Strategy
Part of Ebook marketing: what to keep and what to drop for 2027
Ebook marketing tools to test, starting with the file
Ebook marketing tools to test, starting with the file: Kindle Create, EPUBCheck and Ace by DAISY on the export, then delivery, email and ads trials.
What to take away
- Judge a tool by the file it exports, not by its editing view.
- Run the KPF preview, the EPUB reports and the retailer routes as separate tests.
- Decide on purpose whether the PDF or EPUB itself may appear in search.
- Hash every release file and record the tool version and settings.
- Test reader delivery, email and paid discovery tools after the file passes.
Preview the export, not the editing view
The export is what a reader receives, so test it first. The tools behind ebook marketing cover writing, layout, EPUB creation, validation, covers, metadata and distribution.
Amazon's guide to preparing books with Kindle Create describes importing a source, applying structure and styles, building a table of contents, exporting a KPF file and checking it in the online previewer. It is authoritative for that workflow, not an independent comparison, and it does not promise accessibility on every reading system.
| Stage | Tool | Passes when | Evidence to keep |
|---|---|---|---|
| Production | Amazon Kindle Create | Import, styling, table of contents, KPF export and online preview all finish without errors | The KPF file and a preview screenshot |
| EPUB validation | EPUBCheck | Reported issues are corrected and the check is rerun | The saved report |
| Accessibility | Ace by DAISY | The report is reviewed, issues are corrected and the check is rerun | The saved report |
| Reader delivery | BookFunnel or StoryOrigin | A test link is sent to a reader, the download completes and the access instructions work | The test link and the reader's confirmation |
| Mailchimp | A test signup produces the welcome message, an unsubscribe path and a subscriber-record export | The exported subscriber record | |
| Paid discovery | Amazon Ads | Campaign setup, targeting and reporting are reviewed in a draft without launching a campaign | A screenshot of the draft |
| Retail routes | KDP, Apple Books, Kobo Writing Life, Barnes & Noble Press, Draft2Digital, IngramSpark | Current submission rules and accepted files are checked for each route you intend to use | Required fields and manual handoffs |
Trial tasks by tool job
- Authoringheadings, lists, notes, images, links
- Conversionexport reflowable EPUB, no lost semantics
- Validationfind structural and accessibility failures
- Metadatamaintain edition and channel source of truth
- Distributiondeliver, correct, remove, export records
- Reportingreconcile orders, refunds, royalties, costs
Use a difficult representative chapter in every trial, not a clean sample supplied by the vendor. Include tables, figures, notes, lists, links and language changes, plus the assistive-technology checks the book needs for accessibility.
Validate the EPUB in a fixed order
Run the final EPUB through EPUBCheck and Ace by DAISY, read both reports, fix what they flag and rerun. EPUBCheck is the conformance checker for EPUB publications, and Ace is the accessibility check beside it.
- Export the EPUB and the PDF from the same source.
- Run EPUBCheck on the EPUB and save the report.
- Run Ace by DAISY on the same file and save that report.
- Correct the issues in the source, export again and rerun both tools.
- Open the result in the reading systems you intend to support.
Check the EPUB and PDF outputs separately. Navigation, layout and accessibility can differ between them even when both come from one source, so a clean EPUB report says nothing about the PDF.
A PDF carries logical structure only if structural tagging was added when it was created. Preview the EPUB on more than one device too: flowable text breaks lines and paginates differently depending on the reading platform and user choices.
Decide whether the file itself gets indexed
A PDF or EPUB that shows up in search can bypass the context that explains edition, price, update, rights and accessibility. Decide deliberately whether the file should be indexed, rather than finding out from a search result after launch.
Google lists the file types its search system can index, and PDF and EPUB are among them. That list describes technical support, not a promise that a file will be crawled, indexed, ranked, or shown.
Discovery checks before indexing
- Verify indexability and canonical signals
- Confirm file access and titles
- Check descriptions and structured data
- Add campaign parameters and retailer routes
- Decide whether the file itself is indexed
Publish one controlled book page and a representative sample, then check each of these:
- Identifiers:set an ISBN policy where ISBNs are used and record retailer IDs consistently.
- Metadata:compare titles, descriptions and contributor details across the page and retailer listings; use ONIX where the route requires it.
- Files:inspect the EPUB and PDF separately for navigation, layout and accessibility.
- Discovery:verify page titles, descriptions, canonical signals, file access and campaign parameters.
- Store routes:confirm the retailer pages resolve and the edition, sample and purchase path match the landing page.
Freeze the release set
Keep a hash of every release file, and preserve the tool version and settings used to create it. A reviewer can then confirm that the file in the store is the file you tested, and rebuild it if a retailer rejects an update.
The release set holds the source manuscript, the final publication files, the cover, complete metadata, the rights record, retailer IDs and transaction reports. These must be retrievable without relying on one tool. Move the catalog once, then compare the result with the original.
Fix file problems before promotion starts, since a corrected file is cheaper than a corrected campaign. Other mistakes to fix before promotion sit outside the file, and a wider launch checklist puts this test beside them.
Record what the test does not cover
A pass for production is not a pass for the whole marketing system. Record pass, fail or untested for each task, and attach the file, report or screenshot behind the result.
For example, mark Kindle Create a pass for KPF production if your trial completes source import, styling, table-of-contents creation, export and preview without errors. Record EPUB validation, accessibility, retailer acceptance and marketing delivery as untested unless you test them separately.
A feature earns credit only when an intended user completes the task safely. Save the tested plan, the date, the evidence and the tool version, and verify current product documentation before adoption, because capabilities can change. A file that passes can then go into listing tests.
Common questions
Should PDF and EPUB be generated together?
They can share a source, but each output needs its own layout, navigation, accessibility and device testing.
Does a passing KPF preview mean the EPUB is valid?
No. The preview covers the Kindle Create workflow only. EPUB validation and accessibility stay untested until EPUBCheck and Ace by DAISY have run.
What is the essential export?
The source manuscript, final publication files, cover, complete metadata, rights record, retailer IDs and transaction reports.
