Selling to 12,000 People, Not 12 Million: Niche Audience Targeting

A specialist publisher is not trying to reach everyone. The realistic buying audience for a serious book on a narrow subject might be ten or twenty thousand people in the entire English-speaking world. Every targeting decision follows from that number, and most standard advertising advice does not.

Why broad targeting misleads specialists

Platform automation is built on the assumption that your product has a large addressable market and the system’s job is to find the best segment of it. Give a broad campaign a book on Soviet optics and it will spend the first several hundred pounds discovering, expensively, that most people are not interested in Soviet optics.

For a mass-market product, that learning phase is an acceptable cost. For a specialist publisher, the learning phase can consume the entire budget for the title.

Start from what you already own

Before touching interest targeting, exhaust the audiences you already have. These are always better and always cheaper.

Your customer list, uploaded as a custom audience. Your email subscribers, segmented by which subject they signed up through. Website visitors from the last one hundred and eighty days, split by which book pages they viewed. Anyone who has watched more than half of a video about a related title.

These are people who have already demonstrated the specific interest. Retargeting them is not sophisticated, but it is where the return actually is.

Lookalikes from purchasers, not from everyone

When you move to prospecting, build the seed audience from buyers rather than from all site visitors. A lookalike built on people who spent money on a £40 hardback is a fundamentally different audience from one built on everyone who bounced off your homepage.

Keep the percentage tight. A one per cent lookalike from a purchaser list of even a few thousand people will outperform a five per cent lookalike built on a larger, vaguer seed, because for narrow subjects similarity degrades fast.

Stacked interests, and the association trap

Where platform interest targeting genuinely helps a specialist publisher is in combination rather than in isolation. A single interest is too broad. Two or three overlapping interests describe an actual person.

Someone interested in military history alone is a very large and very mixed group. Someone interested in military history and a specialist magazine in that field and a museum or archive is recognisably a reader of serious books on the subject.

The trap is adjacent-hobby association. Interest in a historical subject overlaps heavily with model kits, wargaming, reenactment, and video games in the same setting. Those are real interests held by real people, and most of them do not buy £40 reference books. Watch the placement and audience breakdowns for segments that generate clicks and no purchases, and exclude them explicitly.

Where the audience already gathers

For genuinely narrow subjects, the highest-intent audiences are not on the big platforms at all. They are in specialist forums, subject society newsletters, subreddits with four thousand members, YouTube channels run by one person with a camera and deep knowledge, and the mailing lists of dealers and archives.

These places do not have self-serve advertising, which is exactly why they remain undervalued. A newsletter placement in a subject society bulletin can outperform a month of platform spend, because everyone reading it has already self-selected. Approach them as a publisher rather than an advertiser: offer review copies, extracts, a piece written for their audience.

The metric to hold onto

For narrow subjects, ignore reach and impressions entirely. They measure the thing you are deliberately not trying to do.

Watch cost per purchase and repeat purchase rate. If a campaign reaches twelve thousand people and produces two hundred buyers who go on to buy again, it has done its job better than one that reached a million and produced three hundred one-time buyers at a higher cost.