There’s a durable stereotype that metadata is clerical.
A small grey office behind the publishing house, where somebody types commas into a form while the real work happens elsewhere, preferably in a room with a mood board and an expensive biscuit.
Charming, in the way your town map would be charming if it omitted all the bridges.
A hybrid book can already have readers walking toward it from two directions while the official map has printed no district name, and no bookseller has hung a sign above a bay of shelving.
Its text has not changed.
Its market has.
So the operational question is whether the map gives both groups a usable route before the local architecture catches up.
Which means separating systems that get bundled into one bland parcel marked metadata.
ONIX carries the record.
Subject codes provide the formal routes.
Retailer categories and search turn some of those values into discovery.
Then a shop makes its own shelving decision, sometimes with the brisk sovereignty of a town councillor who has just discovered a very good endcap.
Metadata routes the book. Classification records where institutions have agreed to point the sign.
Neither system determines literary merit, commercial seriousness, or whether a dragon deserves a kiss.
They determine access.
Here is the map audit.
Do both roads reach the town?
Start with two romantasies.
Each has a central love story and a fantasy apparatus substantial enough to require more than decorative fog.
The first is coded only to Fantasy. Its copy calls the book an epic adventure. Its retailer category uses one of the three slots available. Its keywords are mostly variations on the title.
A romance reader searching for that emotional contract gets no road sign at all.
The second leads Romance over Fantasy, with Fantasy over Romance as a companion.
Its copy puts the relationship promise in the first sentence, without pretending the invented world is a printed backdrop.
Its keyword field supplies relevant adjacent entry terms the cover cannot say for itself.
The novel has not become more romantic, less fantastic, or especially virtuous in a cardigan.
It has become usable from two approaches.
That is the worked difference. Buried romance, versus a dual-code route.
The first asks your reader to know in advance that a book hidden behind an adventure sign will deliver the relationship they wanted.
The second lets the system make that possibility legible.
And the retailer cap is what turns this from a filing preference into a decision. Three store categories. No more.
The subject-code guidance says the same thing. No more than three codes, with the first naming the category that best describes actual content.
That is not permission to name every genre you encountered before lunch.
It is a mapmaker’s constraint.
Your lead code says which district the book inhabits.
Your companions identify the adjacent borders.
So for a hybrid, run the title under both plausible leads before treating either as obvious.
Under Romance over Fantasy, which promise gets the first lamp-post?
Under Fantasy over Romance, which one does?
The test is reader recognition.
Not rank.
And the distinction gets more consequential the moment the trade starts announcing weather.
A romantasy title is folded into its parent totals rather than reported on a discrete line. A dual-coded book counts with romance or with fantasy according to its lead choice.
Which means a report can describe reader demand and the municipal decision about which ledger receives the units, in the same sentence, without distinguishing them.
Your category claim needs its year, its parent line, and the route used by the hybrids crossing between them.
The codebook and the town square
The subject-code list is granular in a way no sensible shop intends to build at street level.
Dozens of major sections. Hundreds of fiction codes alone, with the largest genre subtrees running to sixty or more headings apiece.
A map that detailed is useful precisely because nobody puts all of it on the ground.
The bookshop is a different jurisdiction entirely.
A store maps inventory to a few dozen internal categories tied to actual shelf locations. When the codes are too broad to isolate what its customers ask for, staff curate by hand.
So your title can be Romance over Fantasy in the record, enemies-to-lovers with swords in a retailer search, and sitting in a locally invented display category under a sign designed by somebody who knows what people ask for on Saturdays.
That is not a standards failure.
It is a division of municipal labour.
A library needs an exhaustive scheme that can accommodate every object it has ever kept.
A store needs sales-driven landmarks.
The error is treating a shop’s shelf as an appellate court that overruled your metadata, or treating your metadata as an order compelling a retailer’s bays to comply.
Routing and shelf choice are separate decisions.
The town square can still alter the codebook, though.
The subject codes committee is made up of publisher, retailer, wholesaler, and data-aggregator volunteers. It meets monthly. The list is republished annually, in late autumn.
A proposed heading has to represent a durable cluster. Not a single title, author, or publisher with a very determined publicist.
A recent annual update added more than two hundred headings and revised over a hundred, including a cosy heading inside the fantasy subtree.
And romantasy still has no dedicated code.
Publishers use Romance over Fantasy, or Fantasy over Romance, depending on which element marketing judges dominant.
That is committee lag.
Not ontological failure.
A district can be full of pedestrians, cafés, and three shops selling the same improbable boots, well before the council finds the paint for a sign.
Find the second door, not just the main gate
Your lead code governs the formal road.
Keywords open a second door.
So the diagnostic is: what is the second door?
If nobody can name a relevant search route reaching the adjacent promise, your code order is carrying more of the discovery burden than it can bear.
Relevant keywords affect placement in search results, alongside sales history and rank, and you get a small handful of them per title.
Which means a romance-first code does not have to abandon fantasy-led search language, and a fantasy-first code does not have to make the relationship invisible.
The decisive word is relevant.
Test your candidates in live search before publication, and see whether they surface results a shopper would recognise as relevant.
That is a small but proper piece of civic engineering.
A keyword is not a banner you wave over the town.
It is a door a shopper actually tries.
If it opens onto unrelated properties, it hasn’t routed your book.
It has clogged the alley.
And the prohibition list explains why the field is not a panic cupboard.
Terms redundant with your title, contributor, or category are barred. So are subjective claims, time-sensitive language, misrepresentative content, unauthorised brands, and the names of a retailer’s own merchandising programmes.
A prohibited entry is not a cheeky shortcut around zoning.
It is a policy violation subject to removal.
So for the two romantasies, the audit is concrete.
Put each plausible lead code at the main gate.
Then ask what the keyword slots add that the route does not already say.
If Romance over Fantasy leads, the second door should help a fantasy reader recognise the book’s actual imaginative promise.
If Fantasy over Romance leads, it should help a romance reader find the central contract.
Neither route needs keywords repeating the category, praising the book, borrowing another author’s name, or advertising a temporary sale.
A second door is useful because it opens somewhere else.
Which map travels abroad?
An international edition adds a further map, and the wrong instinct is to assume it simply replaces the first one in a nicer typeface.
The multilingual scheme gives you twenty top-level subject areas and a couple of thousand headings, with fiction gathered under its own letter.
Unlike the North American list, it’s post-coordinated.
A cataloguer combines a base subject with qualifiers for place, time period, language, educational level, interest age, or style.
Which lets your record describe a book more specifically, without inventing a new fixed code for every alleyway in the world.
Mappings between the two schemes have been distributed with the licence for over a decade, so a US-assigned code can travel toward a UK or EU retailer without manual recoding.
The two are designed to coexist, even though the multilingual scheme is eventually intended to supersede the national ones.
So your international plan is not choose the superior map.
It is use the map the receiving town reads.
A romance-fantasy hybrid can be legible in more than one code system, while a bookseller in each market still decides which shelf receives its physical body.
Which is why category migration is never a vote held by one committee.
And the record itself keeps moving.
The carriage format reorganised into composite blocks well over a decade ago and has revised since. Its controlled values change on a much faster timetable than its schema, adding declarations for AI-generated content, partly-digital retail products, and content advice.
Your title’s map is not finished because the manuscript went to press.
Is each field carrying the right cargo?
Not every visible word belongs in every field.
The record separates a book’s description from reviews, prize mentions, and cover-line promotional text, so retailers can independently decide what to display.
The field names are doing municipal work.
Your title is an address.
Promotional copy is a poster.
Which means your hybrid audit should not solve a discoverability problem by turning the subtitle into a small parade.
The description can name both promises.
The codes can route them.
Keywords can make a second door searchable.
Collateral can carry the promotional material in its own block.
Each field gives a different part of town a clean instruction.
Force every field to be the loudest billboard, and nobody can find the address.
The ambiguous corner lot
Some books really do belong at the corner.
Their romance is not a subplot wearing a decorative sash.
Their speculative system is not merely a chandelier.
Both promises govern your reader’s experience.
The temptation is to code one, call the other for fans of, and trust your reader to discover the rest by intuition.
That is how a book gains an admirably pure identity and loses a perfectly available street.
Use several of these as the town-plan test:
- The lead code names the promise that most accurately describes the book’s content, and two companions have been chosen rather than accumulated.
- The title has been run under both plausible leads, to see what changes in the first reader-facing promise.
- The three-category retailer limit is treated as a routing constraint, not an invitation to call every adjacent category home.
- Relevant keywords have been tested in live search, with the second door doing work the category route does not already do.
- Retail copy names dominance plainly enough that no reader is sent toward a happy ending expecting an epic quest, or toward an epic quest expecting a romance the book does not provide.
- The shop’s local shelf is read as a sales and curation decision, distinct from your record and your codes.
- Any claim that the category is rising or falling carries its reporting year, and acknowledges where cross-shelved units are counted.
If several apply, your book is not caught between districts.
It has an address at an intersection.
And if the two promises remain genuinely equal after the audit, let the copy name the dominant reading experience, let the lead code carry it, and give the adjacent promise a companion route and an honest keyword door.
The answer is never both, somehow written in faint pencil.
It is two routes with one clear main entrance.
The shelf may still choose differently.
It may stay a fiction table, become a romance feature, or invent a local sign with more flair than any national standard could manage.
That does not invalidate the district.
Codes are committee work.
Shelves are local government.
Readers are the traffic.
A category exists because people walk there. The signs can be late.