The wrong book-design tool often reveals itself late: you have arranged sixty pages, then discover that the downloadable PDF is a review copy, the captions cannot be updated consistently, or a new paragraph breaks every subsequent page.
Before building the whole book, make a four-page pilot in the tool you are considering. Test the hardest things your book must do, export the result, and make one realistic revision. This is a practical selection method for an illustrated memoir, photo collection or small nonfiction book. It does not require buying several applications or trusting a software ranking.
Write the delivery requirements before choosing features
Ask the intended printer for the page size, interior delivery format, bleed and safe-area requirements, colour instructions and cover process. Record what is confirmed and what is still undecided. A template designed for a different printer may be a useful visual reference, but it is not automatically the correct production template.
Next, describe the work the software must handle. A mostly continuous manuscript needs reliable text flow and styles. A photo essay needs controllable crops and image placement. A bilingual book needs the actual writing systems and fonts tested. Start with the software you already know if it can pass those requirements.
Blurb’s tool-selection article considers format, available time and existing experience. The pilot below adds a decisive question: can this particular project be exported and revised without a fragile workaround?
Prepare a deliberately awkward sample
Use material from your real project where permission allows. Otherwise, create clearly labelled test content. Gather one long heading, several paragraphs, your longest likely caption, one horizontal photograph, one vertical photograph, and a small table or other difficult element. Include accented names or non-Latin text if your finished book will contain them.
Keep a source sheet containing the exact text and image filenames. This lets you check whether an export lost content rather than relying on memory. Use this demonstration plan for an imagined illustrated memoir; it is a test document, not a claim that a four-page bound book is available.
| Pilot page | Content to place | What it tests |
|---|---|---|
| 1 | Long chapter title and 250 words of prose | Heading behaviour, paragraph styles and ordinary reading size |
| 2 | Horizontal image with a 70-word caption | Crop control, caption capacity and image quality |
| 3 | Vertical image beside 200 words | Alignment, wrapping or separate-frame behaviour |
| 4 | Six-row timeline, page number and final sentence | Table legibility, numbering and whether the end of the content survives |
The word counts are stress-test inputs. They are not targets for your published pages. If the format cannot accommodate them legibly, that is useful information: change the page plan, not the text size merely to force a pass.
Build the pilot using repeatable controls
- Set the document to the confirmed page dimensions. Keep the cover out of the interior pilot.
- Create named styles for body text, headings and captions where supported. If the tool lacks reusable styles, record how many places must be changed manually.
- Place images from the original files and check each crop. Do not use a screenshot of the picture as a substitute.
- Apply the actual typefaces and language. Include the longest name or unusual character now.
- Add page numbers in the way you expect to manage the whole book.
- Save, close and reopen the editable project. Confirm that images, fonts and text are still available.
Record where the process stalls. “I could not find the crop control” may be a learning problem. “The crop cannot be positioned independently inside the frame” may be a capability problem. Check the vendor’s current help before treating them as the same thing.
Export the actual deliverable early
Use the export route you intend to use for production, with the printer’s settings. Open that PDF outside the design application. Check the page dimensions, all four pages, the last sentence, caption text, special characters and image crops. Check embedded fonts with the PDF font guide. A successful save or upload alone is not a pass.
One concrete trap: BookWright’s PDF Proof is described by Blurb as a low-resolution, watermarked review file unsuitable for production printing. Do not choose it as the file to send to another printer. Verify the exact export option you need before committing to a provider-specific workflow. See Blurb’s PDF Proof documentation.
Print the pilot at actual size. A readable screen preview zoomed to fill a large monitor does not establish comfortable book typography. If your printer scales the sheet, measure a known dimension and correct the print setting before judging the design.
Run a revision test
Now add a new paragraph to page 1, replace the page 2 image with a different aspect ratio, and lengthen one timeline entry. These changes represent ordinary editing, not an artificial torture test.
Check which pages need attention. Does text continue, disappear into an overflow frame, or overlap the next element? Does the replacement image crop out its subject? Does the timeline stay attached to the correct heading? Change the caption style once and inspect every caption. If you must repair each page separately, estimate that work across the whole book.
| Requirement | Pass evidence | Action if it fails |
|---|---|---|
| Usable production PDF | Correct dimensions, no proof watermark, complete content | Resolve export route before building more pages |
| Readable type | Paper sample works at intended size | Revise layout or format; do not rely on screen zoom |
| Reliable revision | Added text remains accounted for | Use a different text-flow method or shortlist another tool |
| Recoverable project | Saved file reopens with its required assets | Fix storage and dependency handling |
Choose using the failures that matter
Imagine Tool A produces a correct PDF but requires manual caption changes on every page. Tool B handles styles well but you have not found a suitable export. Neither has won yet. Export is a hard requirement; convenience cannot compensate for its absence. Once Tool B’s export is verified, compare the actual revision work you observed.
Do not invent a numerical score to hide an unresolved requirement. Write a decision such as: “Use the current tool for this 24-page image-led book; export passed, and there are only six captions. Reconsider for the planned 180-page text-heavy edition.” That decision is bounded by the tested project.
Keep the pilot, exported PDF, settings and failure notes together. If software updates or the printer changes, rerun the affected tests. You can then submit the pilot for a file check before spending time on the full layout. State which dimensions are confirmed and which decisions remain open.