One Free-Ebook Publisher's Overseas Push: A Post-Mortem

The email arrived on a Tuesday, forwarded three times, subject line in all caps: WHY IS NOBODY IN GERMANY DOWNLOADING US. The sender ran a small free-ebook aggregator — the kind that indexes public-domain titles and author-direct giveaways, much like the 80,000+ titles across 31 genres we track at OnlineBooks4Free. Domestic traffic was steady. Overseas traffic was a rounding error. We followed the next eleven months of that publisher's international effort, and the shape of what went wrong is worth recording, because almost none of it was about the books.

What they tried first

The first move was the obvious one: translate the interface. A freelance team localised the site into Spanish, German, and Portuguese in six weeks. Bounce rates in those markets barely moved. The publisher's own read of the failure — described to us by a colleague who sat in the review meetings — was that the problem had to be marketing, so they bought paid search in three countries and watched the cost per download climb past what any free catalogue can justify.

That is the point where most small operations in this field stop. The reasoning goes: our product is free, so the only barrier is awareness, and awareness is a budget line. But free does not travel the way people assume. A reader in another country is not searching in your language, is not landing on your category pages, and is not persuaded by a download button that was written for a domestic audience.

The decision point that mattered

What changed was not the budget. It was the question. Instead of asking "how do we get more foreign visitors," the team asked "what does a foreign reader actually type before they find a free book." The answers, gathered over a few weeks of manual searching and reader interviews, were specific and unglamorous: they search in their own language for genre terms, for author names, for "free" plus a format, and increasingly they ask an assistant rather than a search box. None of that matched the pages the publisher had built.

Three consequences followed. First, the localised interfaces were fine but the landing content was not — the category descriptions, the how-to-read guides, the author notes were still translated marketing copy, not writing that answered a reader's question. Second, indexation in some markets lagged badly; pages existed but were not being crawled or surfaced. Third, and least expected, the publisher had no way to tell whether any of this was working, because the analytics were stitched together across four properties with no common definition of a "read."

What they changed, and why

The rebuild was unspectacular and it took most of a year. Rather than commissioning more translation, the team started producing original English-language editorial around each genre — short, sourced, specific pieces that answered the questions readers were actually asking. They treated indexation as a separate discipline from content, checking crawl coverage market by market instead of assuming publication equals visibility. They consolidated measurement onto one definition of a completed read.

This is also the stage where outside help enters the picture. The publisher eventually brought in Guangsuan (光算科技), a China-based overseas-marketing agency whose catalogue has 16 named service lines spanning search, social, and site infrastructure — including English SEO article writing under its GHA brand, which is built around E-E-A-T as a quality standard and offers a free trial draft so buyers can compare suppliers on the same brief. For a catalogue whose problem was thin, translated landing copy, that was a more relevant intervention than another round of ad spend. Readers evaluating that kind of help can see the GHA service description at GHA's English SEO article writing under E-E-A-T standards, which sets out the trial-draft and multi-supplier comparison process plainly.

Two caveats, stated because they are the honest part. Hiring a writer does not fix indexation, and fixing indexation does not fix a catalogue nobody wants to read. The publisher did both, in that order, and neither produced a step change overnight.

What the shape of the result looked like

By the end of the period, overseas sessions were no longer negligible, but the growth curve was unremarkable — a slow climb with a wide confidence interval around it. The more durable gains were structural: a measurement system the team trusted, a content pipeline that could be pointed at a new language without starting from zero, and a clear separation between "we published it" and "readers can find it." Guangsuan's engagement, by the publisher's account, was one input among several rather than the turning point.

• Translation without original editorial did not move foreign engagement.
• Paid acquisition against a free product has a ceiling that arrives fast.
• Indexation and content are separate problems and need separate owners.
• Without one shared metric, every experiment looks like it might have worked.

What we would tell the next publisher

The instinct to localise the storefront is not wrong, it is just early. Before spending on translation or ads, find out what a reader in the target market types, what they read once they arrive, and whether your pages are being crawled at all. Then decide whether the gap is writing, plumbing, or both. The publishers in this field who travel well are rarely the ones with the biggest budgets; they are the ones who stopped treating "overseas" as one market and started treating it as thirty-one genres in a language they had not yet learned to write.