the looser the binder, the more freedom you keep and the weaker the link a reader feels. Here they are from loosest to tightest.

A recurring minor presence is the loosest. One figure turns up in every book: a fixer, a bookseller, a letter carrier, a traveler nobody can place. It costs almost nothing. The only constraint is that the figure has to stay alive and reachable. What you get is the faintest flicker of recognition, and the readers who catch it prize it far beyond its size. Brandon Sanderson's Hoid, wandering through the Cosmere books, is the best-known modern fantasy case. Balzac was doing it in the 1830s. The mechanics of a recurring character belong to Relationship Web.

Moral climate is a constant stance toward the material: a world where everyone is forgiven at the end, or a world where nothing is ever paid for. Also among the loosest, and the hardest to make visible. Readers sense it without being able to name it. It binds tone and little else, and on its own it will not move a reader from one series to the next.

Place is a town, a city, a county or a planet. Its cost is geography and distance: everything has to be plausibly reachable, and a book set somewhere else is out. Its yield is that readers notice it faster than any other binder, and it outlasts a change of era. Setting and Maps cover the ground itself.

Period is Regency London, the 1950s Midwest, the fifty years after a war. Its cost is technology, manners and the range of events you can use. Its yield is instant identity on the shelf, while your cast and your geography stay free.

Institution is a police squad, an auction house, a guild, a diocese, a hospital. It costs more than it looks. The institution dictates what kind of plot each book has, and it goes on dictating after you have tired of it. In return it yields renewable cases and a ready excuse for strangers to meet. Factions covers how to build one with interests of its own.

Rule-set is the tightest: a magic system, an alien biology, a physics. It has to hold in every book, permanently. It yields the thing that makes a set of books feel like a universe rather than related books on a shelf. Adding to it late always costs, and a rule stated in book two and contradicted in book six cannot be quietly mended. Magic System owns the design. This page owns the size of the commitment.

A typical working universe runs on two or three binders at once. The useful move is to write down which ones are actually operating rather than adding more. A common discovery: the binder you assumed was doing the work, say the magic system, is inert, and what readers actually recognize is the place and the tone.

The pitch for a universe says you build it once and then write in it for years. Half of that is true at most. Correct advance design is impossible. To do it you would need to know, before book one, which secondary character can carry her own series, which corner of the world readers will ask about, and which rule you will want to break in 2031. Writers who try anyway produce giant bibles and zero books, and that is the worldbuilding trap at universe scale.

Here is the sequence that actually happens. You write a novel. You write a second one set nearby, since the research is done and the world is still fresh. Around book three or four, you notice the books are kin, and from then on you write as if a universe exists. Recognizing it afterward is the standard entrance, not a planning failure. Most famous universes arrived this way.

The usual cost of this path is that the early books carry decisions made without knowledge: an impossible date, a character who should have known something, a rule stated too firmly in a passing line. Three remedies exist, and all of them get used. You can absorb the contradiction, letting it become a fact of the world and working around it; this costs the least and is usually right. You can rank the texts, recording which book wins when two conflict, and stop relitigating; the Continuity Bible page covers ranking, including the formal version used in large multi-author worlds. Or you can revise the early book; this costs the most and mainly suits authors who control their backlist, but it works, and Stephen King's case below shows it. Whatever you do, never wait for certainty. An unpublished universe is a hobby.

The hardest practical question is how far apart the parts should stand, and at bottom that question is about access, not craft. Reading-order anxiety is the single biggest barrier between a reader and a big universe. A reader who cannot tell where to start often buys nothing rather than risk starting wrong, and the share of readers in that position grows with the size of the universe. Success lowers tryability. The Series page owns the version of this problem inside one sequence: orienting someone who arrives at book four, recapping without spoiling. This page handles the problem between series, where the remedies differ.

A front door is any title a newcomer with no prior reading can open, finish and feel satisfied by, with no other book required. One front door means you have a series that looks like a universe. Several front doors mean a universe. Count them; the audit in Try it yourself below walks through it. If the honest count is one, you have a decision to make, not a marketing problem.

Number and name each sub-series separately, so the hierarchy shows. "The Kirkmead Novels, book nine" is wrong. "The Nessa Ruddock Mysteries, book three, in the Kirkmead universe" is right. One running count across four casts tells a reader only that there are many books, and it deters them. A reader must see at once which of two numbers is theirs. The Series page covers what a number means inside one sequence.

Publish the in-universe chronology separately from the reading order, and label both. Running them together is the commonest error. A reading order is a recommendation: begin with X, go on to Y. A chronology is a fact about the world: this happened in 1847, that in 1902. Readers want both and use them differently, and many of the arguments in an author's inbox come from someone reading one as the other. Big franchises often get around to this years late and under pressure.

Write your entry points deliberately, and know which books are not entry points. A book deep in a mature universe that rewards veterans and baffles newcomers is fine. An unmarked one is not. If book seven needs books one to six, its description says so, and somewhere prominent and current there must be an answer to "where do I start?" that points away from book seven.

The where-to-start page is one authoritative copy, kept current, on your own site and under your own control. It gives the order in words, never in cover images alone. It takes an afternoon and costs nothing to maintain, and of everything on this page it is the item most often skipped.

Five neighbors share this ground, and here is where each one takes over. Crossover handles meetings of casts and series: the cameo, the spin-off, the reward for the loyal reader, and the consistency bill that comes with them. Series handles the form itself: types of series, how long a run should be, order inside a sequence, the business case for repeat books about the same people, companion series, and what a finite, finished series yields. Series Arc handles the overarching story, the question asked in book one and answered in the last, and the planting that takes across many volumes. Continuity Bible handles the record: what to note, when to note it, and how to merge two records when worlds join. Worldbuilding handles inventing one world properly and the building-not-writing trap, and its layers belong to History, Culture, Factions, Lore and Maps in turn.

Why it matters

A reader who finishes your series has somewhere to go. Be exact about what a universe buys, since the vague version has sold many authors a universe they did not need. Writers worry about the reader who quits mid-series. The reader who costs you most is the satisfied one who finishes and leaves for another author's books. At that moment a universe hands them a title that is fresh and familiar at once: new cast, same world, same promise. No other mechanism in publishing does this as well.

Your research keeps paying. The period detail, the procedures, the geography, the way an institution works: all of it is already funded. Writing your second title inside a world you know beats writing it in a fresh one for speed and for quality, and by book five the gap is huge. At any rate of output, that gap is real working time.

Cross-sell runs sideways as well as backward. Inside a single series the traffic flows one way, backward, from the latest volume to book one. In a universe the pull also runs sideways, between sequences whose plots never touch and that are joined only by the world. A reader of your historical novel tries your contemporary mysteries and comes back. A lone series cannot do this. The Series page covers how discovery compounds inside one sequence. You will find no figures here on rates, terms or percentages, and that is deliberate: every one of them shifts, and any number printed here would be stale within months. For the current ones, read the retailer help pages and the writers publishing in your category this year. One more condition. The sideways pull works only when the link is visible from outside. Readers detect it through the cover, the description and the metadata, meaning the background data that tells a store what a book is, never through the manuscript. An invisible universe earns nothing. Blurb and Self-Publishing cover how to make it visible.

The cost compounds. Every published book constrains every unwritten one. A single series has that problem too, and the Continuity Bible exists for it. What belongs to a universe alone is that the constraints multiply instead of adding. Here is the arithmetic. With one series of six, a new fact gets checked against six earlier books. With three series of six, a new fact in series C gets checked against the five other C books plus anything in A or B that shares a place, a period, an institution or a rule. Each sub-series you add brings a relationship with every existing series, not one more item on a list. The continuity load outgrows the book count, and authors near book ten or twelve report that the writing feels unchanged while everything around it has grown heavier. The conclusion is to insulate the parts on purpose unless a connection pays for itself. Two series in one city that never touch cost almost nothing. Two series with interlocking plots mean rechecking both whenever either one moves. Link on purpose. Enthusiasm is not a reason. The record itself belongs to the Continuity Bible.

A universe-wide threat shrinks every story that comes after it. Series fiction has a known escalation problem: readers get used to whatever the sequence has survived, and a merely bigger threat produces no fear. The Series page covers that within one series. The universe version is nastier and deserves its own name, and the author usually inflicts it on herself through a book she experienced as a high point. The trap is the universe-wide threat, something that becomes true everywhere: the gods depart, the empire falls, the plague is cured, the war ends. Afterward every untold story shrinks. The prose is no worse. The reader has simply seen the world at its maximum. A murder novel in a provincial cathedral city works until you publish the book in which that city is nearly destroyed, and then the murder is a footnote, and nothing repairs that. The damage runs backward too. Once a universe-wide disaster exists, stories set before it pick up an irony you did not want, and stories set nearby pick up the question of where these people were and why they stayed out. The big franchises carry this permanently, and their fix is explanations nobody enjoys. Two habits protect you. Spend universe-level events very rarely and only when you know the aftermath; before you write one, draft a paragraph about a book set a year later and check that you want to write it. Try it yourself calls that the shrink test. And escalate sideways rather than upward. Tension rises with proximity and cost, and hardly at all with scale. A threat getting nearer to someone the reader loves beats a threat getting bigger, and the universe keeps its size. Stakes and Tension explain why.

Most writers should not build one. Said bluntly, since the topic attracts exactly the authors it damages. The bet rides on your appetite ten years out, not on your readers, and the odds are poor. Many authors who adored a world at book three loathed it by book eight, through no fault of their own; the Series page covers that at length. Do not build a universe on your first or second novel, when you do not yet know your strengths and a decade bound to your first invented world is a costly lesson. Do not build one as a way to avoid writing; the appetite for universe design is often the appetite for not finishing a draft. Do not build one because a successful writer in your genre has one; theirs is usually ten years of slow accretion dressed up as a plan. Do consider one when you have three or more finished books circling the same ground, when readers keep asking about a secondary figure or a corner of the setting, or when you are already mid-series and the next book you want to write is not about your lead. The low-cost bet is a standalone set in your series' world. It costs almost nothing, and it tells you whether the world works without its lead, whether readers follow, and whether you still enjoy it. The outcome is a universe with one extra book in it, and the bet is reversible. Universe-first is not.

Universes outlast books, so settle the ownership early. Sometimes they outlast the writers able to work in them. Four situations need settling. Co-writing first: once two authors share a universe, the record stops being private and becomes joint work, and the question to settle before anything else is who owns the world apart from the books. An unwritten answer is a bad answer. A multi-author world needs a named owner of the setting, a stated approval rule, and a written plan for what happens when one party stops. The most thoroughly documented model in fiction is a formal one, Wild Cards, described below: its credited author is the Wild Cards Trust, a legal entity that holds the property, and its gatekeepers are editors named in advance. The Continuity Bible page covers what sharing does to the document. Licensing outsiders in: tempting, and an underrated trap. Every outsider's book is one more consistency obligation, and approving it is real labor that adds nothing to your own prose. If you do it, approve against a written document rather than memory, and keep a veto. Pen names: a universe can span two, and sometimes should, to keep a grim sub-series apart from a cozy one. The cost of the split is the cross-selling that justified the universe in the first place, since one name has no easy way to point at the other. Decide which matters more. Estates last: ten years of a universe justify drafting a will for it. An estate can license continuations, authorize a successor writer, or do nothing, and "do nothing" is the default when nothing is written down. The working model is a handover arranged in advance and carried out over several books. Take the business side to a local lawyer. The one craft note is that your heir inherits your filing system and will curse it; the record-keeping for a handover, with one recent, well-documented case, sits on the Continuity Bible page.

Documents are not books. The building trap at universe scale is worse than the single-world version in one way: it looks more respectable. Worldbuilding owns the single-world version. Universe building produces documents: a map, a chronology, twelve sub-series with working titles, a diagram relating three magic traditions. Documents feel like progress. None of them is a book. None of them can be sold, read or reviewed. Universe design is procrastination that cannot be finished, because every question breeds another: river trade leads to who controls the bridge, which leads to a treaty, and no fiction gets written. Three rules hold the line. The universe is a consequence, never a prerequisite; enough world for the current book is enough, and the rest gets decided in whichever book needs it. Invent at the point of use and record at once; when a scene needs a fact, make it and note it, and the universe ends up exactly as detailed as the fiction demands. Count books, not pages of notes; eleven planned sub-series beside two published novels is a document, not a universe, and the document is costing you books.

Real-world examples

Stephen King's Mid-World (1975 to 2004): a universe noticed, then exploited

King's linked universe is the clearest published case of a universe recognized after the fact, and he indexes it on his own site. The index itself tells you something: here is an author keeping a list of which of his books connect.

Follow one thread. 'Salem's Lot (1975) is a vampire novel. The priest, Father Callahan, fails and leaves town in disgrace. The story is closed. No sequel is implied and none was planned. The Dark Tower novels open with The Gunslinger (1982), unrelated at the time: a western-fantasy about a gunslinger crossing a ruined world. Then Wolves of the Calla (2003), twenty-eight years after 'Salem's Lot, has Callahan turn up in that other world, alive, with the missing decades explained, and he plays a major role in the final three volumes.

King's official Dark Tower connections page repeats the pattern. Randall Flagg comes from The Stand (1978) and The Eyes of the Dragon. Ted Brautigan comes from Hearts in Atlantis. Patrick Danville comes from Insomnia. Derry and Bridgton recur as places, not plots. None of it was designed in advance. All of it was noticed and then exploited, and that is the normal route, run across thirty years (1975 to 2004) by an author with an unusually large backlist to mine.

Then the bill. In 2003, with the final volumes still to come, King reissued The Gunslinger revised and expanded, longer by several thousand words. The foreword gives the reason: the original held what King called "errors and false starts", especially set against the later books. Here is the purest instance of the price of a retroactive universe. Book one was rewritten to fit a universe that did not exist when it was written. Only an author with that degree of control over his backlist can do it, and most authors lack it, and that is the argument for absorbing or ranking instead.

Set Brandon Sanderson's Cosmere beside it. The Cosmere is the most famous recent universe designed as a universe from the outset. The Mistborn books, Elantris, Warbreaker and the Stormlight Archive share one cosmology, mapped out ahead of most of the writing, and one minor figure recurs across them. The planning is real and so are the payoffs. What the planning did not solve is the entry-point problem, and there it is severe. Several rival recommended reading orders circulate, and "where do I start?" has become a genre of its own. Planning is no cure for access. Sanderson appears elsewhere on this table for unrelated reasons: his magic design on Magic System and his drafting practice on First Draft and Revision.

Tana French's Dublin Murder Squad (2007 to 2016): six books, six narrators, one squad

French built a crime universe the opposite way from a detective series, and her answer to the entry-point question is clear.

Six novels share one institution, the Dublin Murder Squad, and almost nothing else: In the Woods (2007), The Likeness (2008), Faithful Place (2010), Broken Harbor (2012), The Secret Place (2014) and The Trespasser (2016). The narrator changes with every book, and each new narrator was seen earlier in a small part. Cassie Maddox is the partner of the first book's narrator and narrates the second. Frank Mackey heads Undercover in the second and narrates the third. The pattern holds all the way through, and nobody narrates twice.

The transferable point is the binder: one institution, the place behind it, and nothing more. No overarching plot, no series question, no universe-level clock. Almost any starting point works. Each novel is a complete case and a complete emotional arc, and a newcomer forfeits only the pleasure of recognizing faces. Rotating the cast also dodges two chronic ailments of the long detective series: the lead who ages, and the status quo that freezes. A squad is renewable; a detective is not. Nobody has to be held in one job for twenty years.

The chaining is the clever bit. Promote a secondary figure to the next lead and every book becomes a full story plus a trailer for the next, with no cliffhanger and no promise owed. It makes almost no demand on the author, and it ports to any genre that contains an institution. French's later move shows the other standard expansion: a second, separate Irish setting with a retired detective at its center, in The Searcher (2020) and The Hunter (2024). Instead of adding a sequence inside the existing world, she opened a fresh corner. The Flaw page studies In the Woods separately, for a flaw the book will not cure; the two pages argue different things about French.

Becky Chambers's Wayfarers (2014 to 2021): four novels, four casts, four endings

The Wayfarers novels are the purest published case of a universe whose parts stand as far apart as they can.

Four novels share a setting, the Galactic Commons, and its rules: which species exist, how travel works, what travel costs. The Long Way to a Small, Angry Planet (2014), A Closed and Common Orbit (2016), Record of a Spaceborn Few (2018) and The Galaxy, and the Ground Within (2021) overlap beyond that setting as little as Chambers could manage. The second book's lead is a minor figure from the first. The third and fourth have casts untouched by the earlier books, and the fourth is a set of strangers stuck at a waystation on a planet the first book scarcely names. Every novel closes fully. No series arc runs through them, and the set still won the Hugo Award for Best Series in 2019, a series award given to four books with no continuing story. Chambers has said the set is complete at four.

The craft lesson is the trade-off. A binder of place plus rule-set and nothing tighter makes almost every volume a front door, and that is the best commercial position available. The price is that nothing accumulates. A reader who finishes has no unfinished business, only appetite. You can be easy to try or hard to stop; both at full strength is impossible. The Supporting Cast page uses the first book's crew for a different argument, about private agendas inside one book. That page is about the crew. This one is about the shape of the sequence.

Louise Erdrich's North Dakota (1984 onward): a universe with no arc and no numbering

Literary fiction has shared universes too, and Erdrich's is among the largest and most studied.

Love Medicine (1984), her debut, follows several families on an Ojibwe reservation in North Dakota through linked narratives that cross decades forward and back. The Beet Queen (1986) widens the scope to Argus, a nearby town. Tracks (1988) goes back to 1912 to 1924 and the earlier generations of families readers already know. More than a dozen of her novels now inhabit this territory, among them The Night Watchman (2020), which won the Pulitzer Prize for Fiction in 2021, and The Mighty Red (2024).

Two lessons. First, two valid orders exist, the order of publication and the chronology of the families, and readers use both. Nothing needs solving. The situation is inherent wherever a later book may be set earlier, and the obligation is to publish both orders and label them clearly, the two-documents point from above. Second, the connective tissue can be very thin. No arc, no series question, no numbering. The binders are one place, one group of families and one consistent moral climate, and a readership has held since 1984 on that alone. Keep that in mind against the myth of the master plan.

Lisa Kleypas's London (2004 onward): the link printed on the cover

Romance builds shared worlds more systematically than any other popular genre, and Kleypas is a clean case of accretion followed by sale as a universe.

The Wallflowers novels (2004 to 2006) follow four friends in Victorian London. The Ravenels, from Cold-Hearted Rake (2015) to Devil in Disguise (2021), take place in the same society one generation on. The Wallflowers' children are now of marrying age and have become the leads, and Sebastian, Lord St. Vincent, a Wallflowers hero, is present as a father. The Hathaways share the world as well.

The evidence is on the covers. Three Ravenels titles carry the subtitle "The Ravenels meet The Wallflowers": Devil in Spring (2017), Devil's Daughter (2019) and Devil in Disguise (2021). Retailers list the three as their own group under that phrase, and the publisher sells the link on the front cover, pitching the Ravenels to Wallflowers readers and the Wallflowers to Ravenels readers.

The lesson is the one most authors skip: the cross-sell was made visible on the product. Connected books earn nothing commercially unless the connection can be detected from outside, so the link has to be readable on the product and not merely true inside the text. Whether a late meeting between two casts is good craft is Crossover's question.

Wild Cards (1987 onward): the most formal shared world still running

Wild Cards is the most formal multi-author shared world that has survived, and anyone thinking of admitting another writer to a universe should study it first.

The first volume appeared in 1987. The roots were a long-running tabletop game played by a group of Albuquerque writers, and George R. R. Martin and Melinda Snodgrass designed the premise as a frame many people could write inside: after World War II an alien virus is released above New York; it kills most of the people it infects, deforms some, and gives a few superhuman powers. More than forty authors have contributed, across more than thirty volumes, from four different publishers, and the books were still appearing when this page was checked, so look up the present state rather than trusting a count on a web page.

Three features matter, and the first can be copied at small scale. Ownership sits with an entity, not a person. The books are credited to the Wild Cards Trust, and that is unusual and deliberate: the setting becomes a defined piece of property with rules for its use, rather than an understanding among friends that collapses at the first money or the first quarrel. Second, the gatekeepers are named. Martin and Snodgrass edit, so consistency is decided by people named beforehand rather than settled mid-argument. Third, newcomers write to a specification, not a vibe. The Encyclopedia of Science Fiction treats this as central to the form: a document fixes the roles, the places and the limits, and everyone writes inside it. That document is the Continuity Bible's subject, and in a shared world it stops being optional.

The one takeaway: before a second writer enters your world, settle in writing who owns it, who approves, and what happens when one party leaves. Those three written answers separate a working shared world from litigation.

Comics and screen universes: evidence to take, solutions to leave

The best-documented shared universes are in comics and film, and that makes them tempting evidence and a misleading model. Both halves matter.

First the evidence. The modern superhero shared universe begins with Fantastic Four #1 in 1961, together with the practice of characters appearing across titles, which turned a publishing line into one world. Film followed from 2008. These are the largest continuity experiments ever run, and two findings transfer to novels.

The first finding concerns entry points. Comics and film both eventually issued official in-universe chronologies, separate from publication or release order, because publication order had stopped working as a way in. When the rights owner itself posts a timeline on its streaming service, the numbering has been admitted a failure. A novelist with four sub-series should do this early and deliberately, not at book twenty under pressure.

The second finding, continuity load, is stronger. DC Comics published Crisis on Infinite Earths (1985 to 1986), conceived by writer Marv Wolfman specifically to tackle a continuity that decades of stories had made impenetrable to newcomers. DC's own 2018 account, with Wolfman and the artist George Pérez, states the goal: a single unified universe in place of a multiverse Wolfman judged reader-unfriendly. It partly succeeded, created new problems, and was followed by more clean-ups later. Continuity load never plateaus. It accumulates until someone pays to cut it.

Now the limits of the analogy, and they weigh more than the evidence. Comics and film universes rest on four assets no novelist has. Staff: continuity editors, and in the biggest operations whole departments. Reboots: the power to void old stories and restart the numbering, fine for a corporation and ruinous for a novelist whose readers bought those books from her personally. Mass retcons, a retcon being an alteration of established facts, which the audience accepts as a constant low-level drip. And a reader culture that treats inconsistency as a puzzle rather than an error; Marvel has long run a joke prize, worth nothing, for readers who explain away a mistake. Novel readers who find a protagonist's sister renamed between books see carelessness, not a puzzle. So take the two findings and leave the solutions. A novelist's reboot is an apology. A novelist's continuity department is one document and the discipline to open it.

An everyday scenario: three crime novels and the Kirkmead decision

Imagine a writer with three published crime novels. Her lead is Nessa Ruddock, a solicitor in the cathedral city of Kirkmead who keeps getting drawn into other people's catastrophes. Sales are modest and steady. Two things have just happened. Readers have started writing in about Clement Pyle, the diocesan surveyor who has a brief part in all three books and who exists only because book one, chapter four, needed someone to pass on one fact. And she has begun a historical novel set in the same close in 1911, sparked by a story she found while researching book two.

Intended or not, a universe is beginning, and the right questions, in order, are unglamorous.

"Will I still want Kirkmead in ten years?" This tests her appetite, not her readers'. If the honest answer is "I am already a bit tired of the close", the historical novel goes somewhere else and the discussion is over.

"What is the binder?" Place plus institution. The institution is the Haverly Trust, which in the books owns most of the land in the close and the river frontage, so that everyone in Kirkmead eventually deals with everyone else in Kirkmead. The pair is strong: the place is vivid to readers, and the Trust gives strangers in any era a reason to meet. She writes both down, because otherwise she will drift into believing Nessa is the binder.

"How many front doors does this have?" One at present, the first Ruddock novel. The historical novel would be a second, and that is the strongest reason to write it. A Pyle novella would be a third, and a weak one, because readers would arrive expecting Nessa.

"What does the 1911 book make impossible?" The shrink test, run before writing. If the book ends with the Haverly Trust's land confiscated, the engine has been removed from every contemporary book she has not written yet. If it ends with the Trust's power confirmed and a family grudge planted, fuel has been added. Same book, same quality, opposite effect.

"What do the two sequences have to agree about?" A short list only: the date the Trust was founded, the ownership map, the geography of the close, one or two family names. The brevity is deliberate. Anything the two sequences do not share never needs checking.

Here is the whole decision. No master plan, no twelve-book chart, no diagram. One honest answer about appetite, a named binder, a count of front doors, a check on consequences and a short list of shared facts. Then the 1911 novel gets written.

Try it yourself

This takes about an hour and works whether you have two books or twelve. Have your published books at hand.

  1. Write the ten-year question. Finish this sentence: "In ten years I would still want to be writing in ______ because ______." If you hedge, stop here. A universe you are already bored with is the costliest mistake on this page, and a standalone in the same world is a far less expensive test.
  2. Name the binders. List the six candidates: recurring presence, moral climate, place, period, institution, rule-set. Tick the ones that actually hold your books together, not the ones you meant to use. Then ask which ones a reader would tick. If the two lists differ, the reader's list wins, and your descriptions and covers follow it.
  3. Audit the front doors. List every published book in the world. For each, write Y or N: can someone who has read none of the others start here and finish satisfied, with nothing left unexplained? Count the Ys. One Y means a series, whatever the label says. Two or three mean a universe a stranger can enter. Then take each N and ask whether it is marked as a non-entry book, in its description and on your site. An unmarked N is where the one-star reviews that say "confusing" come from.
  4. Make two documents and keep them apart. Page A is headed "Where to start": sentences, a recommended order, and a plain statement of whether order matters. Page B is headed "Chronology": dates, with the world's events listed by date. Host both pages yourself. Never merge them, and never let cover images alone do the job.
  5. Draw the connection map, then cut. One sheet, one box per sub-series, one line between any pair that depends on the other for a checkable fact. Interrogate each line: what does it give you? A line that gives nothing is pure continuity cost. Cut it in the next book by not mentioning the link. Deliberate separation is real continuity work, and the only kind that reduces the load instead of recording it.
  6. Run the shrink test on any planned universe-level event. Before writing an event that changes the whole world, draft one paragraph about an ordinary book set the following year, in the same place, at your normal scale. Read it back. If it reads as small, trivial or impossible, drop the event or shrink it to local scale; the drama almost never suffers. Then the backward version: one sentence on why the people in your other sub-series stayed out of it. If you cannot write the sentence, you have found the problem early instead of in the reviews.
  7. Write the shared-facts list. Shorter than a bible, and doing a different job (the Continuity Bible page covers bibles). One page, holding only the facts more than one sub-series relies on: dates, ownership, geography, world rules, a handful of names. Keep it under a page and read it before anything goes out. The brevity is what guarantees you will actually consult it.
  8. Settle the collaboration question now. Three lines: who owns the world apart from the books, who certifies that a manuscript is consistent, and what happens to the world if you stop. Writing "me" on all three lines is fine. Writing it down keeps the question visible until a collaborator, a licensee or an estate forces it.
  9. Count. Tally A is the novels you have published in this world. Tally B is the pages of notes you have on it. If the notes far outnumber the novels, the finding is about this year's substitute for writing, and not about the universe at all. Then put the answers on one sheet:

Binders (mine / what a reader would say): __________ / __________ Front doors: __ of __ books. Unmarked non-entry books: __________ Where-to-start page: __________ Chronology page: __________ Shared facts more than one sub-series depends on: __________ Shrink test on next big event (a book set the year after): __________ Who owns the world / who approves / what if I stop: __________

Common mistakes and misunderstandings

Believing universes are planned, not noticed. Almost no working universe was designed. Almost all were spotted at about book three or four and only then developed on purpose. Wait for certainty and you get a long wait and a pile of documents.

Labeling a series a universe. If the same protagonist runs through every book, you have a series, however vast the world. The test is whether the set lets a reader complete sequence A and then begin sequence B with fresh people. If it does not, you have a series with good worldbuilding, which is excellent and is a different thing.

Leaving the connection invisible. Huge effort goes into the real linkage and none into making it legible: the cover, the description, the series metadata, your own site. A universe nobody can see cannot cross-sell, and most of its commercial case goes with that.

Planning for readers who start at book one. Readers never arrive in order. Front doors and chronologies exist for exactly that reason. Assuming a book-one start is planning for a reader too rare to matter. The Series page covers the reader who arrives mid-sequence.

Reading a universe-wide threat as growth. Blow up the city, end the war, kill the gods, and you have raised the ceiling once and lowered every story that follows. Writers usually do it at the close of a successful sub-series, with confidence at its peak, and the problem surfaces two books later.

Thinking one more sub-series is free. It adds a relationship with every existing sub-series, and that is far more than one extra set of facts. Continuity work grows faster than the book count. Keeping sub-series apart on purpose is a technique, not laziness.

Using comics and film as a model instead of evidence. They have continuity staff, reboots and an audience that treats inconsistency as a game. You have none of those. Keep their findings, about entry points and about a continuity load that only ever rises, and drop their solutions, since every one of them needs a department or a reader culture you do not have.

Building the universe to avoid the draft. The uncomfortable one. Universe building cannot be finished, feels productive, and produces artifacts you can show people. If you find yourself fascinated by the world's banking system halfway through a draft, the finding is about the draft.

Saying "storycosm" in a pitch. The word is fine for your own planning. In a pitch or a query, say "shared universe". The point is not minor: vocabulary nobody recognizes spends part of your first impression on explaining the word instead of the book.

Related elements