The full record — how publishers have approached selling direct, 1926 to today

The Book-of-the-Month Club (1926) and Literary Guild (1927) built subscription businesses that sold books by post: curation, convenience, and the negative option. Booksellers answered with boycotts and the industry's first regulatory war over direct selling.

Reader's Digest and Time-Life later worked out lifetime value, churn and subscriber inertia on index cards. They called it 'continuity marketing', decades before SaaS vocabulary existed.

• T1 MARGIN
• T4 COMMUNITY
• T5 CONFLICT

Most publisher websites stayed brochures while Amazon absorbed the online demand publishers imagined capturing. The exceptions were instructive: Baen's Webscriptions (1999) sold DRM-free ebooks direct to a community of science-fiction devotees, and O'Reilly translated its direct-mail heritage into a real customer database.

The publishers who succeeded direct were vertical, community-dense, and DRM-sceptical. Everyone else learned that a storefront is not a strategy.

• T1 MARGIN
• T3 PLATFORM RISK

The Kindle and $9.99 pricing taught publishers they no longer controlled the retail price of their own digital products; the agency fight and the 2012 DOJ suit played out in public.

Pottermore proved that sufficiently valuable IP could force even Amazon to link out to a direct store, and its later retreat proved how exceptional the case was. By 2012 the trade press was declaring 'the year publishers embraced selling direct.'

• T3 PLATFORM RISK
• T1 MARGIN

Mike Shatzkin traced the full arc inside his own writing: direct relationships as 'an essential survival skill' (2012), then the sober correction (2015) that for most trade publishers D2C costs more in lost sales than it gains.

Joe Wikert built the opposite case: drop DRM, sell direct, and treat the reader's name and email as the most important data a publisher can own. Every debate since has occupied the ground these two staked out.

• T2 DATA
• T5 CONFLICT

Oyster paid publishers full wholesale rates and could not survive its own heavy readers; it shut in 2015. Kindle Unlimited survived by paying authors from a fixed pool — same value, risk transferred downward.

Publishing drew two conclusions and has repeated them ever since: platform-controlled subscription devalues the book, and publisher-friendly subscription dies of its own economics. Any modern subscription pitch must answer the Oyster question first. Oyster's library of one million titles closed within four years of opening.

• T1 MARGIN
• T3 PLATFORM RISK

Kevin Kelly's '1,000 True Fans' (2008) supplied the theory; Amanda Hocking the proof; Hugh Howey the polemic. Joanna Penn codified the playbook: more profit per sale, faster payment, and the customer data Amazon never shares.

By 2025, romance authors were reporting 60–80% of revenue from direct channels. The vanguard's current edges — direct audio, crowdfunded pre-orders, serial-first funnels — are the leading indicators of what publisher discourse adopts next.

• T4 COMMUNITY
• T2 DATA

Big houses experimented (11 Main, Aerbook, Profile's three consumer sites, Open Road's million-page-view community) while the centre of gravity stayed sceptical.

Then O'Reilly, the publisher with the best direct franchise in the industry, closed its ebook store in 2017. Jane Friedman's verdict: the infrastructure direct commerce demanded might be unsustainable for publishers. GDPR (2018) had a delayed effect: consented first-party data became harder to gather, and worth more.

• T6 BUILD VS BUY
• T2 DATA

In March 2020 Amazon deprioritised book deliveries. Publishers had spent a decade debating platform risk as a theory; now it had a date. Bookshop.org grew 40-fold within months of launching; Microcosm Publishing's mail order rose 384%.

As publishers built lists and CRMs, booksellers noticed, and the 1927 channel-conflict pattern resurfaced in modern form at Winter Institute 2023.

• T3 PLATFORM RISK
• T2 DATA

Brandon Sanderson's Kickstarter raised $41.7 million for four self-published novels, the largest direct-to-reader transaction in publishing history. Hachette became the first Big Five house to transact direct from its own site (December 2023). BookNet found almost all surveyed publishers selling direct.

And Anne Trubek supplied the margin arithmetic: direct sales were a quarter of Belt Publishing's volume, and half its revenue. The industry's own measurement (BookScan) excludes direct sales entirely; the channel is systematically undercounted.

• T1 MARGIN
• T2 DATA
• T4 COMMUNITY

The TikTok ban scare crystallised the fear: 'a large following on TikTok is not a community, it's a platform.' AI answer engines extended the logic: if you're not crawlable, you're not citable. Spotify and YouTube re-ran the platform-risk play in audio.

Meanwhile turnkey D2C stacks reached the mid-market: HarperCollins created an SVP-level D2C role, IPG partnered on a reader app, PRH launched vertical platforms. The century-old build-vs-buy objection ('publishers are not tech companies') has resolved decisively toward buy.

• T3 PLATFORM RISK
• T6 BUILD VS BUY
• T2 DATA