Vellum vs Atticus: A 3-3 Feature Tie and a Cost Crossover

Publishing-forum folklore insists Vellum only pays off for 20-book backlist veterans, yet the amortization math puts the true crossover near six titles. Vellum's $249.99 license carries a $102.99 premium over Atticus's $147, and at $40 an hour, the half hour Vellum saves per title converts that spread into exactly $20 of recovered labor per book. Priced that way, the expensive tool overtakes the cheap one on title six.

The catch is that most authors never get there. Independent survey base rates show the median self-published author releases one to three books a year, so the typical career closes out below the six-title line. For that majority the rational default flips to the cheaper tool: choosing Atticus over Vellum keeps $102.99 — about 41 cents of every dollar of the Vellum budget — while pure amortization never rescues the pricier license, which still runs $12.50 per title against Atticus's $7.35 at twenty books.

What remains is a counting exercise, not a leap of faith. Tally the titles already formatted plus those contracted for the coming year and compare the sum to six; past it, Vellum's saved hours pay, below it, Atticus's $147 does. Report the threshold rounded — round-off error compounds through chained calculations — and remember the feature scorecard ties 3–3, so cost makes the call.

Two Cost Curves, One Crossover

Content for Two Cost Curves, One Crossover is being prepared.

The Published Record

vellum.pub and atticus.com have posted the same numbers for roughly four years: verified against both vendors' live pricing pages in February 2026, vellum.pub lists Vellum Press at $249.99 and ebook-only Vellum at $199.99, while atticus.com sells a single $147 tier covering ebook and print together — prices both companies have held stable since roughly 2021-2022. Two details hide in that record. The spread driving the crossover analysis above is a derived quantity, not an advertised one; neither vendor prices against the other, so no product page will ever hand you the comparison. And the tiers are asymmetric: Atticus's one price includes print, while Vellum reserves print for Press — the $199.99 edition exports ebooks only. Any print-capable, apples-to-apples comparison therefore runs through Vellum Press, never the cheaper tier.

The risk-reversal terms differ in kind, not degree. According to vellum.pub's pricing and FAQ pages, Vellum's free trial allows building unlimited books and charges only when generating final files, with trial exports carrying a watermark. According to atticus.com's published guarantee, Atticus sells outright with a 30-day money-back window. That is output-gated insurance versus time-boxed insurance: Vellum lets you finish a full manuscript before paying anything, with no clock running, while Atticus's window can expire on a slow first project. Neither dominates — they insure against different failure modes.

The speed evidence carries a methods warning. Kindlepreneur's hands-on Atticus review and the Alliance of Independent Authors' tool guides report per-title formatting times for a styled 80,000-100,000-word genre novel of roughly 1-2 hours in Atticus versus 0.5-1 hour in Vellum — self-reported walkthrough figures, not instrumented benchmarks: no repeated trials, no variance, no shared test manuscript. Pairing the ranges' extremes yields a per-title edge anywhere from 0 minutes (one hour versus one hour) to 90 (two hours versus half an hour); midpoint-to-midpoint arithmetic gives about 45. The 30-minute edge anchoring the crossover above sits below that midpoint, so the six-title threshold is, if anything, ungenerous to Vellum — unless the true edge collapses toward zero, where no count breaks even and Atticus wins outright.

Reachability is a base-rate question. According to the most recent ALLi member income survey and Draft2Digital's published author surveys, the median self-published respondent releases on the order of one to three new titles per year. Simple arithmetic puts six paid titles two to six years out at that pace — inside a normal planning horizon, but not automatic. Here the persistent myth dies on contact: Vellum Press is not a trophy reserved for twenty-book backlist veterans. An author clearing three titles a year crosses over in year two; even one per year arrives inside most careers. The honest edge case cuts the other way — stall after two or three books and you never reach the crossover, making Atticus strictly correct.

Last, licensing behavior. Vellum's public changelog documents major-version updates delivered free to existing licensees since launch; Atticus advertises lifetime free updates on its own site. Quote both vendors' statements rather than trusting summaries, because this record is the only available predictor of future licensing behavior — Vellum's is a demonstrated multi-year history, Atticus's a promise younger than the behavior it predicts.

The Feature Ledger

Scored row by row, the Vellum-versus-Atticus ledger ends three to three, and that tie is the finding, not a dodge. Atticus sweeps every row you settle before formatting a single page — price, scope per dollar, platform reach. Vellum sweeps every row you live with afterward — type controls, throughput, maturity. Because the two triads sit on opposite sides of the fixed-cost amortization curve, there is no universal champion here; the overall winner is conditional on title count, and pretending otherwise is how buyers talk themselves into the wrong checkout.

Weight the rows by decision impact and the tie resolves asymmetrically. For a Mac author clearing six or more paid titles, the production-side triad dominates because the acquisition delta amortizes away: spread the premium across enough formatted books and its per-title share trends toward zero, leaving typography, speed, and maturity to govern daily working life. For a one-to-three-title author the logic inverts — the $102.99 premium, about 41 cents of every dollar of the Press budget, is permanent and unconditional, while the production advantages stay speculative until enough books ship to feel them. This is where both folk myths die at once. The cheaper sticker does not always win: Atticus taking three rows does not make it the right tool at title eight. And the Press price does not reserve Vellum for twenty-book backlist veterans: a debut novelist committed to a planned series crosses the identical threshold, because the test counts titles ahead, not years published.

Hybrid publishers should stare hardest at the scope row. Vellum splits its product line so print capability requires the Press upgrade, while Atticus bundles ebook and print conversion in a single SKU. That structural choice marks the row Atticus permanently, and it penalizes Vellum most sharply for authors who need print on a tight budget — a print-inclusive rival at a lower sticker attacks the gated tier directly. Note, however, that the penalty dissolves exactly where the count test activates: at the Press tier, print is included, so a six-plus-title Mac author forfeits nothing by buying up.

Bind every row back to the canonical rule and the ledger compresses to one sentence: below the crossover, buy Atticus; at or above it on a Mac, buy Vellum Press. No single row — not even typography, the row typesetters care about most — is permitted to override the count test. Treat the six-title threshold as a working estimate rather than a certified constant: according to the source fetch log behind this guide, the exact break-even cannot be verified from the vendors' published pages alone, because it swings on variables they do not disclose — update fees, future feature parity, and your realized per-title time savings. The concrete next step: log your formatting time across your next two titles. If your measured saving tracks the benchmark and your horizon clears six paid books on a Mac, the ledger has already told you which button to press.

What the Data Doesn't Tell You

Every timing figure feeding the crossover model is self-reported. The half-hour-per-title edge attributed to Vellum traces to vendor demos and reviewer walkthroughs — nobody has instrumented it against controlled manuscripts, and variance swamps the margin the moment interiors get complicated. Poetry line breaks, nonfiction endnote apparatuses, translated editions, and image-heavy illustrated books all demand manual correction passes neither tool automates cleanly; on those projects the edge can compress to nothing or invert entirely. The sensitivity logic is unforgiving: as the Crunched App breakdown on Medium notes, a breakeven formula doubles as a sensitivity test — halve the per-title saving and the crossover count doubles, erase it and no title count ever justifies the premium.

The base model also misses a hidden interaction on the Windows side. Running Vellum through CrossOver or a macOS virtual machine typically means a recurring subscription fee plus community-reported performance penalties and activation friction — and those penalties land squarely on the per-title time saving that justifies Vellum in the first place. The degradation compounds against the crossover in Atticus's favor, which is why the decision rule's Mac qualifier is load-bearing, not decorative: a Windows author clearing six paid titles hasn't cleared anything until they've timed their own virtualized workflow.

Averages hide capability gaps, too. Community forums carry recurring complaints of interface lag in Atticus on manuscripts above roughly 150,000 words — anecdotal and uncontrolled, but exactly the kind of signal to verify against current release notes before committing. On the other side, Vellum's print output centers on standard KDP and IngramSpark paperback specifications, with thinner hardcover-specific template support; a hardcover-led imprint plan should be checked against release notes, never assumed. Trust no feature table — including the one earlier in this guide — without a dated check against both vendors' current documentation.

Finally, survivorship bias shapes the whole evidence base. Public comparisons skew toward power users with large backlists who rationally chose Vellum; the silent majority who bought either tool, formatted one book, and stopped never writes a review. That selection manufactures both persistent myths — that Vellum is reserved for twenty-book veterans (the arithmetic says otherwise) and that the cheaper sticker always wins (it doesn't past six titles on a Mac). The longitudinal data needed to validate the six-title crossover barely exists, so treat it as a well-reasoned prior, not a measurement.

Before you buy: pull both pricing pages and timestamp them, then run one timed chapter of your actual manuscript through each tool's trial. If your catalog is poetry, endnote-heavy nonfiction, translation, or illustration work, your own stopwatch outranks every demo clock in circulation.

Run the Atticus column end to end: a $147 license plus six titles at 1.25 hours apiece burns 7.5 hours, worth $300 at her rate — $447.00 of total production cost for the backlist. The Vellum Press column opens $102.99 heavier but spends only 4.5 hours, worth $180: $429.99 all-in. Vellum Press wins the six-title scenario by $17.01, which is almost exactly the 0.85 surplus titles past breakeven multiplied by the $20 hourly saving — the premium wasn't barely repaid; it was repaid with most of a book to spare.

Worked Case

Step down to five titles and the sign flips. Atticus totals $147 + (5 × 1.25 × $40) = $397.00; Vellum Press totals $249.99 + (5 × 0.75 × $40) = $399.99. Atticus wins by $2.99. Solving the crossover directly, n* = $102.99 ÷ $20 = 5.15: the switch sits between titles five and six — a knife-edge, not a vague "eventually." Book six is not one point on a trend line; it is the entire decision.

Now stress the model's dominant variable: the wage. Cut her valuation to a hobbyist $25 per hour and the per-title saving halves to $12.50, stretching the crossover to $102.99 ÷ $12.50 = 8.24 titles — our six-book author stays on Atticus. Raise it to a professional $60 per hour and the saving grows to $30 per title, pulling the crossover down to $102.99 ÷ $30 = 3.43 — she should have switched at book four. Identical manuscripts, identical software, opposite answers. What her hour is worth decides the purchase; brand loyalty never enters the arithmetic.

The long horizon confirms the ordering. Amortize each perpetual license across twenty titles and the fixed-cost component alone falls to $12.50 per Vellum Press book versus $7.35 per Atticus book — derived directly from the vendors' posted prices — so the license gap keeps shrinking as n grows while the half-hour-per-title labor edge keeps paying out on every additional book. Nothing in this case requires a 20-book veteran, and that is precisely the myth to retire: a six-novel backlist crosses the line before the author drafts book seven. The mirror error — trusting the cheaper sticker — loses $17.01 on the same table.

Six titles. On a Mac, that single integer is the whole decision; everything else is tie-breaking noise. Held up as a fixed-cost amortization problem, the folklore dies quickly: the durable belief that Vellum's sticker price reserves it for twenty-book backlist veterans cannot survive the arithmetic, because a perpetual license plus the half-hour-per-title edge puts break-even at title six, not title twenty. The mirror error — trusting the cheaper sticker always wins — fails just as hard for any Mac author clearing six. Five threshold rules finish the job.

Rule 1 — Count, don't vibe. Tally the paid titles you will realistically format from March 2026 through February 2028, counting only finished files intended for sale; works-in-progress and "someday" reissues don't qualify. Anchor the forecast to your trailing window: whatever you actually exported in the past 24 months predicts the next 24 far better than intention does, so revise upward only for signed contracts. Below six, buy Atticus at $147. At six or more on a Mac, buy Vellum Press at $249.99. Ambition is the classic contaminant in this tally, which is why Rule 5 exists.

Rule 2 — No Mac, no debate. Vellum runs on macOS, so if you format on Windows or Linux and will not maintain a virtual machine, Atticus wins outright and the debate ends there. If CrossOver looks like a bridge, note that CodeWeavers lists it at roughly $74 per year — a recurring fee, unlike either perpetual license, and one that varies by plan, so verify current terms. Across the same 24-month window that adds about $148 to Vellum's side: $102.99 + $148 = $250.99, divided by the model's $20-per-title recovery rate, gives 12.55. The crossover slides from six to approximately thirteen titles, beyond nearly every indie author's planning horizon — and because subscriptions scale with time, the bar climbs higher every year you wait.

How to Choose Well: Five Threshold Rules

Rule 3 — Ebook-only publishers climb a shorter ladder. Strip print from the plan and the comparison becomes Vellum's $199.99 ebook-only tier against Atticus's $147: the premium compresses to $52.99, which at the same per-title rate lands the crossover at three titles. Publish three or more ebooks and the cheaper-looking tool loses on pure math. Before climbing, verify the vendor's current upgrade terms — whether the ebook-only purchase later credits toward Press decides whether this ladder telescopes or resets, and that policy is yours to confirm, not assume.

Rule 4 — Let genre complexity veto the clock. The half-hour edge is a mean, and means fail at the tails. Endnotes, poetry lineation, tables, and multilingual typography are exactly where the two engines diverge most, so discard the timing assumption entirely for such catalogs and default to Atticus unless you have personally timed Vellum winning on a comparable manuscript. The test costs one afternoon: during overlapping trials, format your ugliest chapter — endnoted history, a sonnet, a Spanish-language epigraph — in both tools and measure the delta yourself. If it inverts, the crossover model was never yours.

Rule 5 — Never prepay for unfinished work. Both vendors let you defer payment until evidence exists, so exploit the structure. Build the complete manuscript in Vellum's free trial and pay the $249.99 only at the moment of first export; buy Atticus inside its 30-day refund window only when a formatted manuscript already exists. This is lazy evaluation applied to licensing: bind the decision to a realized artifact, not a forecast, and a finished book converts your title count from speculation into data.

Run the rules top-down, because each can short-circuit the next: platform first (Rule 2), complexity second (Rule 4), the counts third (Rules 1 and 3), payment timing last (Rule 5). A Windows machine never needs a title tally; an unfinished manuscript never needs a price comparison. The table below settles every remaining case.

Rule 4 — Let genre complexity veto the clock. The half-hour edge is a mean, and means fail at the tails. Endnotes, poetry lineation, tables, and multilingual typography are exactly where the two engines diverge most, so discard the timing assumption entirely for such catalogs and default to Atticus unless you have personally timed Vellum winning on a comparable manuscript. The test costs one afternoon: during overlapping trials, format your ugliest chapter — endnoted history, a sonnet, a Spanish-language epigraph — in both tools and measure the delta yourself. If it inverts, the crossover model was never yours.

Rule 5 — Never prepay for unfinished work. Both vendors let you defer payment until evidence exists, so exploit the structure. Build the complete manuscript in Vellum's free trial and pay the $249.99 only at the moment of first export; buy Atticus inside its 30-day refund window only when a formatted manuscript already exists. This is lazy evaluation applied to licensing: bind the decision to a realized artifact, not a forecast, and a finished book converts your title count from speculation into data.

Run the rules top-down, because each can short-circuit the next: platform first (Rule 2), complexity second (Rule 4), the counts third (Rules 1 and 3), payment timing last (Rule 5). A Windows machine never needs a title tally; an unfinished manuscript never needs a price comparison. The table below settles every remaining case.

What to do next

Frequently Asked Questions

At how many formatted books does Vellum's time savings actually pay back its higher price?

At $40 an hour, the half hour Vellum saves per title converts its $102.99 premium over Atticus's $147 into exactly $20 of recovered labor per book, putting the break-even on title six.

If I ignore time savings and just amortize the license fees, does Vellum ever get cheaper per book than Atticus?

No pure price break-even exists, because both one-time licenses amortize toward zero — even at 20 formatted titles, Vellum still runs $12.50 per book against Atticus's $7.35.

What does Vellum's free trial actually let me do, and is there a catch on the files it produces?

Vellum's free trial allows building unlimited books and charges only when generating final files, with trial exports carrying a watermark.

How reliable is the claimed 30-minute-per-book formatting advantage for Vellum?

Kindlepreneur's review and ALLi's guides report roughly 1–2 hours in Atticus versus 0.5–1 hour in Vellum for a styled 80,000–100,000-word novel, but these are self-reported walkthrough figures rather than instrumented benchmarks, yielding a per-title edge anywhere from 0 to 90 minutes.

Can I run Vellum on my Windows laptop or Chromebook?

No — Vellum Press runs on macOS only, while Atticus covers Windows, macOS, Linux, and Chromebook.

Could I buy the cheaper $199.99 Vellum edition instead of Press and still format print books?

No — the $199.99 edition exports ebooks only, sits $52.99 above Atticus while omitting print, and is therefore dominated in this comparison.

Quick answers

Also worth reading: OK (the original title is strong and accurate): OK (the original title is · Step-by-Step Guide Converting Word Documents to EPUB Format in 2024: Step-by-Step Guide Converting Word Documents · Pandoc Lua Filters: Why EPUB Conversion Time Isn't What It Seems: Pandoc Lua Filters: Why EPUB