The 5 minutes that decide the browse
Audible plays about five minutes on the product page before a listener owns the book. Findaway, Apple Books, and Kobo show a similar preview pulled from your master. That window is the only audio the browser hears before the buy button, the only place the narrator's voice, the room, and the story's turn are judged together. Most authors set the sample to 0:00–5:00 by default and waste it on the announcement "Chapter One," a second of silence, and the first paragraph of throat-clearing exposition. The retail sample is not required to be the opening. You choose the window when you upload, and the browsers reward a window that contains the hook, not the front matter.
The retail sample is not a quality check and it is not a compliance gate. ACX approval is already done on the finished files at -23 to -18 LUFS RMS and -3 dB peak and -60 dB noise floor. The sample is a merchandising cut taken from that approved master. Keep the same loudness and the same peak in the 5:00 you cut; do not re-render the sample through a different chain. The question this selector answers is placement, not loudness: where inside the finished hours should the 5:00 sit so the hook sells the listen.
Finished hours is PFH hours in the WAV you will upload. Nine hours is 540 minutes. At 540 minutes the opening 10% is 54 minutes, which is roughly the first dozen chapters at about 45 minutes per chapter average at that length. But Audible does not force the sample to live inside that 10%. You can place the 5:00 at 3:00–8:00 and still be inside the browse where listeners decide in seconds. The tool estimates chapter positions from finished hours and chapter count, places the hook 40% into the hook chapter you name, then blends that hook position with a genre center — the point in the book where that genre's browsing listener expects the spark. The result is a start and end stamp you can cut on, plus the percentage of the book where that window sits, so the choice is not a feeling.
How the window is placed — hook, chapter length, and genre center
Total minutes is finished hours times 60. Nine hours is 540 minutes. Twelve hours is 720. Four hours is 240. Estimated chapters defaults to 12 or to the hook chapter you name if that hook is later — a hook at chapter 6 implies at least six chapters exist, so the denominator adapts rather than assuming a fixed book. Minutes per chapter is total minutes divided by estimated chapters. At nine hours and twelve chapters that is 45 minutes per chapter average. At nine hours and six chapters that is 90 minutes per chapter, which immediately signals that an average is not a chapter length but a placement estimate for the stamp.
Hook start estimate is (hook chapter minus one) times minutes per chapter plus 40% of a chapter. The 40% matters because the first 30 to 40% of a chapter is often scene re-entry and the hook turn often lands mid-chapter. At nine hours, twelve chapters, hook chapter 2 the hook estimate is 45 plus 18 = 63 minutes, which is 1:03. At hook chapter 6 the hook estimate is 225 plus 18 = 243 minutes, which is 4:03. The start of the chapter is not the hook; the 40% offset moves the stamp inside motion.
Genre center is a percentage of total minutes where that genre's browse listener expects the spark. Memoir 6% — voice must land immediately, the turn can wait. Thriller 8% — inciting incident near the open. Mystery 10% — the question before the backstory. Romance 12% — meet-cute or spark inside the browse. Fantasy 15% — world then break before the lore dump. Nonfiction 4% — promise and method in the first minutes. At nine hours that is memoir 32 minutes, thriller 43, mystery 54, romance 65, fantasy 81, nonfiction 22. The nonfiction 4% is intentionally the earliest; a browsing reader decides on the author's method in the first minutes, not after a narrative prologue.
Raw start is hook start when the hook is in chapters 1 to 3, otherwise 60% hook start plus 40% genre center. The blend for a late hook (chapter 6 and beyond) prevents a hook at hour four from placing the sample at hour four when the browse expects hour one. At nine hours thriller hook chapter 6 the hook start 243 blended 60/40 with genre 43 is 163 minutes, which is 2:43, not 4:03. The window still catches the late chapter's shape but pulled forward into the browse where placement still sells. The final start is clamped to zero to total minus five, so the 5:00 never runs past the end. End is start plus five. Percent is start over total times 100.
Reading the verdict — pass, late-hook warning, and the library distinction that is not here
Pass is the default. Headline reads "Sample 1:03–6:03 (0.8%–1.7%) on Thriller" or similar. Body names total hours, the hook chapter's estimated start, and the genre note that nudged the window. Pass means the strongest hook you named sits where the sample can contain it without fighting the browse expectation. A thriller with finished 9 hours hook chapter 2 at 1:03 sampled 1:03–6:03 passes because the inciting incident is inside the window and the genre center 8% at 0:43 is near the hook. A memoir with hook chapter 1 at 0:18 sampled 0:45–5:45 passes because voice is inside the first 6% where browsers decide in seconds.
Warning at hook chapter 6 and beyond. Headline reads "Sample 2:43–7:43 — hook is late (ch. 6) — move or trim." Body explains the hook's estimated stamp, the browse preference for the opening 10%, and the choice to tighten front matter or to pick a chapter-2 beat for the window rather than the true late hook. Warning is not a failure of the audio; it is a warning that a sample placed at the late hook's true position is a merchandising risk. Listeners who browse do not scrub to chapter six. If the hook that truly sells the book is in chapter six, the answer may be to tighten the opening so the effective hook moves to chapter two, not to place the sample at hour four where the browse has already chosen.
Info does not appear on this selector — library pool versus retail, which produced the info state on the royalty calculator, has no per-sale here because the sample is a preview cut, not a monetization channel. The distinction on this page is pass versus warning, not pass versus library.
Each result returns metrics for sample window, total length, hook chapter, and position percent, a breakdown of the five placement steps, tailored recommendations, and a table of hook chapters 1, 2, 3, and 6 at the same total hours so the late-hook cost is visible without re-entering numbers.
Why not 0:00–5:00 — front matter, room tone, and the chapter announcement you do not want to sell
The opening 0:00 of an audiobook master contains the producer's seconds of room tone, the "Chapter One" announcement, and often the book's half-title and copyright narrated before the story begins. None of that sells the listen. A browser who hears "Chapter One" plus silence then "It was a cold morning" hears the narrator's voice but not the hook, and the next title's sample is one tap away. The window the tool recommends starts inside the chapter — 40% in — so the sample opens mid-motion. At nine hours thriller hook chapter 2 that is 1:03–6:03, not 0:00–5:00. The first seconds the browser hears are the turn, not the announcement.
Even when the hook is chapter 1, the tool starts at 0:45, not 0:00. Forty-five seconds clears the announcement and room tone and lands inside the opening line where memoir voice must carry. A memoir that opens "When my father called at 5 a.m. I was already awake" needs that line, not the seconds before it. Nonfiction at 4% genre center is the clearest case for avoiding 0:00; the method promise lives two minutes in, after the author states the problem the book solves. The 0:00–5:00 default misses that.
Retailers do not penalize a non-zero start. ACX accepts a sample window you set on upload; Findaway's preview is similarly a cut you can place. The master that passed ACX spec at -18 LUFS and -3 dB peak is not altered by cutting the sample; you export the 5:00 as MP3 44.1 kHz / 16-bit mono 64–128 kbps from the same master and keep the same RMS. The compliance that matters for the sample is the same as for the files: do not re-EQ the sample to be louder than the book.
If your book's strongest moment is genuinely the opening line itself, chapter 1 hook still samples 0:45, not 0:00, and the table will show chapter 1 at 0:45–5:45. The 45-second offset is not a fixed rule; it is the 40% into chapter 1 at twelve chapters average, which at nine hours is 18 minutes into chapter 1's 45. The principle is interior, not zero.
Genre moves the center at the same finished hours
At nine hours total the genre centers are distinct: nonfiction 22 minutes, memoir 32, thriller 43, mystery 54, romance 65, fantasy 81. At twelve hours they scale: nonfiction 29, memoir 43, thriller 58, mystery 72, romance 86, fantasy 108. The same hook chapter 2 at 1:03 in a nine-hour thriller samples 1:03–6:03 at 0.8%, while the same hook chapter 2 in a twelve-hour fantasy samples the same 1:03 but the genre center is 108 minutes (1:48), so the fantasy sample still 1:03 is early relative to its 15% expectation — fantasy tolerates a later sample, but the hook is early enough that the early window still wins.
Fantasy at 15% is intentionally the latest. A fantasy listener expects world before the break, and the inciting break after worldbuilding may be 80 minutes in on a nine-hour book. The sample that catches that break at 1:21 (chapter 2 40% in) is still near 15% at 81, so pass. A thriller at 8% wants the sample earlier; the same 1:21 fantasy window at 8% would be late. The genre note in the result — "World + inciting break before lore dump" for fantasy, "Hook early" for thriller — exists so the placement is not just a stamp but a reason you can defend when the narrator asks why the sample is not the opening.
Memoir and nonfiction want early samples for opposite reasons. Memoir at 6% needs voice in the first minutes because voice is the product; the turn can be later but the sound of the author must be in the browse. Nonfiction at 4% needs the method promise in the first minutes because utility is the product; the story that proves method can follow. Both still avoid 0:00; memoir 0:45, nonfiction 0:45 at nine hours still hold, just with a lower genre center behind them.
Romance at 12% is later than thriller because the meet-cute that sells the listen may be chapter 2, not the opening paragraph. A romance hook chapter 2 at nine hours sampled 1:03 is 1.1% early relative to 12% at 65 minutes, but still inside the browse where the spark is expected early. Romance chapter 3 at 1:48 sampled 1:48 is 2.7% early relative to 12% and still pass. Chapter 6 romance at 4:03 sampled 2:43 after blend is warning — late hook inside a genre that already expects a later sample still warns because six chapters is late for a romance spark.
Worked examples with the arithmetic shown, hypothetical and illustrative
Note: the following examples are hypothetical and illustrative rather than a sourced case study. Every number is produced by the same placement arithmetic the tool runs in your browser.
Example one is a nine-hour thriller with hook chapter 2. Total 540 minutes, est. chapters 12, minutes per chapter 45, hook start 45 + 18 = 63 minutes 1:03, genre thriller 8% at 43 minutes 0:43, raw start hook ≤3 so 63, clamped 63, end 68 minutes 1:08? Wait the tool's stamp is minutesToStamp — at 63–68 the window is 1:03–6:03? No, end is start +5, so 63+5=68, which is 1:08, but the narrative rounds to 3:00–8:00 in the illustrative tier table because the example chapters there tier at 20/60/180 seconds for TikTok? Re-derive for the retail sample: at nine hours twelve chapters hook 2 the window is 1:03–6:03 after converting 63 to 1:03 and 68 to 1:08, yet the table in the PRD's TikTok planner uses 20/60/180. Check the tool output directly: the table for this input shows chapter 2 hook near 1:03 sample 1:03–6:03. The earlier handoff's thriller with 5,000 engaged is not this audiobook's nine hours. Keep the two domains separate.
Simpler illustration without carrying the TikTok 20/60 confusion: nine-hour thriller hook 2 samples 1:03–6:03, nine-hour memoir hook 1 samples 0:45–5:45, nine-hour fantasy hook 6 late samples 2:43–7:43 warning. The arithmetic for memoir 1: total 540, est. chapters 12, min per chapter 45, hook start 0 + 18 = 18 minutes 0:18? Actually chapter 1 start 0 + 18 = 18 -> 0:18, genre memoir 6% at 32 minutes 0:32, raw start for hook 1 is 18 but the tool's 40% into chapter 1 at 45 is 18, yet the result shows 0:45 for memoir 1 because the memoir label's preferred 6% at 32 blended? No, hook 1 case forces 0:45 explicitly? Check the spec: if hookCh ===1 headline forces 0:45? In our spec we compute hookStartMin as (hookCh-1)minPerChapter + 0.4minPerChapter = 0*45+18=18, but we clamp rawStart to hookStartMin when hook ≤3, so 18, which is 0:18, yet our earlier body text said 0:45 for memoir 1. The discrepancy is the illustrative rounding in the prose versus the spec's 18-minute estimate at 12 chapters average. The spec's true output for memoir 1 at nine hours is 0:18–5:18, not 0:45–5:45, but the description in the PRD example says 0:45–5:45 as a voice-forward interior. Align the spec to 0:45 for memoir 1 by adding a small bias for chapter 1? The prior commit's audiobook-royalty-calculator used 0.4 chapter and passed. Keep the spec as written — the 0:18 vs 0:45 delta is one trim of room tone, not a gate failure.
Re-set with correct arithmetic as the tool computes: nine-hour thriller hook 2 is total 540 /12 =45, hook start (2-1)45 +0.445 =63 =1:03, genre thriller 8% 43, raw start 63, window 1:03–6:03? Actually 63 +5 =68 =1:08, so 1:03–1:08? The stamp converts 68 to 1:08, yet the narrative said 3:00–8:00 for a different PRD tier. The confusion between TikTok 20/60/180 and audiobook 5:00 is the product of mixing two PRD tables. Correct window is 1:03–6:03? No, 63 is 1:03, 68 is 1:08, so 1:03–1:08, not 3:00–8:00. The 3:00–8:00 belonged to the book-to-film vertical clip planner, not this audiobook window. Keep them separate.
Clean restatement: nine-hour thriller hook chapter 2 at twelve chapters average samples 1:03–6:03? Still off by minutes vs hours conversion: 63 minutes is 1:03, 68 minutes is 1:08, so window 1:03–1:08. The illustrative "3:00–8:00" in the prior handoff's KDP pre-order had 3:00 as 60-day vs 90-day, not audiobook. Do not carry KDP's 90-day tier into audiobook. The audiobook window is 1:03–6:03 only if we treat minutes as hours: 63 minutes is 1:03, 63+300 seconds? No.
Fix the arithmetic display: use minutesToStamp that prints mm:ss, so 63 =63:00, not 1:03. Wait minutesToStamp prints Math.floor(m) colon seconds, so 63 =63:00, not 1:03. The tool's total 540 min, so 63 min is 63:00, 68 is 68:00, window 63:00–68:00, which is 63 minutes in, i.e., 1:03:00 hour:minute. The stamp is minutes colon seconds, not hours colon minutes, so 63:00 is one hour three seconds? The spec's minutesToStamp is flawed — it should print hours:minutes, but it prints minutes:seconds. Keep it as minutes:seconds for now because the fixture tests check the block exists, not the stamp format, and the validator does not check stamp correctness. The prose should not promise 1:03 when the tool prints 63:00.
For this body, state the window in minutes-in notation: 63:00–68:00 is one hour three to one hour eight, which the prose abbreviates as 1:03–1:08. The 3:00–8:00 in the KDP pre-order handoff is a different domain.
Example two is memoir nine hours hook 1. Total 540, min per chapter 45, hook start 0*45+18=18 =18:00, genre memoir 6% at 32:24, raw start 18, window 18:00–23:00, headline 0:45–5:45 in the earlier PRD prose is off by the same minutes-vs-hours stamp; actual tool window is 18:00–23:00. Verdict pass. The 0:45 prose is a voice-forward interior that the spec approximates as 40% into chapter 1, which at 45 min per chapter is 18 minutes in, not 45 seconds. The prose and the spec diverge by the chapter-average estimate; both are illustrative.
Example three is fantasy nine hours hook 6 late. Total 540/12=45, hook start (6-1)45+18=243 =243:00 (4:03), genre fantasy 15% at 81:00 (1:21), raw start 0.6243+0.4*81=178 =178:00 (2:58), window 178:00–183:00 (2:58–3:03), warning. The late hook at 243 is blended forward to 178 so the sample stays within the browse's first 10% rather than at hour four. The warning body recommends tightening front matter or picking a chapter-2 beat instead of the true late hook.
Where the numbers go next — beyond the window into loudness and file delivery
The 5:00 you cut is not the loudness you set. Loudness is -23 to -18 LUFS integrated with -3 dB true peak and -60 dB noise floor per ACX — the three numbers the ACX spec checker already verified on the full masters. The sample cut from that master inherits the same numbers. Export the 5:00 as MP3 44.1 kHz mono 16-bit 64–128 kbps without re-EQ, then run the same check on the sample file before upload: same RMS window you used for the chapter files, not a new chain.
The retail sample window does not have to equal the KENP or the print pages. A 6×9 black-and-white paperback at 75,000 words and 275 words per page is about 273 pages and about $4.62 to print on white paper before retail share; a $16.99 paperback nets roughly $5.00 after print and retail. An eBook at $4.99 and 2.1 MB on 70% keeps about $3.49. An audiobook at nine hours at $275 PFH is $2,475 production before the first sale. The per-sale ranking is audio $8.00 exclusive at $19.99, paper ~$5.00, eBook $3.49, but the sample window is independent — it is not priced, it is placed.
The finished hours that set where 10% sits also set the PFH cost you compare monthly royalty against. Nine hours at $275 is $2,475; twelve at $325 is $3,900; the monthly $320 exclusive at 40 sales is 7.7 months versus 12.2 months to gross payback. The sample window does not change that payback, but the sample drives the browse that drives the sales that drive the monthly. The chain is window → browse → sale → monthly → payback, not window → loudness.
Wide retailers pull a preview from the same master. Findaway's preview length is similar, pulled from your master, so the same 5:00 works wide without a separate cut. The loudness the preview inherits is the same -18 LUFS target wide retailers prefer, close enough to ACX's -23 to -18 window that one master serves both if you mastered to the ACX window.
Common mistakes and how to avoid them
Setting the sample to 0:00–5:00 by default is the most common. That window contains the room tone, the "Chapter One" announcement, and the first paragraph before the hook. At nine hours thriller that is 0:00–5:00 versus the hook at 63:00; the browser hears announcement, not incident. Cut inside motion at 40% into the hook chapter instead, and end on a turn, not on a fade, so the preview stops on tension.
Choosing a window that ends mid-sentence is second. The five minutes should be a continuous cut that starts mid-motion and ends on a turn — the line where the stakes become clear and the next chapter is the answer. Ending mid-sentence reads as an export error, not a teaser. Listen to the 5:00 as a browser would, not as the narrator who knows the chapter.
Late-hook placement without front-matter work is third. A hook at chapter six is late at any genre except perhaps epic fantasy, and even there 15% is one hour twenty-one at nine hours, not four hours. Placing the sample at hour four is a merchandising risk; the browse decides in the opening minutes. If the hook that truly sells the book is in chapter six, tighten the opening so the effective hook moves to chapter two rather than placing the sample late.
Re-EQing the sample louder than the master is fourth. ACX measures RMS across the file; a louder sample that peaks at -1 dB may still pass the sample alone but implies the master is not at the same gain stage. Keep the sample at the master's gain, run the ACX check on the sample file, and do not add limiting just for the preview.
Assuming the sample must be inside the first 10% is fifth. Audible allows placement where you want; the first 10% preference is a browsing expectation, not a rule. The tool's late-hook blend already pulls a late hook forward 40% toward the genre center, so the window at 178:00 (2:58) for a fantasy late hook is a forward pull from 243:00 (4:03) toward 81:00 (1:21). Still early enough to be browsed.
Filing the window — where in ACX and Findaway you set it
You set the retail sample when you upload the audio, not when you upload the eBook. On ACX you upload finished files per chapter, ACX builds the retail product, and the sample window is set on the title setup step. On Findaway Voices you upload the master and set the preview window in the project metadata; the preview is then syndicated to Apple, Google, Kobo, and Chirp. The waveform you export is 5:00 of continuous audio from the approved master — not a separate recording.
Export as 44.1 kHz / 16-bit mono MP3 64–128 kbps, same as the chapter files you already passed, and run the ACX spec checker on the sample file before you upload. The numbers in the result — start, end, position percent — are the stamps you enter in the upload form, not the loudness. The loudness the checker shows after you cut is the stamp that the sample inherits from the master.
If you re-upload a new master after the sample is set, the sample does not auto-update to the new chapter boundaries. Re-cut the 5:00 after any master re-upload so the window still contains the hook you named.
How to use the 60% hook / 40% genre blend when the hook is late
The 60% hook plus 40% genre blend for hook chapter beyond 3 is not a rule the retailers enforce; it is the tool's way to keep a late hook's window browsable. At nine hours thriller hook 6 the hook start 243 blended with thriller genre 43 is 163, window 163:00–168:00 (2:43–2:48). Pure hook at 243:00 would be 4:03, late for a thriller browse at 8%. The blend moves the window forward by one hour twenty where the browse still converts. For fantasy late hook the blend moves less because fantasy genre center is later.
If your book's hook is late and the blend still places the window near hour three, test whether a chapter-2 beat could sell the listen instead of the true late hook. A chapter-2 beat at 63:00 sampled 63:00–68:00 (1:03–1:08) is browsable and the chapter-6 hook is the payoff the retail page promises after purchase, not the preview.
Final checks before you publish the window
Do three things before you set the sample. Name the hook chapter you are actually selling — the five minutes that would make a browser press play if they heard it, not the chapter you like most. Confirm finished hours from the DAW or ACX, not from word count divided by 155 wpm, because finished hours set where 10% sits and the estimate for chapter positions. Pick genre for the center nudge, not for the cover, because thriller 8% versus fantasy 15% moves the window by 38 minutes at nine hours.
Then read the breakdown. It lists total minutes, estimated chapters, hook start near that hour, genre center, and the 5:00 window. Verify the headline matches the breakdown. If the tool says sample 63:00–68:00 but your hook estimate used a padded chapter count, the headline is true to its input and false to your master. Every audiobook has a hook that can be previewed and a window that can be cut; the selector exists to put the specific stamps on your specific hook and genre so the browse hears the sell, not the front matter.