The publish button is the least consequential click in a book launch. The decisions that govern the next two years happen earlier, in the five minutes spent filling in the KDP form, and several of them are awkward to reverse once the listing has accumulated history. Knowing what to do before publishing a book means treating that form as a set of strategic choices, not a data entry task. These choices lock into place once a listing accumulates reviews and sales history, even though the form itself takes only minutes to fill out.
Classification, comps, description, subtitle, price, and edition plan all feed the systems that retailers and distributors use to locate and shelve a title. BISAC subject headings function as “search terms,” “access points,” and “shelving guides” across the trade, according to the Book Industry Study Group’s official BISAC FAQ. That means a misclassified book can stay invisible to its intended readers no matter the marketing budget behind it.
This page is the map, using two real titles, Agatha Christie’s The Murder at the Vicarage and J.C. Ryle’s Expository Thoughts on Matthew, to sort these decisions into two piles: which ones are cheap to change later, and which ones are not, with links to depth on each one.
Classification works this way because it feeds directly into how distributors and retailers surface a title to searching readers, the same logic behind IngramSpark’s metadata checklist, which tells publishers to lock in subject codes before any marketing spend begins. Get the subject codes and categories right, and every later marketing effort has something solid to build on. Get them wrong, and even a strong launch campaign is pushing against a closed door. The sections that follow sort the six pre-launch decisions by how reversible they are, walk through the Vicarage and Ryle worked examples, and show what a finished pre-publication document looks like.
Key Evidence
- BISAC codes function as shelving guides: the Book Industry Study Group’s official BISAC FAQ describes subject headings as “search terms,” “access points,” and “shelving guides” across the trade.
- Misclassification suppresses visibility: Firebrand Technologies reports that “even small misalignments can limit a title’s visibility,” and a book in an overly broad category “may never reach its intended audience, regardless of how strong the content or marketing effort.”
- Keyword revisions move the needle: a 2019 Ontario Creates study found discoverability improvements across every BISAC category studied after authors improved their keyword lists.
- Category volume rewards precision: Penguin Random House’s author-facing guidance notes that authors choose from more than 3,000 available BISAC codes, which is why vague categories do more harm than good.
The Six Decisions You Make in Five Minutes
What to do before publishing a book comes down to six fields on Amazon’s Kindle Direct Publishing (KDP) form, each a strategic decision disguised as data entry. They are classification and categories, comparable titles, the description, the subtitle, the price, and the edition plan covering paperback, hardback, ebook, audio, and any later revisions. Getting familiar with how ISBNs work across editions matters here too, since edition choices ripple into identifier decisions.
Classification is infrastructure. BISAC headings tell retailers “the primary and secondary store sections within which the title will best fit,” according to the BISG FAQ, and authors choose from more than 3,000 available codes, per Penguin Random House’s author-facing guidance. That volume of choice is exactly why vague, safe-sounding categories do more harm than good.
Here’s the honest limit: Amazon publishes very little about how its browse categories and seven keyword boxes actually interact. The only reliable evidence available to an indie author is their own before-and-after observation of category rank and search placement. Every one of these six fields is a decision the reader never sees and the algorithm never forgets.
Which Decisions Are Expensive to Reverse
The sorting principle is simple: the cost of reversal rises with how much history the listing has accumulated and how many downstream systems hold a copy of the data. Some fields you can rewrite on a Tuesday afternoon. Others are locked in the moment a reader leaves the first review.
Cheap to change: the description, the keyword boxes, the price, the A+ content, and the categories themselves through KDP support or the browse-category fields. These can be revised repeatedly, and revising them is the correct way to run tests rather than a sign something went wrong the first time.
Expensive or effectively permanent: the title and subtitle, locked into the ASIN and cited in reviews and links; the ISBN and imprint; the trim size and interior layout once reviews attach to a print edition; and the decision to enroll in KDP Select for a fixed term. Firebrand Technologies treats categories and keywords as fields publishers expect to revise as sales data comes in, while title, subtitle, and ISBN function as fixed points once assigned. A second edition resets a listing’s review history in most cases, which is exactly why the edition plan belongs in the pre-launch document rather than the recovery plan.
Awkward in between sits the comp title list. It is not a form field, but as the worked examples below for The Murder at the Vicarage and Expository Thoughts on Matthew show, it shapes the subtitle, cover, description, and category choices an author then has to live with. A description is an experiment. A subtitle is a commitment.
Firebrand Technologies describes this exact pattern: an author picks a wide, safe-sounding category early, watches the book stall at a low sales rank, and only later discovers a narrower subcategory would have put it on a bestseller list at a fraction of the sales volume. By then, changing category is easy. Changing the title already printed on early paperback stock is not.
A Fast Reversibility Test
Before committing to any field, ask three questions.
- Does it live in the ASIN? Title, subtitle, and format changes touch the record itself.
- Does anyone else hold a copy? Distributor feeds, library catalogues, and review citations propagate slowly.
- Does it accumulate evidence? Reviews, rank history, and sales velocity attach to the listing and rarely transfer.
You might notice this while browsing a category yourself: a book tucked into a bestseller-heavy subcategory looks lost within a page of scrolling, no matter how strong the cover is. There’s no single right way to sequence the six decisions, but skipping the step of writing them down tends to show up later as a rushed launch.
Two Worked Examples: Fiction and Non-Fiction
Generic advice hides how differently the six decisions play out across genres, so it helps to work through real titles rather than hypotheticals.
Take The Murder at the Vicarage as a fiction example. Classification pulls between classic crime, cosy mystery, and traditional detective subcategories. Comps are complicated by the book’s age and by the many competing editions of the same public-domain text already on the market. Penguin Random House’s author guidance stresses that genre categories should match reader expectations at the point of search, which is why cosy mystery and traditional detective send different signals than a generic crime label. The subtitle and edition framing end up carrying most of the differentiation work, since the title alone cannot distinguish one edition from dozens of others.
Take J.C. Ryle’s Expository Thoughts on Matthew as a non-fiction example. Classification runs across biblical commentary, devotional, and historical theology headings. Readers search by author name, Bible book, and purpose rather than by mood, which changes how comps function: they become other commentary sets rather than tonal matches. The edition plan (single volume, multi-volume, or study format) shapes the price band directly. Fiction competes on resemblance. Non-fiction competes on specificity of need. The same six fields carry different weight in each.
What the Fiction Example Shows
Working through the Vicarage example surfaces one clear pattern.
- Crowded shelves punish vagueness: a broad crime category buries a title among current bestsellers.
- Comps set the cover brief: the chosen comparables decided the visual and tonal direction before any design work began.
- Public-domain competition is real: edition framing becomes the main lever available to a title.
Building the Pre-Publication Document
Producing a single document holding all six decisions and the reasoning behind each is the practical answer to what to do before publishing a book. This turns launch day into execution of a plan rather than a series of rushed choices made under deadline pressure.
That document should specify 2 to 3 specific subject codes, following the guidance in IngramSpark’s metadata checklist, alongside a 7 to 10 phrase keyword list covering genre, era, and theme, as recommended in IngramSpark’s guidance on book metadata. It should also list 3 to 5 comps published within roughly the last five years, a description of 200 to 600 words with a strong opening line, a title and subtitle kept under 80 characters for mobile display, and a price band with an edition plan attached.
Common errors worth naming: defaulting to broad categories for perceived safety, inconsistent author name spellings across editions, and skipping comp research entirely. For the parts Amazon leaves undocumented, run a testing protocol instead: change one field at a time, leave it in place long enough to see rank movement, and record the before state. Amazon confirms none of this, so conclusions stay provisional no matter how confident the guidance sounds, including this article’s.
Done properly, this is several days of category browsing, comp reading, and description drafting per title. Readers who want to see what a finished version looks like can review the two full sample reports for The Murder at the Vicarage and Expository Thoughts on Matthew at perivane.com. And if the broader launch sequence around this document still feels uncertain, the patterns in common self-publishing launch pitfalls are worth a look before finalizing anything.
Why This Matters
Classification and metadata decide whether a book is reachable at all, and they get made once, under time pressure, at the end of a long project, the same rushed conditions the 2019 Ontario Creates study linked to inconsistent keyword quality across the BISAC categories it sampled. Authors who spend months on a manuscript routinely spend minutes on the fields that determine who finds it. Getting the reversible decisions wrong costs an afternoon. Getting the irreversible ones wrong costs the book.
Conclusion
What to do before publishing a book is make six decisions deliberately and in writing, then treat the reversible ones as ongoing experiments and the irreversible ones as final. Descriptions, keywords, prices, and categories can be revised as many times as needed. Titles, subtitles, ISBNs, and edition structure carry forward for the life of the listing.
Amazon documents very little of how its own systems weigh these fields, so distrust confident claims about categories and keyword boxes, including any made here, and verify with controlled changes of your own. Work through the six decisions in order, follow the deeper guides linked along the way, including notes on choosing editorial support before that manuscript ever reaches the metadata stage, and keep the document. Future editions inherit the reasoning that way, instead of repeating the guesswork.